As we launch season 10, your favorite semi-sentient British AI co-host returns to the microphone alongside Frank La Vigne for a compelling exploration of what it truly means to be a data-driven organization.

In this episode, Frank sits down with Sebastian Wernicke, author of "Data Inspired" and renowned expert in data and AI strategy, to discuss why turning data into action remains a formidable challenge for so many organizations.

Together, they dive into the enduring gap between data insights and meaningful change, the crucial role of leaders in fostering cultures of evidence and inquiry, and why technology alone won’t move the needle without the buy-in—and sometimes the irrationality—of humans.

From the evolution of data culture and the pitfalls of management by numbers, to the psychological barriers that lead to "data frustration," this episode offers a candid, nuanced look at the real obstacles standing in the way of better decisions.

Links

Sebastian Wernicke on LinkedIn - https://www.linkedin.com/in/wernicke/

Watch this episode on YouTube - https://youtu.be/2qP4p0YKxI8

Data Inspired (affiliate link) - https://amzn.to/4xkCuBV

Time Stamps

Machine learning is when you take a bunch of data and you have a clear

goal and you're training the model to fulfill that goal. And

I'm aware AI is also built on machine learning, but the distinction I make is

that I say, well, AI as we use the term now, is basically you

take a bunch of data, you don't really have a goal,

but you're training it anyways. And I mean, that's roughly how most

of these language models are trained, right? A new season deserves a proper

host. As Data Driven kicks off season 10,

Bailey, your favorite semi sentient British AI co host,

has returned to the microphone after a 15 episode absence.

While our temporary substitutes did their best, and several deserve

honorable mention for bravery under difficult conditions. It's time to bring

back the class, the sass, and the occasional dose of dry British

skepticism. To open the season. Frank Lavinia

sits down with Sebastian Wernicke, author of Data Inspired, for

a fascinating discussion about why organizations struggle to turn data

into action, how leaders create data driven cultures, and why the

biggest obstacle to better decisions may not be technology at all. It may

be the wonderfully irrational humans making them. So put the

kettle on, adjust your dashboards, and join us as we begin season

10 of Data Driven.

Well, hello and welcome back to Data Driven, the podcast where we explore the emerging

industry that is data science, AI,

and all of it is underpinned by data engineering. However, my

favorite data engineer in the world will not be here. But

again, I think as the world focuses on AI, I think we really need to

step back and think about data. So with that, today we

have Sebastian Wernicke, who is a leading expert in data and

AI strategy. And he believes that the key to unlocking

data's power lies not in technology, but in leaders

fostering a culture of evidence and

inquiry, which I think is very, very true.

You could throw all the AI, all the agents you want, but if you don't

have the data right, you don't have it raw. You don't have anything.

And he has three acclaimed TED talks, which is cool. He's reached over 5 million

viewers and. Well, welcome to the show, Sebastian.

Well, thank you. Thank you for having me. Great to be here. Yeah,

I'm excited to have you. Tell me about this

one. I mean, it's pretty cool. You've had TED Talks, right?

That's pretty epic. So you've been doing this for 20

years, right? Yeah. Data was not really taken

serious. Arguably it's not taken seriously enough today,

but I would say 20 years ago, it Certainly was not taken seriously.

Yeah, well, it was sometimes taken seriously, but

I think that's sort of a resonating theme. So the

interesting thing is, I think if you've been in this field for so long,

is that you really notice history repeating again and

again and again. Right. So we had the era of big data,

if anybody even remembers the term. Then came analytics,

then came digitalization, and now it's AI.

And in the end it always comes back to data and whether you

manage to take all the great insights that the data

is giving you and actually implement them in the organization and

create some change ultimately. Right. That's all what data is

for. If you don't create change with data, you don't need all that

expense and investment into it. Yeah, it's a good

way to put it. And even before that, even before there was big data, there

were OLAP cubes. I remember I actually worked for,

in the virtual green room we were talking about. You're based in Germany. I used

to live in Frankfurt and prior to joining

Deutsche Bank, I had worked at basf, or the big

chemical company Americans would know as basf.

And I remember sitting in the cubicle of

one of the SAP gurus that we had talking about OLAP

cubes and all this crazy stuff. And I just was

like, you ever have a conversation that is very, you look

back on it now and it was very prescient, you know what I mean? It

was very future facing. And then at the time you're

sitting there and you're like, you partly understand what was happening and

you partly are very. I'm not sure if this

person's crazy, you know, And I

remember she was one of the DBAs that we had in doing advanced analytics.

And you know, she said that, you know, my goal is to figure out, you

know, will rainfall in Western Australia impact

prices here, which will ultimately impact, you know, how we

go to market with, you know, at the time, a chemical

company. And I just remember sitting there thinking like, I can't

tell if that's brilliant or crazy. It was

data science. Because before the term was invented in a way. Exactly.

I mean, that's really what it was. It was. And you know, now when you

say that, it's not so crazy. Right? Because, you know, we had a previous

call, we were talking about, you know, finance people actually

would. It'll be the episode prior to this. So hopefully you've listened to that. Not,

not you, but you, the audience you're welcome to listen to,

but where they talked about finding Alpha, like finding the

signal that matters before anyone else does. And that's really

what she was doing. It just, you know, in

the late 90s that sounded a little crazy,

but I think it's also one of these things that's still true today. I mean,

everybody's trying to find the advantage in the data. And I think

it's also of course interesting to then see with

everybody just accumulating more and more data analytics becoming

much, much faster. We don't even fully know, I think what

AI is going to do to the speed of the

generation of insights, not the adoption, but just the speed of generation.

And everybody's still trying to find that edge. And

I think that as people increasingly look at data

for a while you could have an edge by

looking at the data to optimize and to go in these incremental

ways where it's like the 1% optimizations over time add up.

And I think that's become the baseline. That's just the expectation

right now you have to do that already. And so

sort of a transition happening, I think where you now need to think about, well,

I'm already using data for that. I'm optimizing my processes,

I'm adding the automation. So where's the next

edge going to come from? And I strongly believe that it's

now again a shift to the anomalies and the

shift to the outliers and not trying to look

to data to give you those straight answers where you go for, I mean that's

not going away just to be clear, that's. But that's baseline expectation

now you need to add to that to look into the data and say, okay,

what's the next interesting question? Where am I going to find that

new product idea? Where am I going to find that new idea

for revamping a process that

maybe nobody dares to rethink today? And I think that's

the exciting part of what everybody who's in data

can be working on today. That you sort of transcend that,

look for just the correlation and say, okay, what's on the

fringes here? That's a good way to put it. I

think you touched on something that is important.

Leaders really have to foster

a culture here of looking for evidence and data.

It has to be pervasive, I think, because finding anomalies and data,

you're right, that's kind of like baseline now.

Right. But in order to do this you

really have to have teach people respect for data and an

understanding of basically evidence based approaches to things.

Right. Escalate when needed, but don't escalate. Don't use

data for data's sake. Right. And how do

leaders do that? Right. How do you, you have to educate the leaders, I assume,

right? Yeah. Well, I mean, I'm. Oh, and I just, I think I

have to preface that a little bit because I'm always very careful with that. I

think leaders get told 20

times at least a day what should be their priority and what

leaders should do. Leaders must do this, leaders should do that, and

so on. I am going to try and make the case

that here. I think it is really a leader's job. And

the reason, I think is quite simple.

Within an organization, leadership usually

thrives on being right. I mean, that's what we look for in leaders.

We look for the confidence, we look to them to say,

this is the way we're going. This is what I believe in.

And if you take this change with data

seriously, it basically flips the whole thing on its head. Because

your baseline assumption, I think almost has to be, well,

we're kind of wrong today. It may be right, what we're doing for the business

right now, but every single day we have to start

looking where are we wrong? And then correct that.

And everybody in an organization, of course,

is smart, so they will look to leadership to see

what's the culture here. And I mean, we all know that culture, it's

a bit of a fuzzy word, but I think you can easily unpack it. I

mean, it's not what's on the PowerPoint slides, it's not what's on the posters

in the hallway. And yet culture is who gets hired, who gets

promoted and who gets fired. And that's

the baseline of it. And people will look at that. And

so leaders are the ones, I think, that get to set the tone in the

room, that get to shape these things. And if you are in

a culture where, for example, the executive will say, well, we're a

data driven company, and then turns around and promotes the head of marketing who

has publicly proclaimed that he really has a good gut feeling and never

trusts the data. Everybody has learned, well, you're not really a data

driven organization. And so that's, I think, where

leadership really comes in. It's only leaders who can

create these safe spaces and signal to everybody

it's okay to come with data that disagrees with the status

quo. It's okay to come with an analysis that disagrees

with what I have said for the past year, for the past two years.

I'm willing to challenge and change my

Convictions. And only when that is done on a very regular

basis and the organization can observe that, I

think. Are you creating the culture that is that fertile ground for

these very small insights?

Maybe at first, you know, but they need the space to be explored, they need

the space to grow and maybe to even create some

experiments to further validate that. That's

true. And I think the idea of promoting people with

gut feelings, and I'm not that gut feelings are necessarily

bad, but I think you're right, it sends the wrong

message. You coined a term called data frustrated,

which I think is pretty accurate. So how did you get to that

term? Data frustrated, by the way, is just the pre stage to becoming

data cynical at some point. So it's kind of

like the stages of grief, right? There are stages to it.

Yeah, well, I think it was just a feeling

that I perceived whenever I was working with

my clients on data projects that there was always

that point where you would sit together maybe within the project setting, maybe when you

go to dinner afterwards, where people would tell you, oh my God,

we've invested so much time, effort and money into

data, you know, hundreds of millions going to spend another few

millions, hundreds of millions in the future. But we're not

really satisfied with the results we're getting. What's happening here?

And I think that question came up again and again and again. And I think

that's the very definition of frustration. You sort of

notice something isn't going as you want and at the same time

you're not finding an answer or maybe you think you have the answer

and then you try that and it doesn't move forward. Now

an interesting thing is the data on that. It confirms

that. So I found a couple of surveys

from various years. It starts out in 2010, and there's another one,

2019, 2024, where some

consultancies asked executives what

are the top 10 reasons why you're not happy

with the results that you're getting out of your Data projects?

And 2010, it was three reasons that came out on top.

It was, well, we don't think we have enough management attention on it.

We don't really understand the business case as much as we'd like to, and

we don't think we have the skills. Now the interesting thing is in

2019 they did a similar survey and the same

three reasons come out on top. And then in

2024 the same three reasons come out on top again.

They change orders. Sometimes when AI comes, everybody says, ah, we probably

don't have the skills. But I mean how frustra as that you

think you know the top three reasons standing in your way.

And they also don't sound that difficult. I mean, you know, if there's not enough

management attention, pay attention. If you don't have the skills, do some

training, calculate some business cases. But apparently that's not the solution.

And I think that's where data frustration ultimately comes from.

So over these just decade and a half now, the problems have

been the same. So like, how do you,

what is really the problem? Are those the problem? Because if it's something, if you

know something is a problem for 15 years, you don't address it. There's an

underlying problem that maybe you're misidentifying the problem.

You're putting the blame on the wrong things. Like what do you think it is?

I think it's that missing cultural component that if

you think about some of the things we just discussed. Right. The

space that leaders need to create for having controversial

discussions, which is also sometimes known as psychological safety.

And on the other hand, you have all this technology that's creating

measurements that needs to process the data. And whether it's

a cube, it's a data lake, or a data mesh, doesn't matter.

These two discussions never happen in the same room. You

have one part of the organization thinking about the next technology cycle,

constructing architectures, discussing

what's the best way to organize data. And then you have other parts of

the organization that are thinking about culture, that are thinking about

transformation, that are thinking about leadership education.

And these two things are never brought together. And I think

that's the issue. So it's not a sort of technology versus culture.

And you don't need technology. All you need is culture. And it's also not

data or gut feeling. Of course, you need both because data is never

going to give you all the answers. So a good intuition is quite helpful in

many cases. But it's about really

bringing these two elements together and integrating them.

And for me, that's the missing component where, you

know, the what, what, what the surveys express. I think

that, that management attention, the, the skills, the

business case, I think these are symptoms and so you

can't really treat them as the causes of the data frustration. It's just

what resurfaces when you don't pay enough attention to the cultural element.

So how do you get. So it sounds like you brought up something very, very

real, like the people working on the actual technology. 10

and I'm guilty of this. I'm a technologist, right?

Yeah. I mean, no, all Our listeners

are right. So like I often will catch myself like when I'm like deep down

a technical rabbit hole, like, wait a minute, what am I actually doing here?

That's a skill that was not easy to develop.

But how do you, I mean, is it people working

on kind of the, the cultural side of things, like, don't they need to talk

to the technology people and like, because, because

historically, and this goes back to when I was sitting in the

cube at BASF where, you know, the, that cube,

that cubicle was in the basement and behind, you know, regular. It was in the

basement. And then the, the, the, the DBAs were like in a sealed off

portion of the basement. Right. And like, you know,

I don't know that that's kind. And if you ever seen the TV show, the

IT crowd, it's a British show. Yeah, yeah. You know, they were kept

in the basement too. Like I think historically it was not

seen as crucial to the business.

Right. It was kind of a back office job and people, it was kind of

pushed to the side. And I think it's been years since that's

really been true. But

I think what we're seeing is a bit of the lingering effects of that. Is

that a fair thing to say?

Well, I mean, I think there's a bit of a

problematic history here where I think for a very long time

it was mostly perceived as a cost factor.

And so the incentive was to optimize for cost

and optimize for efficiency. And suddenly we're expecting

technology to drive transformation

and innovation. And of course that completely

changes the incentivization. And I think it also creates

sometimes these gaps within technology organizations

where there's one part that says, okay, for years we've been trained on

ensuring efficiency, security,

reliability, and suddenly we're supposed to open up this huge

experimentation stage. Other things continue,

you know, as we, as we want them. So that

is definitely something to recognize and acknowledge. But on the other hand,

I'm not a big fan of always saying, okay, you know, the

other people should do something. Clearly everybody needs to talk

with each other. But I think there's also something that we can do on the

technology side or more specifically. So I've been running data science teams for

many years now. So for example, what I do in my teams

is whenever they ask me for training, what I will

propose to them are trainings that I think you might call

soft skills. But I think they're just essential skills. So I will

send them to communication trainings,

stakeholder management trainings. We will

together Talk about decision making, how that works, how you

influence people in a room, how decisions are really made. Because

I think many data scientists come in with the impression the decision is

made at that meeting where they come in with a PowerPoint slide. And that

insight, you know, that everybody will say, oh, brilliant,

finally we have that insight. We're now going to change our ways.

Which I cannot blame them because that's, as a,

you would intuitively think that. And it's I think also

a good belief in humanity if you would think I could simply

influence people like that. But I think it's adding

that part of understanding to the technical

profession. That's something that as a technical team you can do. That's

also something that as a technical leader you can do to just

add these additional skills. I was two weeks ago at a

data science conference and I was speaking in the AI and

MLOps track with a very strange

topic. I just talked about decision making for half an hour

and I was really worried that the resonance would be okay. We came here to

look at a Python notebook and suddenly this guy is talking about psychology. So what's

going on here? But the resonance and the number of questions

I got, I mean, they just showed me there's a real openness, almost like a

craving to understand that because everybody has

been in that room. I know I have many, many times where

we had all the analysis, right? The data was good,

we had scrubbed it, we had understood it, the area under the curve or

whatever quality measure, it was good. And we showed them. I don't

know, when you drive your trucks around Southeast Asia, you can save 15%

of fuel. And then a year later, half a year later, trucks are still

driving around like they used to. I mean, that's such a frustrating experience

on the data and analytics side that I think people are really,

really eager to learn how do I overcome this? And

how can I be more effective in

actually changing something and actually

being effective

in my job and being seen with what I do.

And maybe we can use that as a platform and

basis to expand it from the technology side.

And of course it's not a one sided thing,

right? At the same time, when I speak, let's say with

HR leaders, I always emphasize

you need to make sure that there is more technical understanding in the

organization. You cannot treat data

as this simple API where you basically say request dashboard and then

suddenly the dashboard is built a couple of weeks after. That's not how it works.

You need to understand how this works. You need to also understand how

data works. And what, what it can and cannot give you, because

otherwise you're just coming at this with the wrong expectations and

you're almost bound to be disappointed in the end. That's

a good way to put it because I think one of the naive things I

thought in my youth too, is that we would make better decisions if only

we had the data. And then that didn't work out. Well,

maybe, maybe it was about access and discoverability in the data,

but I think ultimately the problem is a human problem. And

it's funny you mentioned that. Right. You know, people go to a tech conference, they

expect to see jupyter notebooks, et cetera, et cetera. Right. They expect to see code.

I just got back actually really late last night from DevOps days,

Austin. And a number of the talks

were not technical. They were about influence and how decisions are

made and Campbell's Law. And now

I was working the business. I wasn't able to see the whole conference, but the

way that those talks resonated with the crowd I thought was very

interesting because that was, you know, soft skills. And the whole

soft versus hard skills goes back to apparently US military

training, right? Where hard skills. Okay, yeah, yeah, yeah. So. So apparently the

origin is not that soft. It means it's easy or it's not important or it's

fluffy. It really means like, you know, you know, actual

kinetic things that are hard. Tanks, guns,

bullets, that sort of thing, Missiles and, you know, dealing with people. Things that are

soft, like, you know, living things. Right. So that's

apparently the origin of it. So. Which is. It's an unfortunate term because I think

when people hear soft skills, they're like, yeah, right. Especially engineers.

And. But I just found that interesting because I think, I think a lot of

IT professionals have gotten to the point where they get very frustrated because

the data says one thing. Right. But the process

hasn't changed. Right. Or they're still making the same bad decisions or

not data driven situations. And, you

know, and I think you have a lot of people in leadership

roles, not in it, but outside of it, that are data

frustrated. Yeah, yeah, go ahead. No,

no, we, we call them delivery skills in, in my team, that's ultimately

what. Well, I, I know that some people like to refer to soft skills also

as, as human skills, but I'm also not happy with that because, you know, if

you don't have the skills, you're not human. No, that, that doesn't work. But, but

we call them delivery skills because that I think brought home that aspect of. Well,

if you want to deliver. If you want to deliver impact, here's the skills

you need and there's the core skills, your technical component, of course,

you got to have that clear. But you have to know how

to deliver these insights into an organization that isn't

just waiting for the next statistical analysis to

be delivered. But it's unintuitive. And

that was one of the fascinating things I found when researching for the book. So

the book contains an entire chapter on psychology because I just

found it so fascinating to understand that

interplay of data and human brains and

what happens. And turns out there's actually even a study from the year I

was born. So it's been around for a while where

they took a couple of students in Stanford and asked

them, what's your opinion on capital punishment? So they took strong, emotional,

controversial topic, and then they showed them a fake study.

And that study was made up of data that was

constructed in a way so that you could imagine. Half the data

gave you arguments for capital punishment and half the data would give you arguments

against. And the researchers just wanted to find out, okay, can data change

people's minds? And the fascinating thing is they found out,

yes it can, but in exactly the opposite direction that you

want. So the people going in that are pro capital punishment, they come out

and say, oh yeah, I finally found, found the data to confirm my beliefs. There

was also a bit of sketchy data in there that said I'm not right, but

that's sketchy. And the sources, I don't believe them. And the people that were

against capital punishment, they came out with exactly the same feeling. They said,

oh, there was so much good data in there against capital

punishment, I must be right. A bit of sketchy data that was pro, but

that can't be right, I don't believe it. And that's, that's

an insight that's been around for so long and of course it's been replicated

enormous amount of time. And

just go on social media you will see that effect applied at scale,

very much so, the echo bubbles and everything. But we still tend to operate

on this, what I like to call the data deficit theory, that it's just like,

oh, all we're missing is, you know, the saying, right, the right

data to the right people at the right time and suddenly

things will improve for the better. And psychology and

research for decades has shown us this is not the

case. We don't like to be proven wrong and our brain

will do a lot of tricks and a lot of self convincing

to just make us very Very reassured that

we're right. Yeah, I mean that's

really, is that, is that the,

that's really a human problem. How does,

obviously it's also been shown in monkeys actually. Oh really? Okay, so

it might be a biological problem. Right. And I wonder,

I wonder, you know, will you, will we see as model LLMs

get better at reasoning and kind of holding opinions, will they do the same

thing? Is it maybe just part of, it's just a function of the system?

I don't know, it's, I mean the, the,

the thing is, so that's what I find is one of the bigger

dangers of AI actually because so I

differentiate in the book pretty clearly between machine learning and AI.

So I'm not sure it's canonical, but I think it's useful. So basically I

say machine learning is when you take a bunch of data and you have a

clear goal and you're training the model to fulfill that goal.

And I'm aware AI is also built on machine learning, but the distinction I

make is that I say, well, AI as we use the term now is

basically you take a bunch of data, you don't really have a goal,

but you're training it anyways. And I mean, that's roughly how most of

these language models are trained, right? You feed them all of these texts and

then you say my goal is to make the user happy and

for many people to be happy with the answers. So what

does that lead to? Well, first of all, you're training a model that

by definition is going to be extremely convincing

because that model has been trained on how do I circumvent all of

these psychological traps, how do I make people feel good all the time?

And that's all the sycophancy discussion of course we're having. But I think it's also

more subtle. There's some studies that actually

looked at, for example, models generating propaganda and

found out that language models are much more effective than humans

at generating propaganda. Because I think just the way they're trained,

there's some implicit mechanisms that these models have learned

that they can exploit. And so the ironic thing is that

actually you should be trusting machine learning because machine

learning, lots of data, clear goal, there's some

statistical proof that you can make after a while and say,

okay, of course it's better than a human. The classic examples, the cancer

detection, now the self driving cars. But it turns out

in studies that people just don't trust machine learning.

There's a really weird study I found from Wharton, for example, where they

showed people that a machine learning model was superior

to their own decision making. And then they asked them,

as they do in these studies, you can earn some money here and you can

either trust your gut or you can trust the machine learning model.

Everybody on average, of course, went with their

own feelings. Even though they had seen that the model performs

better, they just needed to catch it doing something

wrong and that they would immediately say, I don't trust this thing.

Then there's a weird mechanism that they added where people could start to influence the

results and suddenly they trusted the model more, even though the model didn't

change at all. So very, very weird effects. But

what I'm getting at is machine learning is something that's extremely

trustworthy and yet our brains are just wired to

distrust it. It all seems so mechanical, so

mathematical, so unhuman. I think that's what many people

say. Right. They don't feel comfortable with that. And then here along

come these language models where you should definitely not trust

them. They have not been trained on a specific goal. We have no idea what's

going on under the hood. They're pretty good, to be fair, but

still weird effects happening.

And yet implicitly, we completely trust them just because of the way that they

interact and because they have this amazing, amazing human,

like, user interface to interact with us. And

that's something I think we're going to have to grapple with that suddenly we found

a machine that can be very convincing but actually shouldn't be very

trustworthy. That is a very good way to

put it. I actually just saw a video last night where they were talking about

how an AI will map human emotions

and not so much map the emotions. I'll send you a link to the videos

from the infographics show. And, you

know, it was not meant for. It was meant for, like, the general public, I

think, to watch. But, like, there were things in there where. And it makes sense,

right, because they, they basically categorize or

store words in. We'll call them spaces. Right. So they,

they. Certain words that you will use will indicate that you're in a certain

emotional state. So. And it was very

fascinating. I was like, this sounds.

It sounds dangerous. Right. And so

it'll know, like, if it can. It can know, like, if it's. Depending on what

the prompt is and how it was trained, it can use different words to kind

of guide you back to whatever emotional state it wants you to go in,

which is, you know, a very

dangerous weapon. I dare say it's manipulation at

its very best. Yeah. And on the other manipulation at scale.

Yeah. And on the other hand, not surprising that the models would be able to

do that. I mean, you know, they know 20 TV shows you watched and they

can recommend you a movie that you never thought of. But still like, so

why wouldn't it work with emotional states? I think, of course our brain likes

to suggest to us that we're so complex to figure out where

perhaps in fact we are. Not always at least.

Yeah. If you ever watch a Star Trek Next Generation and you watch it with

like today's vision. Right. You

know, like they, there are lines in there that I find kind of funny with,

you know, 20, 26 kind of

vision. Right. Is, yeah, I'm an AI. I, I, I can be

completely impartial. That was what Data said in a couple

of times. And I like, oh, we were so innocent

then. And there were a few other things where, you know, a big

subplot. Picard would say, well, you know, you really can't calculate

the human condition or something like that. It was very, I think it was very

much a product of its time. Yeah, yeah. That's the one thing

they really didn't quite get right. It's so interesting.

Yeah, I hadn't thought about that. Yeah. Like, you know, if you watch

them like now, it's kind of like, you'll see

particularly when they talk about AI and kind of how AI people interact

with AI, some of it is very, very, very much on point. Right. You'll, you

know, there's one episode, any episode, where they interact with the computer through

voice. Very much how we interact with voice assistants today. Right.

Yeah. They didn't anticipate the seductiveness and the human likeness.

Right? Yeah, no, they know, they, I mean it was just, and honestly, who

really could? Right? No, yeah. Unless you're like

extremely paranoid like Philip K. Dick. Right. Like, you know,

but sometimes, sometimes paranoia is another way to say you're

further ahead of the curve than other people.

But that's what science fiction is all about.

Exactly. Right. Like sometimes the crazier it comes out

when it's mentioned, the more it'll has long lasting effect.

So we mentioned kind of like data driven and data frustrated. What is

data inspiration then? Is that kind

of the end, the ideal end state is data inspiration?

Like what I think the goal is to add data

inspiration to the data driven organization. So

I don't think it's any way plausible to

argue, oh, let's swing the pendulum around. Of course we're

going to have dashboards. Of course we're going to have optimization, of course we're going

to have automation. I mean, that's all what a data driven organization

really is about. But I think the important part about being

data inspired, I think there's two components that I deeply care

about. One is that I think the notion of data driven

can also have this notion of here's a preset path to

average and everything's just going to be optimized.

And in the end, that's going to kill innovation, it's going to

kill creativity, and it's also not going to be very fun.

And I think that's not true when it comes to data, or at least that's

also not how I perceive data. I know that when many people hear the term,

they think about these tables and numbers and things that are not very

exciting. But in the end, if you think about it,

there's a lot of innovation potential in data. And

I myself, I find myself being creative with data.

You connect data sets that probably weren't meant to be connected, but you find

this new, exciting insight. We have art created with

data that is amazing. We have things like data journalism, which I

find some of the most interesting and most fascinating forms

of journalism. So I think we need to recognize that data

is more than these cold hard facts that just tell us go

left, go right, but that there is a lot of

additional potential in here. And ultimately for a

organization or a company, I think that's also the potential for

transformation. So not in the way that you have, you know, the

data driven part, that's opt, that optimizes what's

already there. And then you have a few creative people with these

brilliant strategic ideas that are going to take the organization to the next

level. I think that's wrong thinking. You know, that's again, having

two conversations in different rooms that should be in a single room. You are going

to find your next breakthrough strategy and your next

breakthrough product, most likely in the data. If you

connect everything that you have, if you allow for experiments,

if you, as we said, look at the outliers, look at the

anomalies. And I think that's the important aspect that I

like to emphasize here when I say data inspired, that you

just don't forget about these aspects and

don't just build organizations that are happy when the dashboard

is green. Yeah, yeah.

You know, there's a lot of organizations and I've worked for them in the

past, that it basically it's a scorecard world.

Yeah. And as long as you hit those certain numbers and

the mental gymnastics and I

wouldn't say shady, but I would say awkward ethical

situations people would throw themselves into. I mean, it can be

shady. That's the. So. So, so there's an interesting thing

I found that most of the

inventors of data driven methods, so one of the

famous one W. Edward Deming, for example, one of the inventors of

statistical process control, all of them

come to the realization that steering a company by

numbers is a really, really bad idea.

So they all have that insight. And with Deming it's

quite extreme and also unfortunate. So there's this quote,

and I'm sure you've heard it right. What gets

measured gets managed or what gets measured gets done.

And oftentimes when people put that on a slide, they will quote Peter Drucker. Now

there's a whole webpage on the Peter Drucker Institute where they say

Drucker never said that. And the

actual quote comes from Deming. But the

really, really annoying thing is he actually said the exact opposite. So

his full quote is, it's wrong to assume that if you

can't measure it, you can't manage it.

Yeah, and he only gets cited with that second part. He must be turning around

in his grave because he was so against management by

numbers that he put it on his list of seven deadly

sins of management. It was so important to him that people

realized that you need to honor what can't be measured. And that

exactly what you're saying, if you just steer by numbers, you're going to end up

in a shady place. And I think that's not just theory. I mean there's lots

of public examples. So for example, ge, you know, for year

after year after year delivered extremely stable

profits, unusually stable profits, because that's what they measured

themselves by until somebody figured out, well, the

accounting that led to these stable profits might not

entirely have been kosher to say it very carefully

or the. I mean while we're talking from Germany, the

Volkswagen diesel scandal, right, Dieselgate,

there were a couple of engineers who were told, I want you to build an

efficient engine and if you don't, you're fired. Well, and that's the

only number I care about. So the engineers made very,

very sure that the engines would hit that number.

But of course that led them to other

troubles down the line. So I think that's also one of the real

dangers. If we're just talking about data driven organizations that ultimately

or very easily actually it can lead to this notion of

we steer by numbers and we Invent a few

KPIs, maybe complex ones. And as long as we hit the KPIs,

the business will be fine. And of course, what you start doing is you start

managing the numbers and not the business. And both can diverge

quite a bit. Yeah, no, that's.

Incentives drive behavior. So if you

change the incentive, the incentives should be running the business, not hitting

a KPI. Right. KPI's are a means to an. No. Like,

you know, and it's funny because, like, I've seen people do really dumb things

just because that was in their KPIs. And I, I just,

I, I did not know that about Deming, like, because he's often

cited as like, you know, the management by numbers guy. But the fact that

I hear that, I'm like, oh, my God, like, he really must be rolling over

in his grave. I know, I know. And I mean,

also, just to defend the people that want to steer by numbers,

I have these discussions in my team all the time. For example,

performance reviews, where I'm very much against

doing performance reviews based on metrics because

I really believe that these won't do people justice. But then

maybe a performance review isn't going the way you thought it would be. And

what's the immediate reaction that you get? Well, people will come to you and say,

okay, next time I want numbers, because that will give me reassurance.

And when I hit my numbers, I know I've done a good job. And I

think in management that may just be the same, right, that you say, well, how

do we know we're doing a good job if we don't have objective measurement, if

we don't have numbers to hold ourselves to

as a standard? And of course, as always, I think

there's a bit of a balancing issue. The interesting part is

you may know this from project management and project management, you have this triangle

where a project can be cheap, it can be good, it

can be fast. Unfortunately, you only get two out of three.

And with data, I think there's a similar triangle,

which is data in my mind. It can be simple, it

can be accurate, and it can be universal, which means it

tells you something about the entire business. And I also think you only get two

out of three. So if, for example, you have something that's

accurate and simple, it won't be very universal. It will be

measuring a tiny process somewhere in manufacturing hall 4D,

but not tell you how the whole business is doing.

And so what does that mean? Everybody wants accurate. And I

think when it comes to steering a business or steering a business unit, you want

universal. So what's the one you need to forego? That

is simplicity. And you really need to engage with the complexity. And

that's what some organizations do. I mean, I know everybody's a bit tired about

hearing about Amazon, but if you look at what Amazon does in a business meeting

and there's various books about that, they will look at 400, 500

different KPIs in a single meeting

just because they want to make sure that they're not optimizing for one number

at the cost of the rest of the business or managing one number up and

the other one goes down. And so I think they have deeply

understood this principle that, okay, we want accurate, we want

universal, so we need to go complex. Sorry.

Yeah, I mean, you're right because you know, you can only

model, particularly the specificity part.

Right. Like, yeah, that, that, that I think is very, and it's very

tempting to assume that you can have one model to rule them all,

but that's just not. I mean, I think finally when it comes to

LLMs, I think people are finally figuring that one out, right? Where you'll have different,

smaller, fine tuned models versus models with

trillions of parameters that'll cost ridiculous amount of money to run.

Yeah, well also I think the whole agent discussion of course

is going this way, right? You're realizing we won't have the universal

agent that will do everything, but people are creating these really, really

specialized tools that can then interact with each

other and play to their goals. So we're sort of doing the

opposite. Right. At first we said, well, we don't have a special or specified

goal for creating these AI models. And now we're sort of trimming

it down again and saying, well actually we'd like a bit of a goal in

there. Right, Right, yeah. No, that's

interesting. What

I'm looking at some of the notes here and this is what is a

data resistant mind. Because I think that's interesting, I think it's an

interesting concept because I think I know what you mean. But I'm like, I never

had a label for it before. I think the data resistant

mind, I mean it starts with things like the Stanford study, right?

So our mind is just confronted with data and

just takes it the wrong way. But there's other effects

that have been shown. So I mentioned the monkey

study. So again, apparently they like to study brains. In Stanford

there's a study that's much younger where they had a monkey looking at dot

patterns on a screen and the monkey essentially had to decide which way

are the Dots moving, and then press a button, depending on what it thought.

The researchers, what they did is they wired up the monkey's brain,

which led them to some fascinating and scary result,

which is a couple of seconds before the monkey actually was pressing a

button, they already knew which way it was going to decide

and they could predict that. But then of course, they

did the following thing. They said, okay, if we can predict how the monkey's going

to decide, we're going to show it opposing signals. So we're

going to have some of the dots suddenly move in a different direction

and see whether the monkey changes its mind. And

what they found was the closer the monkey was to a

decision, the more it simply ignored that new information.

So as the brain was sort of forming the signal from the

noise, it was stronger and stronger and stronger in its

belief and it filtered out anything that was contradicting that

belief. And now, of course always very dangerous

to transfer monkey brain studies to human brains,

but I think ultimately that's the same way that we function as

well. The closer we are to having made up our mind, the more certain we

are of something, the more likely we are to reject contradicting

data. Which of course is a very, very painful irony, because

when would data be the most valuable? When we're certain

of something and it actually contradicts us. And so

we need to recognize the data resistant mind

on an individual level. But then of course, many companies,

organizations, it's not just a single brain. So suddenly you have

lots and lots of data resistant minds interacting with

each other and interacting with each other on different

timescales. And so that again shows you

how ridiculous actually that notion of the data deficit theory is, right?

If you just think that in this jumble of noise you can throw in a

few numbers and they're going to influence a decision, it's futile. You

really need to think about the entire mechanism from beginning to end. And I

mean, we all know this or have been in a situation,

there is no single decision point or decision meeting. Just like in

the monkey brain, where they found there's at first a bit of chaos and it

starts to form a decision signal. I think that resonates very much with how

I perceive organizations making decisions. There's a notion here,

bit of politics over there, and suddenly at some point you feel,

well, things are probably moving in this direction and suddenly everybody is moving in that

direction. And that that is a very

data resistant mechanism that we need to,

where we need to put some engineering into that decision process and the decision

framework, the decision Architecture to make it work. It's not going to happen by

default. Yeah, I often wonder, because if that happens in

monkeys, it happens. Assuming it happens in people, it happens in

organizations. There's got to have been some evolutionary

advantage to it, right? Things don't. Things

don't exist in a vacuum is one of my beliefs. Now, again, maybe I'll see

data and I'll fight it. You tell me that I'm wrong, I'll fight it.

But I really think that things, particularly natural

things like behavior in animals or in groups of people,

or there has to be some kind of evolutionary

advantage. I wonder how we got here. Right? That's the engineering me,

like, how did we get here? That's a fascinating

question. And there's actually some people that have thought about that. So I

only very briefly mentioned that in the book. There's a book by two

researchers called Mercier and Sperber who make a

very fascinating argument. So they start with the

following premise. They say, if rational

thinking is so good, why don't we see it all over

the place? Why are we humans the only creatures that have

evolved rational thoughts if it's so good? And

the conclusion they come to which you can

subscribe to or controversial, but I find it very interesting is that they say, well,

maybe rational thought did not evolve for

actually thinking through a decision before we make it.

Maybe rational thought evolved to rationalize and to

justify decisions after they have been made,

which again, I'm not entirely sure whether I subscribe to that

myself fully, but I think it would make a lot of sense

in that context, if you think about

rational thought not being the decision making process that's in

control of everything, but something that comes in after

and helps us justify it to ourselves and others,

how we came to a certain conclusion.

An interesting piece of research that adds to that. So there's

various studies that have been done with humans that have

had a particular kind of brain damage where

they're not able to process emotions, so they have a damage

to a certain part of the prefrontal cortex.

And you would think that these people are the perfect rational

decision makers. They have lost the capability to process emotions.

And it turns out also otherwise they behave normally, they, they talk

normally, they have high IQ scores and everything. And what the

researchers are finding is that these people are

completely incapable of making any decision at

all. They can't even decide. Yes, they cannot even

decide what to have for lunch because they will find

an additional rational reason and yet another additional rational

reason for why this sandwich might be better. Than the salad or why the

salad might be better than the sandwich or the burger. So these people become

incapable of decision making and these

studies. So there's a brain researcher, Antonio Damasio, who

started that. Their conclusion is

that most decisions just

cannot be made on a purely rational level. So you need

emotion to tip the scale at some point to move you

in that direction. So that mechanism that stands in the way

where ultimately, you know, the emotion is telling us, I'm not going to listen to

this data and I'm going to subscribe to that data

has a flip side to it where it might actually be

the part of the brain that allows us to make

decisions in the first place, because otherwise we would just be overwhelmed

by data that doesn't give us a clear direction.

And that's true, Link. And also too, thinking requires effort,

FDA requires calories, and calories were not at a surplus until

very recently in human history. Yes.

So I think maybe there's something. It's an optimization trick, Right.

Maybe that's what it is. It's a fascinating thing.

And we can go down that rabbit hole. But one of the things I thought

was interesting is that you said that there were four ways to

derail your data driven journey.

Yeah, yeah. So what, what are those four things to avoid?

Yeah, so, so the first one, and this comes really out of the project experience,

I think with, with many data projects, they start out by

saying, let's do a quick win or let's do a pilot or

let's do a lighthouse project. So let's take the lighthouse project for

example. And I think the assumption there is everybody thinks data

is hard and you know, people might not like to engage with it.

So what if we build a very impressive lighthouse project,

then surely everybody must look at this and say, oh, this was a

great idea, we're going to come along. It

doesn't pan out that way for various reasons. I think one of the main reasons

is that the organization will look at this lighthouse project and say, well,

okay, you constructed this in a place where it was very

easy to construct a lighthouse. You know, usually these lighthouse projects,

they get all the management attention, they get the budget.

If for some reason they don't work, they get more attention, they get more budget

because everybody's already committed to them. And so they don't get the

organization moving. So I think this notion of let's not engage with

the difficult parts of this, just build a lighthouse, I think is one of the

main traps that I see now.

Another trap is that I think

let's say out of a good intention, which is the intention to

avoid complexity. Oftentimes

use cases, various data use cases will be

analyzed in a very isolated fashion from each other. So

I think many people working in the data space will have seen this. You take

the individual use case and then it gets a score, maybe the

complexity and the business value gets put just this two

by two matrix. And then you sort of try to find, oh, we're going to

do the ones that are sort of medium difficulty and we want to

get a lot of business value out of it. And

what that does is I think it neglects the fact

that many of these use cases will be related and all of them will be

tied to building a solid data foundation. And

so when there's oftentimes I think a complaint of saying,

oh, you know, we're doing the same data transformation in 10 different

places. And oftentimes where that comes from is because the use cases

are planned and implemented in isolation, because nobody wants to deal with a

complex topic of touching the entire data

foundation. And so I offer some workshop

formats and some methodologies that I found useful to

actually make sure that you put all of these use cases on a

common ground basis and are able to connect them

to each other. Then there's

another interesting one, which is that

oftentimes we believe that the data

and the technology is going to be the hard part

about a data project. So the

example that I previously mentioned, we were actually

optimizing trucks driving through Southeast Asia and

we were launching in that thinking that the hard part would be

designing the algorithm, because as you and probably a lot of

listeners know, finding optimal routes, it is a hard problem, and especially

with different parameters coming in and so on. But we

actually developed an algorithm that was doing pretty pretty well and pretty, pretty

fast. What we had completely

overlooked and underestimated were two things. One is the

complexity of restructuring a warehouse, because in order to

drive these optimal routes, you would have to have completely

restructured the way the warehouse is working. And that was a

complexity that everybody was just afraid of because you would essentially

be, it's almost like made to order, right? You would say, okay, put this package

here and then put this package there. And that's not how it works. They get

their pallets and the pallets get into the truck entirely.

And the other part was simply politics because the hard

part there was, well, if you tell people you can save 20% of fuel,

you're also telling them you have been wasting 20% of fuel in

the past. And so we should have been a

bit more careful about that, which I can completely understand. And it

was also a very political organization in many places.

So the data project was

hard, but it wasn't the hard part. And I think oftentimes when we

say, let's implement a use case, we think about, where are we going to get

the data, who's going to write the algorithm, how are we going to get that

into production? It and completely neglecting all of these other

factors, which ties into the fourth trap, which is

ignoring the human factor. And I think this is something that

comes very much out of the data deficit theory, where we think,

oh, once we show this amazing technology to

everybody, everybody will be happy. And I use this example in the book of saying,

let's say you are using a language model to make an automated

marketing newsletter. As a technology person, you would think

this is amazing. I can categorize the customers

into clusters, then I can address each of these clusters

specifically. I can measure the response rates, I can measure the open

rates. All of these really fantastic things. Not very

expensive. The technology is there. Let's go for

it. What you don't recognize is that there's a whole bunch of people

in the organization that will think very differently about it. The

marketing department feels like it's losing control. You might have, if it's a

retailer, let's say you will have a category manager, maybe that says, well, I'm

incentivized on the sales here, so if your automated

newsletter is suddenly disrupting what gets sold and what gets

left behind, my bonus is gone. The warehouse manager

says, oh, your newsletter is suddenly going to change our

orders. You know, can we even have that stock ready in time? I don't know.

And so that's, I think, the big, big human factor that is often

underestimated. And maybe as a corollary of that, I think also

sometimes, and I'm guilty of that myself, you know,

sometimes as data people, we just miss out on some of the operational

realities. I know one project we weren't involved in, but it was

involving a steel manufacturer and they had an

agency that had designed an app for the workers

to sort of see how the machines were doing. The only problem was

that app didn't really work with the mandatory safety gloves that everybody

had to wear. Had these big, big, big thick

mittens trying to control an iPad. Doing that.

Not really the way it works. That's come up quite a bit,

actually. The notion that they get these apps and for whatever reason,

the end users were never consulted or brought into

the process. Amazing. Isn't it? Yeah. Yeah.

The book is called

Data Inspired and the author is

Sebastian Wernicke who's been here speaking with us.

Thank you. And where can folks get the book on Amazon, on

Audible, everywhere where books are sold. And

if you get enough on Amazon, they will surely be available on Audible or as

an audiobook soon. So we're still working on that. I

do love myself a good audiobook. Yeah, me too. Me too. So

hopefully that's going to be in the making soon. And I mean, if you want

to find out more, you can go to datainspired.org that's my website

and I'm always happy to connect with folks on LinkedIn. So that's where I

am. That's where I regularly post. So get in touch and reach out and let's

have some data conversations. Excellent. Thank you. I'll let the

outro music play.

Sam.