Artwork for podcast RevOps FM
The MarTech Developer - Sanford Whiteman
Episode 6 • 6th November 2023 • RevOps FM • Justin Norris
00:00:00 01:00:02

Share Episode

Shownotes

Sanford (“Sandy”) Whiteman is a legend in the Marketo world.

A software developer by trade, he’s one of a small handful of engineers who have focused specifically on marketing platforms and problems - not merely as a one-time project or side interest but as a dedicated speciality.

Sandy is the #1 all-time contributor in the Marketo product community with over 23,500 posts and an honorary community moderator designation.

He's also been a great friend and teacher to me over the years and a collaborator on many projects.

Thanks to Our Sponsor

Many thanks to the sponsor of this episode - Knak.

If you don't know them (you should), Knak is an amazing email and landing page builder that integrates directly with your marketing automation platform.

You set the brand guidelines and then give your users a building experience that’s slick, modern and beautiful. When they’re done, everything goes to your MAP at the push of a button.

What's more, it supports global teams, approval workflows, and it’s got your integrations. Click the link below to get a special offer just for my listeners.

Try Knak

About Today's Guest

Sanford Whiteman is Chief Technologist at FigureOne, a Marketo development agency.

Key Topics

  • [00:00] - Introduction
  • [01:05] - Sandy’s background and how he entered the Marketo world. His start as an IT security person. Developing an app for the textbook publishing industry. First Marketo project: to fix Munchkin. Reverse engineering the Munchkin script. Developing fluency in the Marketo UI.
  • [04:58] - Why Sandy took the Munchkin script seriously. Enjoyment of niche knowledge and desire to discover what is currently unknown. Previous experience reverse engineering Microsoft SMTP server. People even at Marketo/Adobe not fully understanding the product internals. They are not front-end people.
  • [08:17] - Justin’s perspective on when Sandy entered the Marketo community. People guessing or speculating at how things worked technically but without a technical background. The contrast when a real developer focused on the same topics. Sandy’s desire to not guess and to spread accurate knowledge about systems. How early developers scoffed at cloud-based systems and refused to support them. How developer skepticism towards MarTech systems may be rooted in a resentment from the IT team. IT questioning why the company would pay for a system they believe (erroneously) could be built in-house.
  • [11:49] - Why “mainstream” developers who are otherwise talented produce bad work in the MarTech world (e.g., poor integrations) and show a lack of curiosity about it. Disparagement of marketing within the developer community. Lack of understanding of core components of marketing automation: SMTP, landing pages, databases. Belief that these are unsophisticated technology. Internal IT resources haven’t had to manage email servers for 15 years or more. That knowledge has died out, leading to lack of understanding. Perceived primacy of back-end developers over front-end developers. Disparagement of JavaScript.
  • [20:49] - The rationale for investing hours of time investigating obscure issues. How mastery of obscure knowledge becomes useful when those situations do arise.
  • [24:32] - The challenge of being a technical consultant to marketers: you’re downstream of strategy. Discussion of how it feels to work on marketing projects where the strategy is flawed or creative seems ineffective. The political challenges of giving feedback as a consultant. Sandy’s frustrating situation with a client claiming the Marketo Lead ID is PII. How there’s also a prestige that can come with being a consultant.
  • [32:19] - Sanford’s perspective on the product outlook of Marketo and whether the glory days of the platform are behind it. Lack of innovation in recent times. If Sandy was in charge, he would shut down imitation / look-alike products (Adobe Connect, Live Chat) as these won’t successfully replace more full-featured applications.
  • [33:42] - Sandy and Justin build a fantasy roadmap for Marketo. The importance of getting ahead of the market and deeply understanding the needs of the customer. Questioning whether anyone at Adobe is filling that role now with respect to Marketo. Lack of focus on effective marketing. Smaller companies with dynamic founders seem more likely to come up with these ideas.
  • [35:55] - Feature idea: ability to customize the UI in Marketo (like a Visualforce page in Salesforce). The challenge of rebuilding an existing app interface using legacy technology. Why Sky failed.
  • [41:59] - Feature idea: ability to analyze records in the database in relation to each other. (E.g., create quartiles.) Discussion of new Activity Stream feature (outbound streaming API).
  • [45:52] - Feature idea: better data manipulation within flow steps (like what is possible to do in FlowBoost).
  • [46:33] - Feature idea: more control over order of operations within the system. Guaranteed ways to avoid race conditions. Question of whether this is solvable within marketing automation architecture. The ability to have visibility of database changes yet to be committed. How Marketo provides a toolkit that requires developer-esque conceptual understanding.
  • [51:13] - Discussion of AI and its potential impact on marketing technology. One potential application is using it to generate documentation. AI beneficial as a convenience tool (like the computer on a Star Trek ship). Justin expresses skepticism about its generative abilities for marketing content. Challenges for AI with generating or explaining humour.

Resource Links

Learn More

Visit the RevOps FM Substack for our weekly newsletter:

Newsletter

Transcripts

Justin Norris:

Now this may seem totally wacky to anyone born after

2

:

1990 or so, but One of the main

benefits of marketing automation

3

:

in the early days was to liberate

marketers from a dependence on IT.

4

:

So the vision there was that marketers

could create their own emails and

5

:

landing pages, make complex campaigns,

do all the cool stuff they wanted to do.

6

:

And they didn't need to

know any code to do it.

7

:

Did that actually happen?

8

:

I think sure it did to some

extent, but we also saw a few

9

:

other things in trying to do that.

10

:

We saw forms that were breaking, we saw

systems that weren't scaling, integrations

11

:

that failed, and just a lot of different

requirements that couldn't be met by

12

:

the way platforms worked out of the box.

13

:

So that brought us back to step

one, which is that we still needed

14

:

developers, but we needed a special kind.

15

:

Someone who had the technical

skills, but also a deep

16

:

knowledge of the marketing domain.

17

:

So effectively a martech

developer and for my money.

18

:

Today's guest is the best

martech developer there is.

19

:

He's a legend in the Marketo

community, one of the smartest

20

:

folks I know, and a good friend.

21

:

Sanford Whiteman, welcome to the show.

22

:

Sanford Whiteman: Thank you, Justin.

23

:

I was afraid that had

something to do with me.

24

:

Justin Norris: So, Sandy, we've

known each other eight, nine

25

:

years, getting dangerously

close to ten years at any event.

26

:

And it's weird that I've never

really asked you too much

27

:

before about your background.

28

:

Could you just walk us through how

you got from point A to point B?

29

:

Sanford Whiteman: Yeah, sure.

30

:

I used to be an IT person.

31

:

I mean, I was a security

and firewall person.

32

:

I was the hard ass Cisco

guy who wouldn't let...

33

:

Any, ports through the firewall, even

if, supposedly some, some mission

34

:

critical app demanded it, I was paranoid

about, I still am, of course, I bring

35

:

all of this experience to the table,

but I try to not be like that anymore.

36

:

Nevertheless, I did have that background

first, that got boring and stressful

37

:

simultaneously, as a lot of people will

tell you about working in security,

38

:

you know, you're always thinking

that your phone, let's call it or.

39

:

Pager or whatever, or whatever you have

at the time is, is, uh, you know, you're,

40

:

you're basically, you're thinking I missed

something and any call you get, you know,

41

:

you know how it is in our world now.

42

:

But anyway, that was like

a nerve wracking thing.

43

:

And I started working with

a friend on web app stuff.

44

:

And we built this ridiculous

app that still exists.

45

:

I hope nobody hears me say

ridiculous, but it's great.

46

:

It's still runs.

47

:

It's still used by people in the textbook

publishing industry of all the weird.

48

:

Niche industries a Notoriously slow

to digitize world where our app built.

49

:

I cannot tell you how many years

ago it would be embarrassing.

50

:

Like my, my lips can't form the

number of years ago that it was

51

:

built, nor the number of years ago

that it last had a version upgrade.

52

:

But it's still the best one out there.

53

:

But that got a little bit weird because

we couldn't find any more clients for it.

54

:

We had like conquered the few enterprises,

I don't want you to think it makes

55

:

a lot of money, but it's just like

we had covered every large publisher

56

:

because there's so much consolidation.

57

:

And yeah, somewhere in there,

my partner started working for

58

:

a financial services company.

59

:

And after a few years there, he brought

me in to look at their MarTech stack

60

:

because he knew that I knew web stuff.

61

:

So my first experience with

Marketo, my first experience

62

:

was troubleshooting munchkin.

63

:

Like I had never used, I had used

Salesforce or other sales, you know,

64

:

CRM type stuff, but I had never

used a marketing platform before,

65

:

but their munchkin wasn't working.

66

:

It wasn't associating leads from

clicked emails, you know, the

67

:

classic thing it's supposed to do.

68

:

So I took this weird route in where I

was brought in not as the reluctant,

69

:

you know, you get the web team and they

hate to work with marketing and there's

70

:

all that kind of resistance to doing,

but it was like, you have to solve this.

71

:

This is your job undistracted.

72

:

So I kind of, you know, I want to say

reverse engineered, but but figured out

73

:

a lot of things about the tech side.

74

:

of Marketo first before

doing anything else.

75

:

And the form side, like you said,

you know, seeing that there's this

76

:

open, adaptable, customizable thing.

77

:

This was still like forms 1.

78

:

0 days, but I could tell, like,

this is a thing where there's

79

:

nobody, there's no Martech

developer, there was no such thing.

80

:

You know, I, I might've been one

of the few people in the world

81

:

at that time that just was.

82

:

Allowed to only do MarTech development,

but it was only because, you know,

83

:

big shout out to my friend, Brian,

if you ever, here's this, he brought

84

:

me in and he said, give this person

this project and this project only.

85

:

And from there, of course, you know, I

got Marketo admin access and, you know,

86

:

and then, and then there was a while

that people assumed that I knew how

87

:

to use the Marketo UI, you know, smart

lists and smart campaigns and flows.

88

:

And I kind of didn't, I knew how

to use the API and I was answering

89

:

any question anybody had about

that I could answer, right.

90

:

But it took me a few years to then become

the person that most of us are, you

91

:

know, with the fluency on the backend.

92

:

And yeah, so since then I've been

trying to be the master of, to retain

93

:

the mastery, you know, of the tech side

and the backend campaign side, which is

94

:

hard because you can go down a rabbit

hole of troubleshooting some tech thing.

95

:

And that takes hours or days that other

people are using in the UI itself.

96

:

But anyway, so that's kind of the arc.

97

:

And now, luckily there's enough

little problems to solve all over

98

:

the place that that's what I do.

99

:

Justin Norris: So to zoom in on

that first experience where you were

100

:

essentially, you know, it sounds

like locked in a room and said, don't

101

:

come out until you fix Munchkin.

102

:

I mean, the fact that you were given

that opportunity is one unique thing,

103

:

but then I think a lot of developers

would maybe get it done and move on

104

:

and somehow you, you sort of like to

take something apart, understand all

105

:

the component pieces, appreciate it and

know how to put it back together again.

106

:

And you didn't dismiss it as like, it's

just this crummy little marketing script.

107

:

Sanford Whiteman: Right,

108

:

Justin Norris: I'm kind of above that.

109

:

Somehow you, you really saw

something to make you go deep into

110

:

that and got interested in it.

111

:

And I'm just curious,

where does that come from?

112

:

Is that a you thing?

113

:

Or did you perceive actually I

could really get good at this

114

:

and then nobody else is doing it.

115

:

Like, did you think of it that way

or was it just, this is interesting.

116

:

I'm here.

117

:

Let's figure it out.

118

:

Sanford Whiteman: I don't deny

that there's a motivation.

119

:

An unconscious motivation of probably

no one else knows how to do this.

120

:

I don't think of that like, you know,

from a career standpoint, more of a

121

:

I want to be the one who knows it.

122

:

But I'm not even thinking that,

I'm just like, I'm feeling it.

123

:

Before I did the Marketo stuff,

there was one other thing that I

124

:

was like super expert weirdo in.

125

:

And it was the Microsoft SMTP server that

came with IIS five and six and seven, I

126

:

guess, but I had stopped by the time is

IIS six and I was on the MSDN communities

127

:

and got so deep inside figuring out

how this closed source thing worked.

128

:

I mean, you want like a

hard reverse engineering.

129

:

It's like, you don't get to see

the Munchkin JS library and sort

130

:

of try to unminimize it and figure

out what's going on inside it.

131

:

Like you have nothing else to go on

except timing things and like the way

132

:

it interacts with the file system and

windows registry and all kinds of little

133

:

things like that, but I became obsessed.

134

:

Like I wanted to be the one

who really understood it.

135

:

Cause I kind of, I could.

136

:

Kind of get the feeling like to your point

that there were people providing generally

137

:

correct answers, but trying to gloss

over the fact that they really didn't

138

:

understand why something worked or not.

139

:

So that was kind of an automatic thing.

140

:

That was probably the most immediate

precedent for my Marketo and Munchkin

141

:

thing was like, I need to know how

this works because it feels like

142

:

there's a missing store of knowledge,

even within Adobe then Marketo.

143

:

You know, like that's, that's the tragedy

to me is like, I felt like no one, and

144

:

this has grown over the years, no one

within the company that owns this product

145

:

currently, at least knows how it works.

146

:

And that is the kind of thing that just

irks me at some core level, I guess.

147

:

So that's, what's kind of kept me going.

148

:

If I felt like, and I think this

is kind of why I could never do the

149

:

same thing for Salesforce because.

150

:

Over at Salesforce, you're like,

okay, the engineers know how it works.

151

:

You know, we're just playing around

in the community and trying to show

152

:

off expertise or whatever, but it's

like, you're never really going

153

:

to get, you don't get that feeling

like you're, you're going to reveal

154

:

something to the people that work there.

155

:

But I think with Marketo, there's that

sense of like, I can actually show

156

:

the people who built it, how it works.

157

:

And not just because of

institutional memory and.

158

:

Stuff like that, or, or, you know,

people departing, but because sometimes

159

:

they don't really, uh, and I say this

with as much kindness as I can muster,

160

:

they don't really understand how

JavaScript and how browser stuff works.

161

:

So they're not really front end people.

162

:

So you really kind of do have to

say the way your forms library is

163

:

built, struggles, or does interesting

things under these conditions.

164

:

And it's not like you're have anyone

inside has been like, well, of course I

165

:

knew that it's more like, yeah, well, we

kind of wanted to get it out the door.

166

:

It is like a drive.

167

:

It is that feeling of

like, no one knows this.

168

:

I'm not interested in, in doing it.

169

:

If I feel like someone knows it.

170

:

Justin Norris: It's a mountain to be

climbed that no one has climbed before.

171

:

I want to describe to you, from an

outsider's point of view, my perception

172

:

when you came onto the scene.

173

:

Because at the time it felt like there

were very few technical people, hardly

174

:

anyone that I remember in the community.

175

:

And it was really a lot of demand gen

marketers or former demand gen marketers.

176

:

Some people maybe that self identified

as marketing ops and so people would

177

:

come up with solutions and things or

maybe Speculate and it was always sort of

178

:

impressive people who seem to have more

knowledge of the inner workings But they

179

:

maybe didn't really know or didn't have

the expertise and then all of a sudden he

180

:

emerged and it was like Oh, you're someone

actually technical and it was kind of

181

:

like, you know a bunch of gophers trying

to drive a car or something and all of a

182

:

sudden a formula one racer comes out and

is Like actually here's how you do it.

183

:

So to your point, I think It was

a really valuable service and you

184

:

alluded about your security leanings

and trying to purge yourself of that.

185

:

But I think that's part of the recipe

from my point of view is that nobody else

186

:

really even understood the principles of

what a DDoS vulnerability would be, or,

187

:

or, you know, like those things, like

people just didn't think about that stuff.

188

:

And then you, at least for me, as

an enterprise marketing operations

189

:

consultant really helped give me

that vocabulary to then sort of

190

:

absorb into my lexicon and take to

clients and be like, well, actually.

191

:

To me, all those components became part of

the recipe of what you did that is unique.

192

:

Sanford Whiteman: Yeah.

193

:

I think you're right that one thing I

wanted to do is say, let's not guess also.

194

:

That's another like painful thing

when you see like a really well

195

:

intentioned person who nevertheless

is just putting up a best guess and

196

:

it's like, I don't think it's that.

197

:

Let's know as much as we can and like

have a good faith effort to really,

198

:

you know, bring together knowledge

of databases and networks and just

199

:

general enterprise architecture.

200

:

I, I, you know, I give full credit to the

people who didn't have that background

201

:

though, because they didn't have the, I

mean, frankly, like years of experience,

202

:

which you can't make that up out of

nowhere, but they also had, they, they,

203

:

you know, still most of the people in mobs

are technically minded, or I guess maybe

204

:

I can say like logically expert, but they

don't have a hands on experience because

205

:

they didn't really come from that world.

206

:

It's hard to extract yourself from the

world where you're, uh, You know, the

207

:

jerk in the IT bullpen who doesn't want

to do any work for marketing or sales.

208

:

That was the thing that I was alluding

to at the beginning is like, I come

209

:

from the original cohort of people

who, when you said you were going to

210

:

move to a cloud CRM said, well, so

much for getting any help from us.

211

:

You, you know, you do that.

212

:

You'd say to the sales team, you

go to Salesforce, don't call us.

213

:

Like I was that guy.

214

:

Now it's silly because I'm

dealing with the IT people.

215

:

And I'm the person saying, you

need to take this seriously.

216

:

This is still part of the company.

217

:

This is not.

218

:

You can't actually ignore it.

219

:

And a lot of that is coming not from

pure cost, I think, which sometimes

220

:

is valid, but it's coming from that

still lingering, almost decades

221

:

old resentment from the IT team.

222

:

Why are we paying for a thing I

believe we could build in house, that

223

:

I believe we had a better version

of before, you know, when I was.

224

:

Young and strong and, you know,

had the energy to do stuff.

225

:

That's a funny interaction sometimes.

226

:

Cause people will say

like, we could build this.

227

:

What, what does this thing do?

228

:

We could build it in house.

229

:

Like I've had people say that about

Marketo, which is just so insane to me.

230

:

I'm like, how many developers are

you going to put on this full time?

231

:

But again, the flip side is I

remember being that person being

232

:

looking, especially the earliest

cloud solutions and thinking as a web

233

:

developer, we could totally do this.

234

:

Why are we paying for it?

235

:

And, you know, those

questions still swirl.

236

:

Well, you're paying for it because you

don't have enough time to have a full

237

:

dev and QA staff working on it, and you

wouldn't save any money on that anyway.

238

:

Justin Norris: build a workflow

every time marketing wants one.

239

:

And,

240

:

Sanford Whiteman: right, right,

241

:

Justin Norris: a campaign.

242

:

I mean, so speaking of these

developers, whether it's sort of your

243

:

classic IT developer at like a big

company, or whether it's your cool,

244

:

you know, Silicon Valley software

engineer type, these are smart people.

245

:

And yet they, many of them, let's say,

have produced a ton of really bad work

246

:

when it comes to MarTech, whether that's

Applications with horrible or unusable

247

:

Marketo integrations that they've

built in house or, you know, people

248

:

doing, people doing unspeakable things.

249

:

But these, these are people who

had the same opportunity as you,

250

:

let's say, to take apart Munchkin.

251

:

They haven't stepped up to it, and

it's not, as, as you've said to me

252

:

in the past, that they're, it's not

that they're like bad developers or

253

:

something, but, like, what is it?

254

:

Is it a lack of intellectual curiosity?

255

:

Is it a feeling that it's beneath them?

256

:

Why, why do traditional

developers struggle when it

257

:

comes to the Martech stack?

258

:

Sanford Whiteman: I mean I think

there is a disparagement of marketing.

259

:

They don't think it's

doing anything really.

260

:

If you somehow have told yourself that

SMTP, which they don't understand and

261

:

landing pages, which unless they're real

front end people, they don't understand.

262

:

And all the database and workflow

stuff and queuing stuff that a

263

:

sophisticated MA has, you know.

264

:

If you've told yourself that

that's not real technology or it's

265

:

unsophisticated technology, then

you're going to keep sticking with it.

266

:

Even if somebody shows you like I've

had this moment, maybe you have two

267

:

where someone has been skeptical.

268

:

They've taken that same view.

269

:

Oh, what does this platform do?

270

:

What, you know, it just looks

like it's a landing page

271

:

builder or something like that.

272

:

And you show them on a good day,

you show them some features and you

273

:

can actually blow somebody's mind.

274

:

Like I've, I've had to give like

some props to people who have

275

:

evolved some people, like Didn't

realize how sophisticated the

276

:

Marketo to Salesforce sync is,

277

:

for example, although that's

still, you know, firmly within

278

:

the realm of sales and marketing.

279

:

It's still it's clearly a piece

of engineering that's well done.

280

:

And, um, and I've seen people

like realize that they thought

281

:

they knew how email works.

282

:

And they don't.

283

:

And that's one of those things where

they talk to the IT team and say, can,

284

:

can't we just send this all ourselves?

285

:

And the IT team will be like, well, not

really, because, uh, you can't really

286

:

do that with, you know, Outlook, um, and

like, start to realize that there is a

287

:

sophistication to the Marketo And, or,

you know, or, or HubSpot or anything, you

288

:

know, queuing and resource balancing and

even the multi tenant architecture itself.

289

:

Like I've talked to people who don't

really, cause they're, they're still

290

:

like in house developers, so they,

they've never built a cloud app.

291

:

So they don't think about

what it would be like.

292

:

Now, Marketo struggles at multi tenancy

as we know, but nevertheless, the idea

293

:

that, no, it's not just, they don't

give you a machine, you know, like

294

:

that resource balancing across the

different customers is something that.

295

:

In house developers usually Don't care

about, they actually will have an app

296

:

that could be overwhelmed by one business

unit, but they don't really think

297

:

about that as a concern because they're

like, well, it's a fireball offense.

298

:

If you try to overwhelm the internal

line of business apps, we're

299

:

not going to worry about that.

300

:

But the fact is like doing that and having

a public facing site, all the reputation

301

:

management, email, reputation management,

the last internal it people or developers

302

:

in general understood anti spam tactics

and things like that was probably 15.

303

:

Years ago or more, I guess

there's a long way of saying.

304

:

To be optimistic about it, when

those complexities are revealed

305

:

to people, many of them have an

epiphany and they realize that it

306

:

actually is sophisticated technology.

307

:

But to your original point, I think you're

right that there's just the thought,

308

:

immediate thought is this is nothing.

309

:

It's simple.

310

:

Why do we even need it?

311

:

Just gets like caught up in

312

:

Justin Norris: The lightbulb that

just went off to me with what you

313

:

were saying is that through, like what

I mentioned in the beginning, like,

314

:

we're going to free you from dependence

on IT and you can do your own thing.

315

:

And then we did our own thing for a while,

and then all of a sudden IT forgot a lot

316

:

of that knowledge that they used to have

317

:

Sanford Whiteman: Yeah.

318

:

Yeah.

319

:

Uh huh.

320

:

Uh huh.

321

:

Justin Norris: And also, like,

like you said, that there is a kind

322

:

of primacy of back end developers

over front end developers.

323

:

Like,

324

:

Sanford Whiteman: Mm

325

:

hmm.

326

:

Mm hmm.

327

:

Sure.

328

:

Justin Norris: requires

more skill and expertise.

329

:

And yet, you can, you know, write

really crazy things in Python, but

330

:

maybe you don't know how to debug

a web script or figure out why a

331

:

form isn't submitting properly.

332

:

Sanford Whiteman: I mean, you spend

your time disparaging JavaScript

333

:

and saying it's a joke language.

334

:

That's, I mean, that's the

most ubiquitous language.

335

:

It is the language of the web.

336

:

You know, there are confusing

things about JS, but it sure

337

:

as heck is not easy to learn.

338

:

And I think a lot of those people are

just running away from how hard it is.

339

:

They just assume, especially now to get,

you were going to ask about like AI stuff.

340

:

They always think, Oh, well, I can just

run it through chat GPT to translate

341

:

my Python, but you can't because

there's no client Python that you

342

:

could write that would trend, you know,

like there's no, I think it's great.

343

:

It's almost like.

344

:

It's brilliant because it's a little bit

of job security too, because you, if you

345

:

don't have some JS skills to start with,

there's like nothing you can translate,

346

:

you know, automatically to that world.

347

:

You have to understand all of these

arcane and honestly, if you read like

348

:

the, the ECMAScript standard, where I'm

always going to like prove why things

349

:

don't work, whatever, the ECMAScript.

350

:

slash JavaScript standard or like,

you know, HTML five, like official W3C

351

:

standards, they're really hard to read.

352

:

Like you have to, you might wave

your hands and be like, Oh, I don't

353

:

need to know that it's dumb, but I

mean, it just doesn't quite work.

354

:

Like it's not, it's not dumb.

355

:

It's actually.

356

:

Arcane and hard and not

as satisfying as Python.

357

:

If you're doing like, you know, quant

stuff in Python or, you know, any kind of

358

:

backend work, you're probably going to get

more immediate satisfaction out of that.

359

:

But for me, coding is never

about immediate gratification.

360

:

For me, I think that that's one of

the things that makes people scared,

361

:

scared, like you said about, you know,

scared of troubleshooting a form.

362

:

Like, you know, somebody opens

a thread on Marketo Nation and

363

:

I'm always like, okay, I know I'm

going to figure out how this works.

364

:

I'm not going to let this just

sit there or, you know, why,

365

:

why, why the form is breaking.

366

:

But I'm also like.

367

:

This could take two to three hours and

sometimes it does just running it through

368

:

a proxy server, figuring out, simulating

the conditions that might have led to

369

:

a form sometimes not submitting and

sometimes submitting, you know, that

370

:

those, those anomalous or seemingly

random behaviors are some of the hardest

371

:

things you just don't really get that

if you're a backend developer, you don't

372

:

have, I don't want to generalize there.

373

:

I mean, you could introduce bugs that

create what seems to be anomalous

374

:

behavior, but there's nothing To

compare to the weirdness of web

375

:

development, like to the whole, like,

I think it's randomly failing to submit

376

:

the form, but then I'm always like,

there's no such thing as randomness.

377

:

It has something to do

with network conditions,

378

:

Justin Norris: truth is out

379

:

Sanford Whiteman: figure out what it

is, you know, but I talked to people

380

:

who are intrigued by my seeming

specialty as a MarTech developer, and

381

:

some of them are currently developers.

382

:

And I will say, well, one of the

first things you need to do is get.

383

:

your feet wet again in that thing you

probably ran away from a few years ago or

384

:

five years ago or in school or whatever

that thing you think is elementary you

385

:

can't skip the knowing how web browsers

work part of web development and that's

386

:

another thing i mean doing this stuff

for a long time i was there i wasn't

387

:

good at it i i mean i sucked at it but

i was there and in fact i was scared

388

:

because it was like early javascript it

was like too impenetrable to me when i

389

:

was you doing security stuff people that

were also like web developers out there.

390

:

There were, I know this guy, um, he'll

never hear this, Vince Allen, one of

391

:

the greatest JavaScript developers of

early JavaScript that I've ever met.

392

:

Like, like insane.

393

:

I can't even, the things that he could do,

he was, we worked for a startup together.

394

:

I was the back end database guy or

chief technology guy, whatever it was.

395

:

Short lived startup.

396

:

He was the front end guy.

397

:

He was, he was way overqualified because

the, the app didn't really need the

398

:

skills he had in what we used to call,

you know, DHTML or, you know, dynamic

399

:

interactions and things like that.

400

:

So good.

401

:

And they were, and I was like horrified

at the idea of how far he was from

402

:

me or how far he was even from most

people building webpages at the time.

403

:

I think he went on to work

for Google internally.

404

:

He might've even retired by

now or something like that.

405

:

But yeah, there was that.

406

:

Period where you really had

to know what you're doing to

407

:

push early JS to its limits.

408

:

But knowing there were people who

were masters is really important.

409

:

And then you had the whole

period of frameworks and people

410

:

not, that's another thing.

411

:

People not really learning JS

happened 10, 15 years ago, they

412

:

would learn jQuery or Whatever.

413

:

And, uh, so because they never really

learned it and then you plop them into

414

:

something where it's like, well, Marketo

forms do somewhat unfortunately use jQuery

415

:

under the hood, but you're not going

to get away with just understanding the

416

:

core idea of like, you know, building

a jQuery and bootstrap powered site.

417

:

Like you have to get real jazz.

418

:

Like I said, again, The idea

of, you know, uh, those race

419

:

conditions, the timing issues,

weird cross origin security things.

420

:

You just knowing jQuery doesn't

help you figure them out.

421

:

And that's a lot of what I do because

it's this form, you know, we're not

422

:

getting leads because this form doesn't

submit, it doesn't matter how good

423

:

you are at jQuery, it's unlikely that

you're going to be troubleshooting

424

:

that at a code level, you know, you're

troubleshooting at a network level and

425

:

getting deep into browser internals.

426

:

Justin Norris: Reflecting back

on your comment about being

427

:

the person who will take...

428

:

Three hours to answer some

obscure question on the one level.

429

:

You're like, you have

a good billable rate.

430

:

Why would you invest that amount of

time, which represents money into

431

:

something obscure and maybe who else

will ever have that exact question.

432

:

And yet to me, it's because you

have figured out those things

433

:

that no one else has figured out.

434

:

That you're my phone, a friend, or

even to this day, I shared in some

435

:

community that we're both in, I was

having some trouble with an accented

436

:

character, not rendering properly.

437

:

And I'm like, I just, something in my

brain, I had my Sanford sense go off.

438

:

And I'm like, I think

he's talked about this.

439

:

I do my little site search of your site.

440

:

And there it is, was the answer.

441

:

And so it's like, it doesn't

matter until it matters, but

442

:

then it really does matter.

443

:

And if no one

444

:

Sanford Whiteman: Mm hmm.

445

:

Justin Norris: figured that out

before, then like, you're screwed.

446

:

You have like

447

:

Sanford Whiteman: I mean, yeah, it's like.

448

:

I will go maybe forever, but sometimes

for years before using one of my

449

:

discoveries for billable work.

450

:

Like you said, in the meantime, I can

help other people out like you with it.

451

:

What I don't want to be, I don't want

to be in the situation where there's

452

:

an emergency and I need to spend those

three hours on my own client's work.

453

:

So it's almost like a, and perhaps unwise

for my time management, but I'm like,

454

:

I'd rather spend three hours now on the

chance that I have something to have

455

:

in my pocket for two years from now or.

456

:

When a client has that question, okay,

I can only bill for an hour because

457

:

I'm not going to lie and say it took

me the three hours then, but at least

458

:

I don't have the panic that I would

have when I know that my client's

459

:

boss is breathing down their neck.

460

:

I know that the form doesn't work.

461

:

I know about the consequences.

462

:

It may be unwise.

463

:

Frankly, I'm not a very practical,

I'm not a practical person.

464

:

I mean, if you only knew, I

mean, I, the amount of time that

465

:

I spent like for things that.

466

:

Have no meaning at all.

467

:

But yeah, it's, if I have any defense

at all, which is rare, if I have any

468

:

defense, it's that someday someone will

ask me to do it, you know, and they are

469

:

my client and I get to give them the

immediate answer, what I really want.

470

:

I would love to be in

an advisory position.

471

:

Knowing this stuff is key

to product development.

472

:

But this is a frustrating thing

about the role that a MarTech

473

:

developer plays for a client.

474

:

I'm taking the load off the

internal product team, right?

475

:

Like sometimes if someone was contacting

me yesterday, it's a flourishing, you

476

:

know, well funded startup, whatever.

477

:

The head of product said we

can't ask their web devs to do

478

:

anything for marketing anymore.

479

:

So I'm going to ask you because

I know you know how to do it.

480

:

I was like, great.

481

:

But I also had that kind of sinking

feeling of like, okay, now I'll never

482

:

that's yet another isolation point.

483

:

You know, I will never know what the

product team is doing and whether the

484

:

product team could make use of deep

knowledge that is relevant to any web app.

485

:

And I feel like I do have a lot of

that knowledge, whether it be encoding

486

:

knowledge or, you know, that deep browser

timing knowledge and so on, that a lot

487

:

of commercial web apps suffer from not

having a person like that on the team.

488

:

But it's just, that's the way it is.

489

:

Like I'm, I'm, I'm there to take the

load off that team, not to Confer

490

:

with that team about issues of

the same type that their main site

491

:

might have, but it is what it is.

492

:

I guess that's that's another.

493

:

Sometimes it's like, you know, doing a

marketing thing on the never mind the app

494

:

part of it, but, you know, the corporate

site, the brochure site, if you will.

495

:

I have to kind of hold my tongue a lot.

496

:

I'm like, you know, it probably

would be good if you didn't

497

:

have a race condition for your.

498

:

Other stuff, or you didn't have, you

know, it would be good if you weren't

499

:

loading a hundred different remote

services on this page, but that's not my

500

:

business, my business is make sure the

embedded Marketo form works and that's

501

:

Justin Norris: tough, the tough part of

any consulting, you know, where you got

502

:

me like, I'm just, don't look over there.

503

:

Don't look

504

:

Sanford Whiteman: Right.

505

:

Exactly.

506

:

Yeah, it's the don't.

507

:

Yeah,

508

:

Justin Norris: ahead.

509

:

Sanford Whiteman: Well, well, you know,

you've been in this situation, you

510

:

know, when you're, you're building out

a campaign and you look at the email

511

:

content and you go, Oh God, that's so

ugly, but you can't say it because it's

512

:

not, or even like a new logo, you're

like, Oh, say nothing, say nothing,

513

:

Justin Norris: So, so this, this touches

on a question that I wanted to ask and,

514

:

and to some degree it was a motivator for

me to leave consulting, at least leave

515

:

it for now and go back in house, which

was in the role of being a technologist

516

:

that supports marketers, you're, you're

always coming in downstream of strategy.

517

:

Like there's always a strategy

and a goal and a campaign goal.

518

:

And sometimes, I guess I, I believe in

marketing, and you mentioned earlier, you

519

:

believe in, in marketing and its value,

it's not that we're saying marketing

520

:

is not good, but there is a lot of bad

marketing, or there's campaign ideas

521

:

that don't work, or there's things that

don't matter, and so you can architect

522

:

very elegant technical solutions, but

you're building, uh, you're building a

523

:

cannon, but it's being aimed at the wrong

524

:

Sanford Whiteman: yeah,

525

:

Justin Norris: it's never

gonna hit, and how do you feel

526

:

about that, just in general?

527

:

I

528

:

Sanford Whiteman: I mean, you

try to sometimes nudge people.

529

:

It depends on the

relationship with the client.

530

:

You know, sometimes you can say,

well, I've worked in this area

531

:

or I've actually done marketing

myself and kind of bring that in.

532

:

But yeah, it's.

533

:

It's painful because it's, it's not

what you're being retained to do.

534

:

And here's another thing the outside world

doesn't understand the difference between

535

:

marketers and marketing operations people.

536

:

I mean, you could dab 10, 000 people

off the street and one of them

537

:

might understand the difference just

based on their experience, right?

538

:

So.

539

:

But I think, again, it's almost a curse

of being an ops person with a good

540

:

design eye or at least, you know, non

zero design eye or campaign planning

541

:

or just audience understanding persona.

542

:

But I, I feel we're, we're

just going to have to shut up.

543

:

You know, I can, I can think of

countless times where I've talked

544

:

to a client and I'm like, well,

I don't know the industry well.

545

:

Or as well as they do, but I'm still

like this offer for these people.

546

:

This is non identifiable information

offering like a 10 Amazon gift card.

547

:

It's one example, just one of a million

examples, but it's another idea where I'm

548

:

like thinking about the persona purse.

549

:

Why is it that who needs the 10 car?

550

:

I know your B2B every one of your decision

makers is make mid, you know, six figures.

551

:

What?

552

:

Maybe they know better than I do, but

there's, but that would just be, again,

553

:

that was a very superficial example, but

it's the kind of thing where I have to

554

:

go, Oh, so how, Oh, today we're working on

the 10 Amazon gift card for CEOs campaign.

555

:

All right.

556

:

Justin Norris: the thing that's going

557

:

Sanford Whiteman: not

going to say anything.

558

:

Right.

559

:

Exactly.

560

:

It's like review.

561

:

It's also like review our, it's

the 10 Amazon gift card for CEOs.

562

:

If they provide a review on, you

know, one of the, I'm like, okay.

563

:

Let's figure out how we can

get the campaign timing to work

564

:

because I can't do anything.

565

:

And, but what, what do

you think about that?

566

:

I mean, like, is there any room for,

for expansion into that world or

567

:

are we just doomed to shut up and.

568

:

Justin Norris: I mean, I guess for me

that was a motivation to like return to

569

:

an in house role where ostensibly you

have some input on strategy and you're

570

:

having that discussion about what should

we do versus just how do we do it?

571

:

Like should we build this bridge over

there versus how do we cross this span?

572

:

Sanford Whiteman: right, right,

573

:

Justin Norris: it's important to have

both and yeah, this is one of the

574

:

things I wrestle with because You know,

there's sometimes there's people that

575

:

are like, we need to get over there.

576

:

And then there's the people

that say, let's do it.

577

:

And there's a division of labor there.

578

:

And it works for a reason, because if

you have, you know, the person that's

579

:

hitting rocks with a pick hammer

every five seconds being like, wait

580

:

a second, is this the right thing?

581

:

Like

582

:

Sanford Whiteman: right.

583

:

Justin Norris: So there's roles and

responsibilities for some reason.

584

:

At the same time, maybe I just

felt I needed to go outside

585

:

that for a bit, which results in

being less technically oriented.

586

:

Like I'm more distant from

some of the technical things.

587

:

Then I was not irreparably so,

but yeah, it's, it's, it's tough

588

:

Sanford Whiteman: Yeah.

589

:

I mean, often it's also like you

can feel the political, even if you

590

:

could say something you don't want

to, because you want, you don't want

591

:

to be blamed for making a suggestion

when it's not your role too.

592

:

Like sometimes I found that.

593

:

Uh, I have found that a client might be

genuinely receptive to your suggestion

594

:

and even skeptical in their own right

about things, you know, that have been

595

:

handed down to them from marketing.

596

:

But the moment you enter into that

relationship with them, you, you might

597

:

get the feeling that they're going to

say, well, Sandy agreed about this and

598

:

all of a sudden you're implicated in

making the suggestion that everybody

599

:

else from a strategy standpoint

rejected or something like that.

600

:

Justin Norris: To me the best form of

marketing is demonstrating expertise

601

:

in public and building trust.

602

:

If I need a MarTech developer, you're

the only person I'm going to reach

603

:

out to because you care a hundred

times more than any other person.

604

:

And you demonstrate it on a weekly, if not

daily basis, like more than anybody else.

605

:

Sanford Whiteman: I had a, an interesting

situation like this with a client recently

606

:

that is not directly related to marketing.

607

:

I was working with a party who

was working for their client.

608

:

You might have seen me post about

this on Mops Pro's Slack channel.

609

:

I did a little vent, a tiny one line

vent about it because I thought it was

610

:

so silly, but yet disturbing and silly.

611

:

The client had said that the Marketo

lead ID is PII and that It can't be

612

:

included in, it wasn't actually in URLs.

613

:

It was actually for internal

purposes that they couldn't key

614

:

things off the Marketo lead ID.

615

:

I found myself in one of those situations

where I couldn't hold my tongue because

616

:

it wasn't really a marketing question.

617

:

It was a, you know,

privacy strategy thing.

618

:

I'm not a specialist in privacy, but

something when it just sounds completely

619

:

preposterous and wrong, and you just,

you can't say nothing, you know?

620

:

So, and I still haven't

gotten a satisfactory answer.

621

:

I, I said they, they needed some.

622

:

Massive, yet insufficient workaround

because their client had told them

623

:

that the Marquette Elite ID is PII.

624

:

And I was like, okay, we can do this.

625

:

Maybe do this.

626

:

It's never going to be as,

it doesn't make any sense.

627

:

And I kept saying like, again, to repeat,

I really have never heard anybody say

628

:

anything like that in my entire career.

629

:

That

630

:

Justin Norris: Cause it's a system

631

:

Sanford Whiteman: Generate.

632

:

Right.

633

:

A randomly generated, well, they said,

okay, to be clear, they said the same

634

:

thing about the Marketo unique code.

635

:

So it's not, you might be able to, I

don't, wouldn't buy it, but you might

636

:

be able to make the argument that the

auto incrementing numeric lead ID has

637

:

some sort of predictability to it.

638

:

In fact, one of my clients once made

a good point about that, that you

639

:

can predict the sequence of lead IDs.

640

:

So whatever, it's still the same as

PII, but it's like a security thing

641

:

more than a privacy thing by any means.

642

:

The client applied the same fractured

logic, in my opinion, to the

643

:

Marketo unique code, a guaranteed

unique alphanumeric sequence that

644

:

has no meaning outside of Marketo.

645

:

And it's just hard to, once you hear

things like that, actually expand it

646

:

to a whole realm of like what other

poor decisions is somebody making?

647

:

So it actually, maybe it shut me

up more because I was like, okay.

648

:

If they're that wrong, I'm not

talking about other things, but I

649

:

thought it was a good example of like

something crossing the line where

650

:

I couldn't be quiet about it, even

though it's very far from the direct

651

:

responsibility that I had in the project.

652

:

I think that you, in your current

role, I would, I would imagine you're

653

:

in the room where people are saying

things like that, and Justin can.

654

:

Raise his hand and say, that sounds

nonsensical you guys, but you can't

655

:

do that as a consultant because

it's like you kind of just, okay.

656

:

Justin Norris: there's a flip side

to that because sometimes I noticed

657

:

this, not a commentary on my current

job, but when I first left my in house

658

:

role to go to consulting, there is a

prestige on the flip side that comes

659

:

with like, well, they're a consultant.

660

:

So there's this like halo, you

know, that surrounds the consultant

661

:

Sanford Whiteman: Uh, yeah, true, true.

662

:

Justin Norris: That you don't

have when you're in house.

663

:

But I think, at a company of my stage,

the questions are a lot more go to market

664

:

strategy than the thorniness of PI.

665

:

Not that we don't care

about PII, but it's...

666

:

I think it's maybe only in a

massive enterprise sometimes

667

:

where you have the luxury.

668

:

Sometimes it can become

669

:

Sanford Whiteman: things like that.

670

:

Yeah,

671

:

Justin Norris: a naval gazing

exercise for lack of a better word.

672

:

I want to ask you, I mean we both

have a long history with Marketo

673

:

and have based our careers in

various ways on that platform.

674

:

What's baked into the question is

that there's obviously a lot to

675

:

like about Marketo, we appreciate

it and all of that, but what's

676

:

your perspective on the platform?

677

:

Are the glory days behind it?

678

:

We talk about the things that are the

best about it and it's still like...

679

:

The smart campaign, the Salesforce

integration, like things that are really

680

:

from the early days when it had a lot of

vigor and energy as a product company.

681

:

And it's hard to think of innovations

that come close to that in recent time.

682

:

So I don't want to go down too far the

path of just venting against Marketo,

683

:

Sanford Whiteman: No, no,

684

:

Justin Norris: what's your product

outlook and what do you think they need

685

:

to do to remain relevant as a platform?

686

:

Sanford Whiteman: we were speaking

about these add ons like integration

687

:

with Adobe Connect webinars and the

seeming struggle of getting live chat

688

:

as a replacement for Drift and so on.

689

:

If I were in charge, I would shut

down anything that seems like

690

:

it's trying to be a look alike

for a pale imitation of something.

691

:

I think in both of those cases is a

customer retention concept, right?

692

:

Like they're not charging more

for those, which is both to their

693

:

credit and also makes it seem.

694

:

Even more bizarre, because the idea

is we're going to use this instead

695

:

of drift, and so we're going to save

all the money on drift, except it's

696

:

also nowhere near as good as drift,

so that's never going to happen.

697

:

Like, it

698

:

doesn't

699

:

Justin Norris: do live chat.

700

:

It doesn't do...

701

:

Yeah, I agree.

702

:

So actually, let's rebuild Marketo.

703

:

Like, we'll build their

roadmap on paper for free.

704

:

They can just take it and use

705

:

Sanford Whiteman: right, right.

706

:

Justin Norris: So we kill the

chat, we kill the Adobe Connect

707

:

thing, whatever that is, so

708

:

Sanford Whiteman: What, you

know, what would we build, what

709

:

would we build in its stead?

710

:

What is the thing that isn't trying

to keep up with the Joneses in a

711

:

feeble way, but is actually, you know,

712

:

long desired?

713

:

Justin Norris: that, to me is

like the critical thing, right?

714

:

You know, it's the whole idea of getting

ahead of the market and where it's

715

:

going and understanding the needs, you

know, the Steve Jobs esque quality of

716

:

understanding the thing that people

could want before it even existed.

717

:

I don't know if there's anybody there

that is thinking about the customer

718

:

that hard and it comes back to the

question of building technical solutions

719

:

for things that don't actually...

720

:

Work that well.

721

:

A lot of it's like, does anyone care

about what is effective marketing and

722

:

what actually is revenue producing and

then trying to build solutions that

723

:

make that easier and more effective.

724

:

I think there are tools out there that

really work well, like the meeting

725

:

booking link things where you can.

726

:

Book a meeting on a website.

727

:

To me, that's a real customer convenience.

728

:

It's beneficial.

729

:

It takes less time.

730

:

It's removing friction

from the sales process.

731

:

It's a tool that solves a

legitimate customer problem.

732

:

And I wonder, is there any new

features coming out that actually

733

:

solve new customer problems?

734

:

There's all the, Oh, this

thing is inefficient.

735

:

Can we like clone flow steps or

save smart campaign templates?

736

:

You know, just little, little things

that all the convenience features

737

:

that people ask for that they never

make it to production because there's

738

:

like, well, what incremental revenue

is associated with that feature

739

:

aside from customer happiness.

740

:

There's like all that category of stuff.

741

:

And then there's like the big innovations,

which are really hard to think of.

742

:

And it seems that it's usually smaller

companies with dynamic founders who

743

:

care about the problem and obsess

about it, come up with those solutions.

744

:

Like, I don't necessarily know what

they are, but I haven't seen those

745

:

things coming out from any of the

big players in marketing automation.

746

:

This isn't just a Marketo thing.

747

:

Sanford Whiteman: Here's a

thing that we know from AWS.

748

:

I mean, Amazon's obviously been around

for a long time, but that idea that at

749

:

Amazon internally they use their own

APIs so they can't go around the APIs and

750

:

avoid discovering the bugs in the APIs.

751

:

I think that You know, Marketo and now

Adobe has never made that commitment.

752

:

They've never decided we're going to have

a flexible full featured API that we are

753

:

going to dog food for our own interface,

which would of course allow them to

754

:

allow you to create, Oh, here's a thing.

755

:

Okay.

756

:

I'm going to interrupt myself

because I do think this is something

757

:

that is, it is absurd that they

have not put this in the pipeline.

758

:

And it's related directly to the

idea of having a public Marketo

759

:

UI API or an SDK for building

visual add on to the Marketo UI.

760

:

Justin Norris: right.

761

:

Sanford Whiteman: That is, I've gone

from defending it and saying it's

762

:

not their business, they never wanted

to do that, to kind of thinking

763

:

it's shocking after all these years.

764

:

And to get back to our whole discussion of

front end developers, that an interested

765

:

front end developer, one who is motivated,

Can't even do it if they want to.

766

:

They cannot say I am going to spend

time customizing my organization's UI.

767

:

You can do it in Salesforce.

768

:

You can't do it in

769

:

Marketo.

770

:

That's a silly

771

:

Justin Norris: step, but really

like, like, like skinning and, uh,

772

:

and even like layout and stuff like that.

773

:

Sanford Whiteman: Well, I, I, yeah,

I think SSFS is actually a way of

774

:

avoiding these better technologies.

775

:

I mean, like SSFS is creating an API

that others can consume, but it's low

776

:

effort for Marketo themselves actually

rewriting and we know how hard this

777

:

Justin Norris: Like a visual

force page, the equivalent of a

778

:

Sanford Whiteman: Exactly.

779

:

Exactly.

780

:

With a move from classic to, you

know, the aura lightning components

781

:

in salesforce was a major, major,

major undertaking for salesforce.

782

:

I cannot imagine how many,

you know, hours pure expenses

783

:

they had to get to that point.

784

:

And it's not the greatest

thing I've ever seen.

785

:

But, but it does seem like adobe

is uninterested in that level of

786

:

customizability and extensibility.

787

:

It's not just custom.

788

:

It's extensibility.

789

:

It's the ability for

someone to say, okay, Okay.

790

:

I want to create a new reporting

all with my code, but consuming

791

:

the same back end data.

792

:

And I want it to appear as

a tab within the Marketo UI,

793

:

Justin Norris: It makes Marketo

more of a platform, right?

794

:

And less of like, well, that's just a

point tool of execution and it generates

795

:

some first party data, which I'm going to

take out and do that work somewhere else.

796

:

Sanford Whiteman: but I just

want to, I want to caution.

797

:

So I don't sound crazy as a

front end person, and I know the

798

:

reputation of projects like that.

799

:

Rewriting from scratch, which is what

this would entail, even if it's only

800

:

the front end, it's one of the things

that the greatest, you know, The

801

:

legendary programmers dedicate chapters

of their horror, you know, books of

802

:

horror to decisions, rewrite things

from scratch, which is fascinating

803

:

because it means that this notion

rewrite from scratch is like dismissed

804

:

by the greatest developers in the world.

805

:

It implies that a given product and

its current, like general iteration

806

:

cannot be rewritten in place.

807

:

It has to be rewritten

and, or there has to be.

808

:

A substantive evolution to

what constitutes a new app in

809

:

order to make such an effort.

810

:

And to not get slapped by the reality

and the disasters that always follow that

811

:

seemingly always follow such a decision.

812

:

So I, I get it, but it's, it's

such a funny thing to know that

813

:

it's like, if you say rewrite from

scratch, you'll get made fun of by.

814

:

You know, experienced gray

beard devs, as they say,

815

:

Justin Norris: So you have to

rebuild the car like one piece at

816

:

a time until you get a new car.

817

:

You can't

818

:

Sanford Whiteman: you can, yeah,

819

:

Justin Norris: car.

820

:

Sanford Whiteman: modular or you can say

it's a new app, but the idea is you don't

821

:

try to rewrite the current functionality

in another language or with another

822

:

framework or in another, another scenario.

823

:

And you might say dot,

you know, dot next can be.

824

:

And that's, you know, Mark

and I tried that with Sky and

825

:

Justin Norris: so I was just

826

:

Sanford Whiteman: the

whole idea is like rip it.

827

:

Right, right.

828

:

But they never really ripped it all out.

829

:

So,

830

:

um, they weren't, they were

831

:

Justin Norris: a lot of people wonder

why Sky failed, like, why did they

832

:

back away and leave us with this,

like, Franken UI that seems like it

833

:

will never be fully transitioned,

it's the weirdest thing, like, maybe

834

:

Sanford Whiteman: it is really weird.

835

:

Justin Norris: it had issues, but there

was a significant investment, there

836

:

was, like, it was going somewhere, it

seemed, and I don't think the reactions

837

:

were terrible to it in the community,

And yet they backed away somehow, maybe

838

:

for adoption reasons, and they're like,

let's do this in between, and you can

839

:

toggle, but we're gonna update, like,

parts of the UI at a time, and now,

840

:

I've gotten used to it, but if you

back away, it's a horrifying thought

841

:

to have a UI like that, like, for an

842

:

Sanford Whiteman: It was a,

843

:

Justin Norris: I can't think of

another app that looks that way.

844

:

Sanford Whiteman: I think that I

mean, look, I don't have any inside

845

:

information, so that's that, but

they decided to get away from the

846

:

XJS framework they're currently using

and try to use another framework at

847

:

the same time against the same API.

848

:

I think that's part of it.

849

:

They didn't, they didn't want

to like completely rewrite.

850

:

And again, that sounds like a bad thing.

851

:

I'm saying it's bad to like

completely rewrite, but they

852

:

didn't want to do a v next, right?

853

:

They wanted to do the sky on the

current, when I say like, it's like the

854

:

API for the front end, that is to say

the public API that isn't documented,

855

:

but allows for fetching things like,

you know, smart lists and flow steps

856

:

and all the stuff that happens.

857

:

I think they wanted to keep that

same API, but consume it using

858

:

a different front end framework.

859

:

And that falls into the

category of never do that.

860

:

That's, I think the most experienced

developers and product people would

861

:

say, no, you're going to have to,

you're, you're doing Marketo V2.

862

:

That's, you cannot do UI

V2 on top of Marketo V1.

863

:

You have to do a full evolution.

864

:

I don't

865

:

know why they didn't do that.

866

:

Like I've seen, I've seen that succeed.

867

:

I mean, MailerLite, small consumer

player, mostly, they've got V2, and

868

:

there are substantive differences,

but they obviously have a team

869

:

that I think it's, you know.

870

:

Privately held small team of

probably more loyal people and

871

:

less distraction and whatever.

872

:

And yeah, but that was, that

was, uh, that's super weird.

873

:

I feel like if I had to put

something on the roadmap, it

874

:

would be the customizable UI.

875

:

It would be the ability to say you

can now build directly into the UI.

876

:

That's the kind of thing where

you could get your Salesforce

877

:

lightning developer interested.

878

:

That's where you get a front end person

to go, Hey, this is actually cool.

879

:

I can plug into the menus.

880

:

I can, you know, bring this data in

and visualize it in different ways.

881

:

Another thing I would do if I

were in charge of product is start

882

:

looking at whole database metrics.

883

:

And you and I worked on this

together for when we did that time

884

:

with the time series thing, right?

885

:

That the fact that we did do

that together for the listeners.

886

:

This was a way of pulling a database

into an external time series

887

:

database through which we could do.

888

:

Things that seem simple, like sorting

the people in the database by their

889

:

level of, let's call it, web engagement.

890

:

So we could say that they

were in the four quartiles.

891

:

You could say this person is in the top

quartile of web engagement, or you could

892

:

say this person is in, you know, the third

quartile of email clicks, or whatever.

893

:

It's not that Difficult to do in a

modern database, even one that is not

894

:

specifically designed for time series

data, the one that that you and I use,

895

:

Justin, is obviously a time series

database, but you can now do it in

896

:

Postgres and in, I think, to some

degree in MariaDB, which It's like

897

:

MySQL, which Marketo uses, but for

whatever reason, they never thought,

898

:

hey, maybe somebody wants to see the

database, see the relationships between

899

:

people within the Marketo database.

900

:

To me, it's like the lowest hanging fruit.

901

:

If I had all the Marketo data in a

modern Postgres 11 plus or whatever

902

:

version it would have to be, if I had

all that stuff and I wanted to write a

903

:

cool query that would blow somebody's

mind for a new evolution of Marketo.

904

:

I mean, that's what I would do.

905

:

I would do split by quintile, you know,

basically letting the leads in your

906

:

database interrogate each other as to

which ones are higher or lower value.

907

:

There's some degree I think that they

do that with like the heat ratings

908

:

and, and relative scoring.

909

:

Yeah.

910

:

But,

911

:

Justin Norris: and

912

:

Sanford Whiteman: but it's not, but

it's, but it's like, it's opaque.

913

:

It doesn't need to be opaque.

914

:

You should.

915

:

I mean, that's kind of needlessly,

it doesn't need to be magic,

916

:

kind of fake AI, right?

917

:

You should be able to run those queries,

I think directly split the database

918

:

by any number of time based, you know,

contributions to the value of a person.

919

:

But I don't know what stops them

from doing that because here I am

920

:

quite sure their backend supports it.

921

:

They just never kind of think, well, like

we don't want to give another dashboard.

922

:

We'll say you can pull

the data out yourself.

923

:

And they do actually, you know.

924

:

You know about the

activity stream feature?

925

:

I haven't used it.

926

:

Justin Norris: no, what is it?

927

:

Sanford Whiteman: The activity stream

feature, which is supposedly free,

928

:

lets you syndicate the rows of the

activity log in near real time to

929

:

a web service of your choosing.

930

:

Justin Norris: So this

is like an API feature?

931

:

Sanford Whiteman: Yeah.

932

:

But the cool thing is it's outbounds.

933

:

It's like a streaming API

934

:

where you give it an

endpoint and it will send.

935

:

The activity log lines to that end point.

936

:

I mean, it's asynchronous surely, but

it's like roughly as they come in, they

937

:

get relayed to an external endpoint.

938

:

This is the kind of thing that

939

:

Justin Norris: got the

notes up here, the developer

940

:

Sanford Whiteman: you can kind

of see as a database architect.

941

:

I know that those features like

change data capture and row based

942

:

replication and that kind of thing

are all part of modern databases.

943

:

So it seems like it was something that

makes sense for them to have introduced.

944

:

The thing is it still requires

that you have in house people

945

:

and you build a service that can

receive, you know, it's, it's.

946

:

It's the furthest thing you

can get from a new dashboard.

947

:

It gives you the building blocks to build

real time dashboards in outside, you know,

948

:

in third party utilities by streaming

the data to them, but it contains

949

:

nothing else except the raw log lines.

950

:

So it's

951

:

like,

952

:

Justin Norris: is a good feature.

953

:

I'm just looking through the notes.

954

:

This is a good feature.

955

:

I clearly haven't been following closely

956

:

Sanford Whiteman: Oh no, I

mean, look, I haven't had it

957

:

enabled in any of my instances.

958

:

So I'm

959

:

Justin Norris: this just come

960

:

Sanford Whiteman: know,

um, six months ago, maybe.

961

:

Justin Norris: Okay, well I

962

:

Sanford Whiteman: Yeah, I'm

going to, I'm going to try it.

963

:

Yeah.

964

:

So, I mean, that,

965

:

that, that,

966

:

Justin Norris: really, because like it

says, it solves a lot of problems where

967

:

you're constantly auditing and, you know,

or polling rather for changes and the

968

:

backlog that that creates and, you know,

the challenge with real time triggering.

969

:

So my top two, for me at least.

970

:

There's always the limitations

that you encounter in the

971

:

raw workflow capabilities.

972

:

Like, oh, you can't quite

manipulate data that way.

973

:

If I think of Workato, which

to me is like Marketo, but

974

:

split apart into 12 dimensions.

975

:

'cause you can work across multiple

systems and also the data pill capability

976

:

and recipes, the ability to hold in data

from a previous step, interrogate it, use

977

:

formulas to manipulate it in various ways.

978

:

Kinda like what mu munchin or munch.

979

:

Whatever

980

:

Sanford Whiteman: Yeah, yeah,

981

:

Justin Norris: munches, that very short

lived and obscure thing where you could,

982

:

I guess kind of what Flowboost tries to

do, but inside of Marketo, to me again,

983

:

is like a very obvious thing to do.

984

:

The ability to better manage race

conditions, I think you know, like

985

:

a lot of the gray hair you get in

Marketo, if you care about doing

986

:

your job well, is how do I make sure

that I'm not tripping over myself?

987

:

There's a lot of different ways to do

that and a lot of different hacks and

988

:

things but the kind of feeling that you

can never fully have the system under

989

:

control and maybe this is an impossibility

you tell me as a developer to like that

990

:

you always have this multiple layers

of the database and things committing

991

:

at various times so you can never fully

have perfect control but then the ability

992

:

to handle that in some more elegant way

than like doing your best but always

993

:

feeling that there's an outside chance

that something might trip over itself.

994

:

Sanford Whiteman: hmm.

995

:

I mean, I do think it's kind of an

unsolvable, even though it's it's

996

:

not like a distributed system where

distributed systems have even, you

997

:

know, increased impossibility of

getting of having a predictable

998

:

settled state at any one point in time.

999

:

But I think.

:

00:47:38,486 --> 00:47:42,626

The Marketo design philosophy seems

to be to dumb things down and not

:

00:47:42,636 --> 00:47:46,546

expose the reality of what's happening

under the hood to the end user.

:

00:47:46,576 --> 00:47:50,466

And I think that just knowing

that, I mean, maybe this is crazy.

:

00:47:50,466 --> 00:47:53,486

I don't know if any commercial

platform has ever done this, but.

:

00:47:54,021 --> 00:47:56,991

Having visibility in database

changes have yet to be committed.

:

00:47:57,081 --> 00:48:02,341

You know, imagine a super awesome spy

view where you could see as, leaving aside

:

00:48:02,341 --> 00:48:07,331

the overhead of providing this view that,

but like where you could see in real time

:

00:48:07,331 --> 00:48:09,061

as things are written to the database.

:

00:48:09,486 --> 00:48:12,266

Or or or databases as databases.

:

00:48:12,266 --> 00:48:14,116

Don't pretend they're

not databases, right?

:

00:48:14,186 --> 00:48:16,846

And I caution people about like

I say, don't call the custom

:

00:48:16,846 --> 00:48:18,286

objects, a custom object table.

:

00:48:18,296 --> 00:48:19,466

So I'm kind of a hypocrite here.

:

00:48:19,466 --> 00:48:20,996

Like I'm saying, like,

don't call it a table.

:

00:48:20,996 --> 00:48:22,396

We don't know how it's implemented.

:

00:48:22,396 --> 00:48:23,366

It could be four tables.

:

00:48:23,366 --> 00:48:26,056

It could be, you know, joined together

relationally, whatever it could be.

:

00:48:26,206 --> 00:48:31,191

But Surely there is a happy medium between

this dumb, down, I don't really know

:

00:48:31,191 --> 00:48:35,701

what's happening under the hood, I just

have to guess, based on the fact that

:

00:48:35,701 --> 00:48:39,531

it's got a database with, like, there's

gonna be a row level lock, right, like,

:

00:48:39,541 --> 00:48:43,011

that's the kind of thing you and I know

about for, for years now, it's like,

:

00:48:43,441 --> 00:48:50,021

it's a SQL database, so we know Only

one row is being updated at one time.

:

00:48:50,131 --> 00:48:54,931

We know that if two low steps try

to update the same row, one of them

:

00:48:54,931 --> 00:48:58,491

has to go first and the other one

has to go second, even if they're

:

00:48:58,511 --> 00:49:02,101

updating the same, or especially

if they're updating the same field.

:

00:49:03,066 --> 00:49:05,316

cannot specify the order of those updates.

:

00:49:05,406 --> 00:49:07,476

I don't know what kind of visualization

you could have of that, but it

:

00:49:07,476 --> 00:49:10,996

just seems like there's been a

concerted effort to like conceal

:

00:49:11,056 --> 00:49:13,216

these essential database realities.

:

00:49:13,466 --> 00:49:16,926

I think you put it really well when

you said you never really feel like you

:

00:49:16,936 --> 00:49:19,706

know what's going on with the market

and it's like, you don't, you don't.

:

00:49:19,956 --> 00:49:23,886

I feel, and less of a Salesforce

person, but I feel more secure about

:

00:49:23,886 --> 00:49:28,736

what's happening, the state of a

Salesforce org at point N in time.

:

00:49:29,011 --> 00:49:30,451

Than I do about Marketo.

:

00:49:30,481 --> 00:49:31,861

Like, what's really happening?

:

00:49:31,861 --> 00:49:33,301

What's, what's in a wait step?

:

00:49:33,301 --> 00:49:34,351

What's waiting to commit?

:

00:49:34,351 --> 00:49:35,611

What's about to be overwritten?

:

00:49:35,701 --> 00:49:36,031

Right?

:

00:49:36,031 --> 00:49:38,731

Like, what, what looks the

way, certain way right now?

:

00:49:38,736 --> 00:49:40,201

But when I refresh, it's

gonna look different.

:

00:49:40,201 --> 00:49:42,241

There's, there's, there's,

there is that uncertainty.

:

00:49:42,381 --> 00:49:43,811

Justin Norris: no debug log, right?

:

00:49:43,811 --> 00:49:47,481

So you give marketers these developer

tools, but it's not like a MailChimp

:

00:49:47,481 --> 00:49:50,431

where it's just like, here's a glossy

interface and you have these little

:

00:49:50,431 --> 00:49:53,541

boxes that you can play with, but you're

kind of, everything else is locked down.

:

00:49:53,861 --> 00:49:55,501

It does actually give you a tool set.

:

00:49:55,501 --> 00:49:56,451

And I think that's why.

:

00:49:56,781 --> 00:49:59,671

I love Marketo, why lots of people

love Marketo, it's like, whoa,

:

00:49:59,671 --> 00:50:02,661

like I can, I'm not a developer,

but I can do developer esque

:

00:50:02,896 --> 00:50:03,186

Sanford Whiteman: Yeah.

:

00:50:03,256 --> 00:50:03,546

Yeah.

:

00:50:03,661 --> 00:50:06,151

Justin Norris: this platform, but then

you're like, whoa, actually developer

:

00:50:06,151 --> 00:50:09,121

esque things are hard, because I have

enough rope to hang myself with, and I

:

00:50:09,121 --> 00:50:11,851

can screw things up, and it's like, okay,

I'm going to get better, and then I'm

:

00:50:11,901 --> 00:50:15,601

going to learn from folks like you who

really understand developer principles, so

:

00:50:15,601 --> 00:50:20,451

things like dry, you know, not repeating

yourself, centralizing things, or word of

:

00:50:20,451 --> 00:50:23,771

operations concerns, or You Whatever, all

those kinds of things that I've picked

:

00:50:23,771 --> 00:50:27,051

up, but then at the end of the day, it,

it only goes so far so that you don't have

:

00:50:27,051 --> 00:50:31,401

like the full control that maybe I presume

one does in the code application where

:

00:50:31,401 --> 00:50:32,631

you're really writing things yourself.

:

00:50:33,061 --> 00:50:33,871

And so I guess that's it.

:

00:50:33,891 --> 00:50:37,671

It's like, maybe it's unrealistic to

think that you really could ever do that.

:

00:50:37,671 --> 00:50:38,831

Like have your cake and eat it too.

:

00:50:38,841 --> 00:50:42,811

It's like, I can go so far, but

no further within this interface.

:

00:50:42,941 --> 00:50:43,841

Sanford Whiteman: I think that.

:

00:50:44,186 --> 00:50:48,276

The design principle of make it

seem like it's a technical person's

:

00:50:48,306 --> 00:50:49,396

platform, which it is to do it.

:

00:50:49,396 --> 00:50:51,366

Well, you have to have technical gifts.

:

00:50:51,406 --> 00:50:52,006

There's no doubt.

:

00:50:52,056 --> 00:50:55,136

I think that a person who can build

a great instance like you, you

:

00:50:55,136 --> 00:50:58,306

know, other others of your caliber

could also be great developers.

:

00:50:58,316 --> 00:50:58,636

If you.

:

00:50:58,951 --> 00:51:00,451

Want it to be bored out of your skulls.

:

00:51:00,656 --> 00:51:01,656

Justin Norris: In a separate timeline,

:

00:51:01,881 --> 00:51:02,161

Sanford Whiteman: Yeah.

:

00:51:02,301 --> 00:51:05,851

And I think there's a sense that it

doesn't satisfy anybody completely.

:

00:51:05,901 --> 00:51:09,351

If you're a not very skilled or

don't have those gifts, you're

:

00:51:09,351 --> 00:51:12,991

going to be overwhelmed and

basically have a continuous anxiety.

:

00:51:13,041 --> 00:51:16,261

I mean, it's a brilliant platform

because it kind of sells itself

:

00:51:16,261 --> 00:51:18,781

based on its technical flexibility.

:

00:51:18,811 --> 00:51:22,321

If you don't disclose how much

consulting help you might need after

:

00:51:22,321 --> 00:51:27,851

you buy it, it seems like the solution

to really any B2B marketing needs.

:

00:51:28,156 --> 00:51:31,946

Justin Norris: let's close with the buzzy

question, which is around AI, and of all

:

00:51:31,946 --> 00:51:36,086

the people that I personally have a chance

of interacting with on a semi regular

:

00:51:36,086 --> 00:51:39,376

basis, you're probably the person that

maybe best would understand what actually

:

00:51:39,376 --> 00:51:43,946

might be happening with AI, which has

become just horribly hyped, and everything

:

00:51:43,946 --> 00:51:45,706

has those two letters attached to it.

:

00:51:45,823 --> 00:51:49,093

Marketo's tried to use what was

formerly called machine learning.

:

00:51:49,148 --> 00:51:52,868

A number of times, I think, mostly

it seemed unsuccessfully, like

:

00:51:52,908 --> 00:51:56,588

predict your audience or deliver

the right content to the right time.

:

00:51:56,638 --> 00:51:59,688

Maybe those capabilities are more

achievable now in a meaningful

:

00:51:59,688 --> 00:52:02,268

way with the tools we have today,

the technology we have today.

:

00:52:02,268 --> 00:52:05,291

But I guess even beyond Marketo,

just in the broader realm of

:

00:52:05,291 --> 00:52:06,911

MarTech, is this hype to you?

:

00:52:06,921 --> 00:52:11,631

Is this a real thing that is going

to have an impact on our discipline?

:

00:52:11,651 --> 00:52:12,361

What do you think?

:

00:52:12,901 --> 00:52:17,541

Sanford Whiteman: Well, I mean, I'm not

one of these, you know, scoffers who's

:

00:52:17,561 --> 00:52:19,761

like, Oh, it's just one big database.

:

00:52:19,821 --> 00:52:21,891

It's just the most powerful

database ever created.

:

00:52:21,971 --> 00:52:25,481

Speaking of LLMs and you know,

in general, I'm intimidated by

:

00:52:25,491 --> 00:52:26,671

and impressed by a lot of it.

:

00:52:26,671 --> 00:52:30,551

But from a, again, from a marketing

that's much marketing ops standpoint,

:

00:52:30,781 --> 00:52:36,121

my first, the first dread I got wasn't

about what I do for a living or what

:

00:52:36,121 --> 00:52:39,231

you do, but the idea that people

are like, I'm just going to feed.

:

00:52:39,531 --> 00:52:44,681

I'm going to ask at GPT to write my,

uh, you know, taglines to write my

:

00:52:44,701 --> 00:52:49,271

content for my, for my page based on,

and I'm like, and people are obviously

:

00:52:49,271 --> 00:52:54,071

doing that by the thousands, taking a

critical eye on the output of, you know,

:

00:52:54,121 --> 00:52:58,071

chat, GPT and, and, and others from a

marketing copy standpoint, it's like,

:

00:52:58,191 --> 00:53:00,451

I have not seen evidence that it's.

:

00:53:00,876 --> 00:53:05,896

Better, faster, or, I don't know,

impressive than a copywriter

:

00:53:05,906 --> 00:53:08,826

would be, but the idea is to

put copywriters out of business.

:

00:53:08,876 --> 00:53:10,486

It's the same thing that's

happened in many industries.

:

00:53:10,526 --> 00:53:12,826

It's not that the technology is

actually better than the people.

:

00:53:12,906 --> 00:53:17,296

That, I think we need to get past the

fallacy that it's better than people.

:

00:53:17,711 --> 00:53:20,431

But what it could in theory

be is way cheaper than people.

:

00:53:20,451 --> 00:53:21,241

And from a,

:

00:53:21,801 --> 00:53:25,081

you know, business standpoint,

that's, that's all that matters to me.

:

00:53:25,281 --> 00:53:28,971

It's like, don't try to sell me on

AI written documentation being better

:

00:53:29,011 --> 00:53:30,411

than handwritten documentation.

:

00:53:30,781 --> 00:53:35,191

You can sell me on it being faster

and almost right in certain cases.

:

00:53:35,201 --> 00:53:37,911

But what concerns me, even, even

from the standpoint of documentation,

:

00:53:37,911 --> 00:53:41,081

where you kind of do think that more

is better, you know, that's like

:

00:53:41,081 --> 00:53:43,631

with documentation, the less is more.

:

00:53:43,971 --> 00:53:45,381

Concept is completely false.

:

00:53:45,531 --> 00:53:46,491

Absolutely not true.

:

00:53:46,531 --> 00:53:48,501

I am firmly opposed to that.

:

00:53:48,531 --> 00:53:51,071

I'm, you know, that I hate

when things are dumbed down.

:

00:53:51,131 --> 00:53:55,061

The idea of expanding product

documentation via AI is fascinating

:

00:53:55,061 --> 00:53:58,641

to me because it sounds like

super awesome, great net.

:

00:53:58,976 --> 00:54:04,256

When something's wrong in the

docs, the load on people for doing

:

00:54:04,256 --> 00:54:08,656

that thing wrong might actually

on net, you know, be negative.

:

00:54:08,666 --> 00:54:12,216

So it's hard to say like the

Marketo product docs have

:

00:54:12,646 --> 00:54:15,166

errors in them that an AI.

:

00:54:15,426 --> 00:54:19,536

Based doc generator couldn't make so

that that's actually, that's something

:

00:54:19,536 --> 00:54:23,296

where I have to, I have to concede, you

know what, if those docs were written

:

00:54:23,296 --> 00:54:26,706

by chat GPT, it couldn't make certain,

I'm not saying it couldn't make new

:

00:54:26,706 --> 00:54:27,936

errors, but it couldn't make the errors.

:

00:54:27,936 --> 00:54:30,576

Like there's a, I don't know if you've

noticed this in the product docs, like

:

00:54:30,626 --> 00:54:34,456

they're curly, uh, curly quotes in the

code samples that have to be straight

:

00:54:34,456 --> 00:54:36,326

quotes, or it's going to be a fatal error.

:

00:54:36,366 --> 00:54:40,516

They're like errors in the MIME

type structure for the Marketo that

:

00:54:40,516 --> 00:54:44,236

could never happen, but with a, an

AI, you know, documentation library.

:

00:54:44,711 --> 00:54:48,411

But would it make other mistakes and

assumptions that would then become the

:

00:54:48,431 --> 00:54:53,161

stuff of support tickets and confusion

and the need to have it gone over by

:

00:54:53,161 --> 00:54:56,071

someone and, you know, you know how

technical documentation is, there's

:

00:54:56,111 --> 00:54:59,521

maybe one person in the world who

knows whether something is true or not.

:

00:54:59,621 --> 00:55:01,021

From my perspective, like.

:

00:55:01,581 --> 00:55:03,681

You know, Martech development,

like front end stuff.

:

00:55:03,781 --> 00:55:07,191

I would love to know that some of the

stuff that I've written on my blog

:

00:55:07,191 --> 00:55:10,731

and put out there in the world ends

up being digested by chat TPT and

:

00:55:10,731 --> 00:55:13,691

becomes part of the corpus of things

that other people are learning from

:

00:55:13,701 --> 00:55:14,871

because it's open source anyway.

:

00:55:14,871 --> 00:55:19,041

And, you know, I hope that people

are able to use other people's

:

00:55:19,251 --> 00:55:21,431

work and get good code out of it.

:

00:55:21,766 --> 00:55:25,416

That's kind of the advantage of all

the time that I spend diving into the

:

00:55:25,416 --> 00:55:28,756

underpinnings, the internals of Marketo

stuff or just browser stuff in general.

:

00:55:29,076 --> 00:55:34,166

If you publish that and it becomes part

of Chat TPT's learning, I think that's

:

00:55:34,166 --> 00:55:35,936

great because it has more of a payoff.

:

00:55:35,966 --> 00:55:38,626

That sounds to me a little bit too

rosy, but in a way it has more payoff

:

00:55:38,666 --> 00:55:41,266

to spend three hours on something

and know that people will be able

:

00:55:41,266 --> 00:55:44,536

to learn from that in their own code

bases as opposed to having to search

:

00:55:44,536 --> 00:55:46,186

for you and find it on your blog.

:

00:55:46,186 --> 00:55:47,066

That's, that's a positive.

:

00:55:47,466 --> 00:55:47,956

How about you?

:

00:55:48,021 --> 00:55:49,181

I mean, what's what's

:

00:55:49,296 --> 00:55:51,686

Justin Norris: I'm, broadly

aligned with your point of view.

:

00:55:51,716 --> 00:55:54,936

The way I've started to think about

AI a little bit is like, I'm a Star

:

00:55:54,936 --> 00:55:58,376

Trek fan, probably not familiar with

the computer on a Star Trek ship.

:

00:55:58,766 --> 00:56:00,966

You know, where you're like,

computer, create for me a model

:

00:56:00,966 --> 00:56:02,256

of X, Y, Z, and then do this.

:

00:56:02,266 --> 00:56:05,036

So it's just like, you're

giving it props, it's performing

:

00:56:05,046 --> 00:56:08,686

computation and it's producing an

output, which is very convenient.

:

00:56:09,086 --> 00:56:12,726

And so I think with things like that,

where it's providing convenience

:

00:56:12,736 --> 00:56:17,486

and allowing you to interact with

information and do things with it and

:

00:56:17,496 --> 00:56:21,436

model things in a, some kind of virtual

space that would otherwise be hard to

:

00:56:21,436 --> 00:56:26,156

do, or summarizing things for you, or

like, you know, providing, distilling

:

00:56:26,166 --> 00:56:27,926

action items from a call, whatever.

:

00:56:28,291 --> 00:56:28,711

Totally.

:

00:56:28,741 --> 00:56:33,861

The idea that it will ever

produce marketing copy or content

:

00:56:33,921 --> 00:56:34,291

Sanford Whiteman: Mm hmm.

:

00:56:34,791 --> 00:56:37,501

Justin Norris: that is useful

seems really hard to imagine.

:

00:56:37,511 --> 00:56:40,021

A, because all the stuff I've

seen it do is just like really

:

00:56:40,331 --> 00:56:42,121

basic and uninteresting.

:

00:56:42,131 --> 00:56:44,541

And B, because it's

guaranteed to be average.

:

00:56:44,551 --> 00:56:48,511

It's always taking the average of

the average of the average of a

:

00:56:48,921 --> 00:56:49,481

Sanford Whiteman: Mm hmm.

:

00:56:49,531 --> 00:56:49,841

Justin Norris: points.

:

00:56:49,841 --> 00:56:52,158

And so, does evolution just stop there?

:

00:56:52,288 --> 00:56:53,878

Because it's not going to...

:

00:56:54,338 --> 00:56:58,138

It doesn't have that spark of

intuition or understanding of human

:

00:56:58,138 --> 00:57:01,368

nature that's going to say, like,

here's like a really funny video.

:

00:57:01,428 --> 00:57:03,868

Is AI going to be capable

of that sort of creativity?

:

00:57:03,868 --> 00:57:05,338

I just doubt it.

:

00:57:05,348 --> 00:57:09,228

So the idea of using it to

generate content to me is, seems

:

00:57:09,448 --> 00:57:10,708

contradictory on the surface.

:

00:57:10,898 --> 00:57:11,568

Who knows

:

00:57:11,698 --> 00:57:14,848

Sanford Whiteman: No, I, I, I,

no, I, I, I, I tend to agree.

:

00:57:14,848 --> 00:57:19,028

I mean, I'm, I'm, uh, I do follow

a couple of very funny comedians.

:

00:57:19,078 --> 00:57:21,198

Um, they're really more

like written comedians.

:

00:57:21,198 --> 00:57:27,028

Like, You know, punchline Twitter

comedians who, who show the futility of

:

00:57:27,038 --> 00:57:30,838

chat GPT at understanding jokes sometimes.

:

00:57:30,868 --> 00:57:33,888

And it's very like, it

definitely supports your concept.

:

00:57:33,898 --> 00:57:37,448

They're like, they'll feed it

an ancient vaudeville joke and

:

00:57:37,458 --> 00:57:39,208

ask it to explain the joke.

:

00:57:39,308 --> 00:57:44,088

And it just hallucinates a

completely wrong reason that

:

00:57:44,088 --> 00:57:45,038

it's supposed to be funny.

:

00:57:45,098 --> 00:57:47,418

Justin Norris: But I think to cap

that point, I think humor is the

:

00:57:47,418 --> 00:57:53,098

perfect example of like something so

basic that a child of two or three

:

00:57:53,523 --> 00:57:55,093

can understand a joke and laugh.

:

00:57:55,133 --> 00:57:57,733

And yet that the most

sophisticated computer models

:

00:57:57,918 --> 00:57:58,268

Sanford Whiteman: right,

:

00:57:58,553 --> 00:58:01,553

Justin Norris: properly explain, which

to some extent is a reassuring notion

:

00:58:01,553 --> 00:58:05,513

that there are some things that always

will remain sort of beyond that reach.

:

00:58:05,813 --> 00:58:08,443

I think we do have to wrap, but this

conversation is everything I hoped it

:

00:58:08,443 --> 00:58:11,233

would be, in the sense of just the,

:

00:58:11,288 --> 00:58:11,808

Sanford Whiteman: it was great.

:

00:58:11,928 --> 00:58:13,048

I could do this all The time.

:

00:58:13,173 --> 00:58:16,379

Justin Norris: lateral thinking, the,

various topics that we covered, and

:

00:58:16,394 --> 00:58:19,006

I'm glad to be able to share that

with the world first and foremost.

:

00:58:19,016 --> 00:58:22,746

Second, I appreciate you and all, all that

you give to the community and that you

:

00:58:22,856 --> 00:58:26,166

help me with personally being my phone

a friend, more times than I can count.

:

00:58:26,726 --> 00:58:28,606

And, uh, and sharing

your knowledge so freely.

:

00:58:28,606 --> 00:58:28,956

So,

:

00:58:29,236 --> 00:58:29,596

thank you,

:

00:58:29,791 --> 00:58:32,421

Sanford Whiteman: Well, you know, I

mean, you taught me originally the

:

00:58:32,421 --> 00:58:36,531

stuff that made me able to pretend to

understand Marketo from the back end.

:

00:58:36,531 --> 00:58:38,521

So I think you did, um,

:

00:58:38,616 --> 00:58:39,236

Justin Norris: we're a fraction

:

00:58:39,281 --> 00:58:40,821

Sanford Whiteman: uh,

way, way, way, way back.

:

00:58:41,031 --> 00:58:46,731

Um, but no, it's really to be on, I

think it's my first official podcast.

:

00:58:46,871 --> 00:58:51,271

So that's something if I ever start one

of my own, I will ask about your tech

:

00:58:51,271 --> 00:58:53,131

stack once you get it figured out here.

Links

Chapters

Video

More from YouTube