Artwork for podcast FinOpsPod
Tony Johnson - The One About Unit Economics
Episode 11 • 30th August 2022 • FinOpsPod • FinOps Foundation
00:00:00 00:33:41

Share Episode

Shownotes

Episode 11 Tony Johnson - The One About Unit Economics

Tony Johnson, fresh from presenting on Unit Economics at FinOps X, talks with Joe about how to start and mature using unit economics to manage your cloud environment.

Transcripts

TJ:

Hello, this is Tony Johnson and this is FinOpsPod.

Joe:

I'm Joe Daly.

Stacy:

Hey, and I'm Stacy case, and this is FinOps,

Joe:

Pod

Stacy:

pod's such a jerk.

Stacy:

You say this is FinOpsPod that's how we're starting today, folks.

Stacy:

So, Joe,

Joe:

yes.

Stacy:

What is the topic today and who did you speak with?

Joe:

I spoke with.

Joe:

Anthony Johnson, Cloud, Business Manager from Box about unit economics.

Stacy:

Okay.

Stacy:

So couple things, unit economics, I know is a trigger word for you.

Stacy:

so we'll have to get into that, but Anthony Johnson, who, if anybody's in

Stacy:

the community also sees as TJ Johnson.

Stacy:

Please tell me that the whole conversation is really just

Stacy:

about what his name really is.

Joe:

You know, every episode we start with a hi, I'm say my name, and this

Stacy:

my name, say my name.

Stacy:

Sorry.

Joe:

No it's great.

Stacy:

Okay, go on.

Stacy:

you start out with, my name is, and this is FinOpsPod.

Joe:

yeah, so I kind of tricked, Anthony, Tony, into, saying what his preferred

Joe:

name is by saying, you just go ahead.

Joe:

And say, hi, say your name and this is FinOpsPod and he did it

Joe:

and he called himself Tony Johnson.

Joe:

so I think Tony Johnson is his preferred.

Stacy:

Well,

Joe:

but he says he's . Very open.

Joe:

He's very flexible.

TJ:

if you wanna call me Anthony, a lot of people call me, Anthony people

TJ:

call me Tony people call me TJ.

TJ:

it's a very flexible naming convention, so to speak

Stacy:

All right

Joe:

there

Stacy:

have

Joe:

Mystery solved.

Joe:

We can end the podcast.

Stacy:

Done in the next week's podcast.

Stacy:

What is Joe short for?

Joe:

You won't believe what Joe stands for.

Stacy:

Okay, so we're talking about unit economics.

Stacy:

So I joke kind of that this is a trigger word for you.

Stacy:

Knowing you as part of this community, anytime someone said unit economics, it

Stacy:

wasn't just an eye roll on your part.

Stacy:

It was like a full body eye roll, like, uh, unit economics.

Stacy:

And like, I hate this.

Stacy:

I don't believe in it or whatever it is that you would say you were very

Stacy:

frustrated with the term unit economics.

Stacy:

So before we talk about the podcast, can you explain what

Stacy:

your frustration was with this.

Joe:

Let me read out of the good book, Cloud FinOps by J.R.

Joe:

Storment and Mike Fuller.

Joe:

As we open the book up, you gotta flip through all the good parts

Joe:

that have quotes from me in it.

Stacy:

Jeez louise

Stacy:

. . . Joe: yeah, true story.

Stacy:

If you walk up to me with your Cloud FinOps book, I will autograph the

Stacy:

pages that I'm quoted on for you.

Stacy:

So that that'll be a fun

Stacy:

more than

Joe:

do

Joe:

Think there are two or three, I don't know.

Joe:

There's like a whole chapter at some point.

Joe:

Anyway, Chapter 19 is Managing To Unit Economics, colon FinOps Nirvana.

Joe:

And it's a good chapter.

Joe:

Explains the concept.

Joe:

I'm not against the concept of unit economics that you link your

Joe:

cloud, spend to something valuable.

Stacy:

Mm-hmm.

Joe:

My complaint was, and I had a lot of good conversations with Mike

Joe:

Fuller on this is that as a leader of a centralized FinOps team, I have

Joe:

so many products and solutions and metrics, and there's so many different

Joe:

expenses that go into those solutions and metrics, not just cloud, that it

Joe:

wasn't really valuable to me to have a unit economic for each one of those.

Joe:

I didn't manage to 800 different numbers.

Joe:

I thought the way it was explained might have been misleading

Joe:

to the person in my shoes.

Joe:

However, if I was an app team, or if I had just one or a few solutions,

Joe:

like I thought that'd be great because I would be super concentrated.

Joe:

I would have so much more data, to be able to build those

Joe:

metrics into meaningful things.

Joe:

I personally don't think unit economics is the answer to every

Joe:

single FinOps question out there.

Joe:

And there are folks that do think that is the case.

Joe:

I think there's a lot more details and steps and thought processes

Joe:

that go into those sort of things.

Joe:

and it's okay.

Joe:

There's room for disagreement.

Joe:

FinOps is large enough that we can all not agree on things and that's okay.

Stacy:

To quote, Rob Martin, whenever there's a disagreement or a question

Stacy:

about FinOps, the answer is it depends.

Stacy:

So I think even with unit economics, it depends.

Joe:

A hundred percent.

Joe:

Now, Mike Fuller, if he was here on the phone with us, he would

Stacy:

oh, on Joe.

Stacy:

We have a caller.

Stacy:

We have a caller right now.

Stacy:

Oh yeah.

Stacy:

Caller Mike Fuller.

Stacy:

What do you have to say, Mike Fuller.

Stacy:

What would you be saying right now?

Joe:

Hi, I'm Mike Fuller.

Joe:

And I think Joe Daly is absolutely correct on everything he's saying.

Stacy:

That is probably the worst Australian accent and Mike Fuller

Stacy:

impersonation I've ever heard.

Joe:

Hi.

Joe:

I'm wearing my golden bedazzled AWS certification jacket

Stacy:

You better stop while you're not ahead.

Joe:

Mike Fuller.

Joe:

He would say, Joe, you're taking this far outta context and

Joe:

that's not what we meant at all.

Joe:

And to his point, they are updating this chapter, in the second

Joe:

edition of the book that J.R.

Joe:

and Mike are currently actively working on.

Joe:

They're updating this chapter to make me have fewer things to complain about.

Stacy:

I mean, that's probably not the only reason they're updating it.

Stacy:

I would imagine they're also updating it based on, you know, I don't know the

Stacy:

community and comments and just the maturity of the FinOps practice itself and

Stacy:

evolution, you know, that type of stuff, but probably a significant thing is so

Stacy:

you would stop harassing them as well.

Joe:

I think all the updating with new technologies and lessons learned.

Joe:

Yeah.

Joe:

That's a side benefit.

Joe:

Those get in the book too.

Stacy:

but it's Joe Daly effect.

Joe:

Yeah

Joe:

but I reached out to TJ, because he did a presentation at FinOps X and a chalk

Joe:

talk on the topic and it was fantastic.

Joe:

It really addressed a lot of my concerns and complaints about how

Joe:

unit economics gets presented.

Joe:

He breaks it down.

Joe:

He shows that there's a path.

Joe:

And you know, I started thinking, well, maybe I was

Joe:

doing unit economics all along.

Joe:

When talking to him about the maturity curve of it and you don't

Joe:

always get to that, how much revenue do you generate per cloud cost?

Joe:

But you find meaningful things to do, and there's a whole

Joe:

maturity learning curve through it.

Stacy:

Cool.

Stacy:

Well, I'm actually really glad that you had this conversation with TJ

Stacy:

because having kind of these two opposing views, and it's almost like

Stacy:

he probably was trying to convince you of something that you have

Stacy:

very strong feelings on, obviously.

Stacy:

I'm excited to hear this and see what happens

Joe:

Joe Daly interviewed TJ Johnson.

Joe:

And you can't believe what happens next.

Joe:

So I'll start it off by saying I have been a unit economic skeptic,

Joe:

throughout my FinOps career.

Joe:

not that I don't think they're valuable.

Joe:

It's just, I I didn't think they were as applicable.

Joe:

My statement on that was, is that as a centralized FinOps team lead, there

Joe:

was no one number that I could use.

Joe:

I have hundreds of applications I'm supporting , you know, all those

Joe:

application teams seem to me that it would be far more useful for

Joe:

them, but not as useful for me.

Joe:

I also feel like people have leveraged unit economics as a get

Joe:

out of a tough answer, easy answer.

Joe:

Someone will say, Hey, how do I do this?

Joe:

And someone will be like, have you tried unit economics?

Joe:

And I'm like, well, I want to see more meat in that answer than unit economics.

Joe:

Now having said this, you presented a breakout session, a really good breakout

Joe:

session in Ballroom B at FinOps X.

Joe:

And that presentation converted me a bit.

Joe:

I was like, oh, I was like, because you addressed all , my skepticism about it

TJ:

Fantastic.

Joe:

Yeah.

Joe:

You did a great job of showing the growth of.

Joe:

You don't just, you don't just run outta the gate, divide it by revenue

Joe:

and cost by revenue or whatever.

Joe:

but let's get to that in a bit.

Joe:

Not only did you do a breakout session at FinOps X and a chalk talk at FinOps X on

Joe:

unit economics, that were both phenomenal.

Joe:

You've been working with a working group inside the FinOps Foundation, and recently

Joe:

released an Intro To Unit Economics.

Joe:

All that to say you are very deep in unit economics knowledge.

Joe:

So for the listeners who are not as deep in unit economic knowledge, can you help

Joe:

define what is unit economics and kind of talk about the process the working group

Joe:

went through in defining unit economics.

TJ:

Yeah, absolutely.

TJ:

When we first started off a couple months ago, we did not have one definition.

TJ:

I think we had five or six or seven different definitions because it was

TJ:

really kind of a, a call out to say, Hey, how would you define unit economics?

TJ:

And then we would take.

TJ:

All of the input and try to come with a solution.

TJ:

Right?

TJ:

What does it mean?

TJ:

And how can we, together and collectively, look at defining a single definition.

TJ:

And I think that was really difficult early on, because we

TJ:

all found kind of our own little definition to be applicable to us.

TJ:

Right?

TJ:

And so when we talk about what that means today, we look at the

TJ:

different models that we put together.

TJ:

I think one of the words that kept standing out, in almost

TJ:

everybody's definition was value to the business in some way.

TJ:

when you look at the actual definition that we put out there, it's really.

TJ:

Not just unit economics, cause unit economics is a universal concept.

TJ:

you can use it for marketing purposes.

TJ:

You can use it for sales purposes.

TJ:

But when you talk about, FinOps, we're narrowing that down to

TJ:

cloud unit economics, right?

TJ:

Cause when you do these calculations, it's typically some cloud

TJ:

infrastructure number that gets added to the process in some way.

TJ:

And so, when we look at this, we say, okay, how are we going to define it?

TJ:

And so, we talk about, a system of profit maximizations based on

TJ:

objective measures, and how well your organization is performing.

TJ:

That's the basic kind of viewpoint that we said with not getting

TJ:

too technical with anything, but trying to make it a broad concept.

TJ:

And we all agreed to this because, like I said, each of us

TJ:

had our own perspective on it.

TJ:

Just for example, my perspective was, you know, we talk about a cloud unit

TJ:

economics calculation, so put it in terms of numerator and denominator.

TJ:

You're taking something and dividing it by something else to come up with a value.

TJ:

The numerator is really the cloud measurement of whatever

TJ:

you're going to have.

TJ:

And then the denominator is either an engineering value measurement

TJ:

or financial value measurement or business value measurement.

TJ:

And the outcome of that calculation is your cloud unit economics value

TJ:

measurement that you're going to use to manage some part of your business

TJ:

perhaps, or to give you insight into kind of what you're gonna do.

TJ:

And then you take that to the next step further.

TJ:

And you say as an organization, Hey, my KPI is gonna be a

TJ:

cost to serve a customer.

TJ:

I want it to be less than 5% of my overall whatever.

TJ:

But that gets you to this whole separate KPI conversation for me.

TJ:

You know, it's a defined target, it's an end kind of thing.

TJ:

I look at it more as, what I call a KMI.

TJ:

And a KMI is a key momentum indicator and,

Joe:

key momentum?

TJ:

Momentum indicator.

TJ:

Yes.

TJ:

So when you look at this, you know, you have a visualization.

TJ:

Most visualizations in FinOps will give you four pieces of information.

TJ:

It can give you awareness, like, how am I doing?

TJ:

And then if you want to look at it further, you can

TJ:

say, how do I get better?

TJ:

Which is really optimization stuff.

TJ:

And you can say, am I actually getting better?

TJ:

And that's really the KMI piece.

TJ:

Cause you can define over a time series.

TJ:

Whether this particular measurement is getting better over time.

TJ:

And what that allows you to do is it allows you to then make incremental

TJ:

decisions, to adjust to a more positive outcome of what you're looking for.

TJ:

And the KMI says, okay, have I reached my goal?

TJ:

Right?

TJ:

And that's the last one have I reached my goal?

TJ:

So with those four components in visualization: 'how am I doing?', 'how

TJ:

do I get better?' 'Am I actually getting better?' which is after

TJ:

the fact of the optimization, and then 'have I reached my goal?'

TJ:

So within all of that, you can define a cloud unit economics

TJ:

that really helps the business.

TJ:

Now, again, it's either gonna be typically engineering focused or

TJ:

financial focused or business focused.

Joe:

That's a good explanation.

Joe:

These sort of measurements in the cloud play more into that DevOps mindset,

Joe:

where the constant data's flowing in, feeding back into the decision

Joe:

making of the application team, the finance team and the business.

Joe:

So that's great.

Joe:

But before we get too deep in there,

TJ:

Sorry, did I, go too deep there?

Joe:

no, we both did.

Joe:

it's great, but the thing that I really appreciated about your presentation

Joe:

was that you showed a maturity model of how to start because I

Joe:

think a lot of folks may not be able to get to, costs per transaction.

Joe:

You know, huge enterprise may not be able to flow all that information

Joe:

down to a cloud team or an app team.

Joe:

But there are steps of here's how you start.

Joe:

You don't come up with cost per transaction on day one.

Joe:

Can you talk about the maturity model that you presented?

TJ:

Yeah.

TJ:

What's great about that is, is it changed from the time that I had

TJ:

presented into the stuff you're gonna see in the white paper.

TJ:

And it changed because I first started with a crawl, walk, run, and then based

TJ:

upon feedback, I added a pre crawl.

TJ:

So the pre-crawl is, you know, I'm not doing unit economics.

TJ:

And as we discussed on the summit in August, we looked at some of the

TJ:

patterns of adoption for unit economics.

TJ:

And we found that 50% are not doing anything yet.

TJ:

And so they would fall under the pre crawl, right?

TJ:

They're not even thinking about it or there's some issue with acceptance

TJ:

in the organization or they don't have the right data or whatever, for

TJ:

whatever reason they don't have it.

TJ:

And that's why the pre crawl started in, because I was always of the opinion

TJ:

that somewhere in the organization, some form of unit economics is taking place.

TJ:

And it may not be you.

TJ:

Right?

TJ:

And that's one of the first things that I recommend for people to do is search

TJ:

out allies that are doing this already.

TJ:

And typically it's gonna be somebody, that's providing usually from finance.

TJ:

Uh, usually it's providing some data to a business unit owner

TJ:

on how they're doing right.

TJ:

And that becomes a potential measurement for you to include, cloud information

TJ:

infrastructure, especially if you are a cloud first or a cloud native

TJ:

organization and you're focused on that as a path for your future.

Joe:

So find the dashboard that the team's using, that the leader of the

Joe:

team is using and then find out who makes that dashboard and go talk to them.

TJ:

Yeah.

TJ:

And talk to them and start looking at how you can contribute to what they're

TJ:

providing , and, honestly, if it's meaningful or not, you know, that's

TJ:

what you wanna figure out as well.

TJ:

Cause it may not be meaningful to them.

TJ:

And then you say, okay, great.

TJ:

I know this is already happening here.

TJ:

Let's see what else we could do that would be meaningful for a business

TJ:

unit order or for finance or whatever.

Joe:

That's such a beautiful point.

TJ:

So we start with that, then we go to the crawl, right?

TJ:

And so the crawl, you know, we were talking about distinctions between

TJ:

unit cost and unit economics.

TJ:

And what was the distinction between them?

TJ:

Typically unit cost is part of the calculation for unit economics.

TJ:

And so there's a, there's definitely a difference.

TJ:

Right?

TJ:

And so unit cost was the crawl phase.

TJ:

Now you've gotten to your calculations, whether you're taking a blended rate off

TJ:

of your invoice and you've got your unit cost or you're mapping it to the actual

TJ:

contract and you can have an unblended unit rates where you have different rates.

TJ:

So you need to figure out how it's deployed and how you're buying

TJ:

it then and figure that out.

TJ:

So you've got your unit cost.

TJ:

That's at the crawl.

TJ:

So if you're able to produce a report around that, great

TJ:

congratulations you're at crawl.

TJ:

And then you start looking at the next phase.

TJ:

Walk was really what I call the unit cost integration, where now

TJ:

you're applying that unit cost somewhere within other calculations.

TJ:

It's not really a unit economics kind of decision making measurement.

TJ:

It's really, to help you understand and create more awareness around

TJ:

what's going on in the organization.

TJ:

An example would be I'm taking my forecast by unit and I'm converting it

TJ:

to a budget based upon the unit cost numbers that I precalculated, and that

TJ:

allows for things to happen better.

TJ:

So I'm integrating the unit cost into this process.

TJ:

And so therefore I would say you at probably a walk phase and then

TJ:

you move to the run phase where you have more about cost to serve.

TJ:

So you were talking about cost per transaction is a cost to serve model.

TJ:

And that's when you're really integrating that with another measurement.

TJ:

So now you have two measurements coming together that are now giving

TJ:

you this new calculated number that says, okay, whether you're doing

TJ:

well or not, or whatever your KMI or KPI around that is going to be.

TJ:

And that's something that you now track differently.

TJ:

So it's not the unit cost piece.

TJ:

That's how you start separating it.

TJ:

At least in my mind, that's how I started separating it out.

TJ:

It makes it easier to understand the progression of where you're at.

TJ:

So if you're using a unit cost, integrating it with another

TJ:

measurement, you're doing unit economics.

TJ:

So these are all kind of things that I've experienced in different levels,

TJ:

maturity, working at Accenture, and Citibank and other companies, and

TJ:

really trying to help people understand what their journey is as well.

Joe:

So on your experience, what were some of the first unit economics that,

Joe:

that you measured and how did they mature?

TJ:

So the crazy part is I have a strong finance background.

TJ:

So as an accountant and chief accountant, I would produce those, and we're

TJ:

talking a long time ago, I'm old.

TJ:

So we're talking about the eighties right at this point.

TJ:

So, I was doing that, not knowing that there was a term for it so to speak.

TJ:

And so you get involved in these processes You know, once you start

TJ:

creating one, what's really interesting is that you now have a demand for more, cause

TJ:

you're like, oh, I can get that number?

TJ:

That's really cool.

TJ:

What else can you do?

TJ:

Right?

TJ:

And so I became kind of that specialist for the management team that was

TJ:

able to create new kind of metrics that the company hadn't seen before.

TJ:

And it was really helpful for them to understand how to manage it

TJ:

with those types of measurements.

TJ:

So I started, in my finance career before I shifted over to tech.

TJ:

And then eventually to FinOps

Joe:

Well, that explains why you were able to convert.

Joe:

I got my accounting background as well.

Joe:

It took another former accountant to convert me.

Joe:

Another thing I think you and the working group do a really good job of

Joe:

explaining there's cost to produce, versus cost to sell and show that you

Joe:

can compare that cost to those other types of costs for an application.

Joe:

Tell me more about that.

TJ:

So, cost to produce, what we're looking at from an overall perspective

TJ:

is saying, okay, what does it cost to get me to the point where I'm in production?

TJ:

So all non-production workloads can be used as a cost within

TJ:

the cost to produce calculation.

TJ:

So I'm saying, okay, to build and test this particular workload.

TJ:

It costs me, you know, $85, whatever, but I have a similar stack with this

TJ:

other team and it cost them $400.

TJ:

What's the difference?

TJ:

Why?

TJ:

Why did it cost them so much more to produce a very similar kind of workload?

TJ:

I need to start looking at what are they doing differently?

TJ:

Is it a labor issue?

TJ:

Are they overstaffed?

TJ:

Is there efficiency in code or not?

TJ:

And so you start diving into these additional questions when you start

TJ:

looking at these type of comparisons.

TJ:

Now from cost to serve model, you've got everything in production.

TJ:

So this is the cost to really serve.

TJ:

The way that I separated that out allows me to think about, okay, I

TJ:

know my production costs are gonna be, you know, at this point a little bit

TJ:

larger than my non-production costs.

TJ:

Although sometimes I've seen the opposite where non-production is 60% of the overall

TJ:

spend and production is only 40% cause you have all these different environments

TJ:

that you're putting your resources into.

TJ:

And if you have sandbox, dev, test, stage, that all adds up.

TJ:

You have to be really efficient about your non-production to save costs as well.

Joe:

I think that's fantastic, the idea of comparing similar non-prod environments.

Joe:

And when the unit economics are not lining up between the two that's a

Joe:

sign to ask questions and investigate.

Joe:

Is it labor?

Joe:

Is it code?

Joe:

What is it?

Joe:

And can the two teams learn off each other and, and help leverage that.

Joe:

That's where the FinOps is helping drive collaboration.

TJ:

Yeah, you know, I find, there's a social context of FinOps as well.

TJ:

It's in that communication stuff.

TJ:

And how do you build those relationships with the different team members?

TJ:

And do you have a communication plan?

TJ:

Do you have a content plan?

TJ:

What's your frequency of communication?

TJ:

How are you doing that?

TJ:

When you have conflict, are you using emotional intelligence

TJ:

in the conversation?

TJ:

Right?

TJ:

Cause sometimes, we can get drawn into a conversation that

TJ:

maybe doesn't work for us.

TJ:

Right?

TJ:

And how do you then handle that?

TJ:

And I think the social aspect of FinOps is really critical because

TJ:

that's how you get consensus.

TJ:

That's how you get, the collaboration and you work together to get to a

TJ:

point where it's good for everybody.

TJ:

And I think that's that's missing in some way that we don't talk enough about.

Joe:

I a hundred percent agree with you because, there's so many ways to

Joe:

solve a problem, however, there's not one way to solve every problem.

Joe:

And it really depends on the context of what the app team is.

Joe:

What is the architecture of the solution?

Joe:

Can the solution handle a modern architecture

TJ:

yeah.

Joe:

Should it have been moved to the cloud in the first place?

Joe:

All questions that are easy to answer after the fact, however,

Joe:

you know, that doesn't necessarily make things easy for them.

Joe:

I think if you can help drive visibility to all sorts of

Joe:

different reasons that a team may be succeeding or may not be succeed.

Joe:

That helps.

Joe:

I love how you're framing this.

Joe:

You're using these metrics, KPIs, economics to help drive

Joe:

transparency to find those problems.

Joe:

I think that's a really, really mature way of looking at it.

Joe:

That's fantastic.

Joe:

So you're at Box now, we talked a little bit about your career path.

Joe:

Cause you are the chief business, actually

TJ:

So I'm the cloud business manager.

TJ:

So, so what the heck does that mean?

TJ:

Right?

TJ:

I am responsible for the coordination of all of this

TJ:

information with all of the teams.

TJ:

I also start special projects like sustainability is one of them.

TJ:

And so looking at how do we then communicate that?

TJ:

In my role at Box, and it hasn't been necessarily every role that I've

TJ:

worked in, but specifically at Box, communication and the social aspect

TJ:

is super important because there's just so many people to work with.

TJ:

In engineering we have thousands of people.

TJ:

But it comes down to a few, a dozen, that, the relationship needs to be

TJ:

there to move forward, like to gain acceptance on a new mandatory tag

TJ:

or something like that, that needs to go throughout all of engineering.

TJ:

Things like that you have to have those relationships where you can have these

TJ:

rather candid conversations and sometimes difficult and complex conversations.

TJ:

But if you don't have a good relationship with the other teams, it's gonna

TJ:

take you much longer to be successful.

TJ:

And, you know, this is all about speed.

TJ:

If we can accelerate together, we'll be able to do more together.

TJ:

And so this comes down to having those good levels of

TJ:

communication and relationship.

Joe:

Yeah So how many unit economics metrics are you leveraging?

Joe:

Are those internal to you?

Joe:

or are they used to support those teams that you're communicating with?

TJ:

We're still defining what we wanna move with.

TJ:

Right.

TJ:

We're migrating to the cloud.

TJ:

Right.

TJ:

The challenge is that, when we look at how we do any calculations

TJ:

or any measurements in context to all the other things that are happening,

TJ:

do we have enough information?

TJ:

Let's say for the migration process to say, I have enough.

TJ:

To give you a number that you can action on, right?

TJ:

And you may not have enough data.

TJ:

You may know the calculation that you want to do.

TJ:

You just don't have enough data that's really gonna be meaningful and material

TJ:

enough to impact the business yet.

TJ:

And so we're at that goofy stage where we're in between that being

TJ:

able to say, okay, here's something that we know that you can action on.

TJ:

Here's something that is interesting, like awareness for now, but when

TJ:

we get more data, then it'll be more meaningful for you.

TJ:

So we're at that stage now throughout Box to be able to do that.

TJ:

Now, there, there are other teams, business unit teams and sales

TJ:

teams, that, that have that number, that special number for them.

TJ:

Right.

TJ:

As I was mentioning before, somebody's doing it somewhere else in your

TJ:

organization already on some level.

TJ:

But specifically the ones that are affected by our FinOps team, we're

TJ:

growing those constantly, but yes, many of them are internal today to our team.

Joe:

So how do you do that?

Joe:

Because you're absolutely right.

Joe:

If you don't have the relationships, you're not gonna have any credibility

Joe:

for anyone to accept these numbers.

Joe:

So how, what are the processes you're doing to develop those

Joe:

relationships and mature those numbers.

Joe:

Are you presenting them monthly?

Joe:

Who are you presenting them to?

Joe:

How are you growing them?

TJ:

Yeah.

TJ:

So it started with, hopefully as accurate as we can, awareness.

TJ:

Right.

TJ:

So let me tell you what's going on for you, . And those conversations

TJ:

dove into many different directions, everything from architecture, to, you

TJ:

know, you have 90% waste on these five clusters, what are we gonna do about it?

TJ:

And how can I help you get better at what you're doing?

TJ:

So, I see our goal as supporting the business and whatever the business is.

TJ:

And if we can give you information that helps you do that, that's

TJ:

really, success for us, cause we want to be able to make you successful.

TJ:

And we take on this, persona of, let me help you get to where you need to

TJ:

be and be the most successful you can.

TJ:

Right.

TJ:

And success is measured in lowering cost, efficiency, a couple different

TJ:

metrics that we track internally.

TJ:

We do a fantastic job and our team is just amazing.

TJ:

I've been able to turn around visualizations in four hours for a

TJ:

leader and really have it be meaningful.

TJ:

So I'm just so proud of my team.

Joe:

So that's great.

Joe:

You're using the business metrics, cause some businesses might be

Joe:

in growth stage and their goal may be to get as many customers

Joe:

as possible, as wide as possible.

Joe:

Others may be like, no, we're a stable company.

Joe:

We need to focus on cost control.

Joe:

Or we're going through this project or this transition of something.

Joe:

And they just need to do it in a controlled and managed fashion.

Joe:

And that's what they're putting their value on.

Joe:

So you're able to use those metrics to help the teams focus and prioritize versus

Joe:

, you know, no waste is good waste, but some waste is more important than other.

TJ:

Yes.

TJ:

Yes.

TJ:

We look at a system generated optimization that comes

TJ:

from Cost Explorer, whatever.

TJ:

And then we have what we call a Curated Optimization.

TJ:

So these are things that we look from an architecture standpoint, we look from

TJ:

a different perspective than a, than just a formula and say, okay, if we did

TJ:

this, now we can really do this so that we have what ifs of moving from one

TJ:

instance type to another that isn't on a recommendation algorithm, but it's on

TJ:

our radar simply because, we look at our contract pricing, we look at different

TJ:

things and we say this is probably a better way for us to move forward.

TJ:

and let's look at what that option would look like and what

TJ:

would the savings be for that?

TJ:

So that's more of a curated kind of optimization and we do at

TJ:

any given time, we've probably got about 20 of those going on.

TJ:

So we're trying to provide alternatives for the teams as well and not just come

TJ:

up with 'Oh right-size down a t-shirt size and you're good to go' because, what's

TJ:

a term my kids use, so that's so 2010.

TJ:

But so it's gotten to the point where we've developed tactics around

TJ:

optimization and we've got over 150 tactics across nine different categories

TJ:

that we can use for our savings program.

TJ:

We have five direct programs and we have four indirect programs.

TJ:

And so as we go through each of these, different tactics, you know, we're able

TJ:

to then shift as the business shifts.

TJ:

Maybe we can't do any, , physical optimizations because

TJ:

of a business reason.

TJ:

So we'll shift to a policy optimization.

TJ:

how do you change this policy that then now affects how you want

TJ:

to run your services that then lowers the cost of the service.

TJ:

So things like that.

Joe:

And then you can use those different playbooks that you're basically play

Joe:

calling kind of and get buy-in by showing the impact on the unit economics.

TJ:

we are in the NFL preseason now.

TJ:

So I think you know, playbooks and calling is actually appropriate.

Joe:

I like that.

Joe:

Wow.

Joe:

That's really cool.

Joe:

I'm really intrigued by having those predefined plays that, we're gonna go with

Joe:

a policy decision here because, they're a pet and they can't handle a reboot

Joe:

right now or something along those lines.

TJ:

Yeah.

TJ:

Procurement is also one of the direct one, right?

TJ:

So how well can you negotiate?

TJ:

I think one of the things I said at the session was, don't

TJ:

always take your contract at face value everything's negotiable.

TJ:

But negotiate to how you consume the cloud.

TJ:

This is a really important piece and it sets the stage

TJ:

for then how we deploy, right?

TJ:

Cause let's say for our, for argument's sake, you're able to get a killer

TJ:

deal on a certain, family of compute.

TJ:

you're gonna focus on that family of compute and then build upon three or

TJ:

four different sizes, small, medium, and large in that family, because

TJ:

you have such a great deal on that.

TJ:

And you're gonna see now, of course you'll have exceptions, but is probably

TJ:

gonna be like 90% of what you consume is gonna be within that family, cause

TJ:

you gotta killer deal off of that.

TJ:

And so that's, those are things that you would look at in a procurement model.

Joe:

Yeah, so true.

Joe:

if you don't talk to your procurement team, your partners, if they're not

Joe:

your best friends, they should be your best friends in the company.

Joe:

Because one contract can invalidate the most efficient code there is.

TJ:

hundred percent.

Joe:

So where is the working group going with unit economics next?

Joe:

We got the intro to unit economics released.

Joe:

It's on FinOps.Org, for folks to start absorbing.

Joe:

Where does it go next from here for folks who are just getting in and

Joe:

starting to go up that maturity model.

TJ:

Yeah.

TJ:

So that was really our sprint one end, right, is to to produce the

TJ:

document and publish it on the site.

TJ:

So sprint two is going to further define the maturity and a couple

TJ:

other topics that we wanted to bring up for sprint two.

TJ:

We definitely want feedback.

TJ:

On the sprint one information that's published up there , and, really

TJ:

get somebody to further define it.

TJ:

But I think it's extending more use cases so people can relate

TJ:

to hopefully their industry.

TJ:

A lot of people are looking for examples of how people can implement

TJ:

it and how people have implemented it, to help them do their own process

Joe:

all right.

Joe:

The call to action is to help.

Joe:

If you have examples of how you are leveraging unit economics,

Joe:

join the working group and share.

Joe:

Well, thank you so much, Tony.

Joe:

This was a fantastic conversation.

TJ:

I appreciate the invite.

TJ:

Thanks so much.

Joe:

Continuing to convert me into a solid believer.

Joe:

I was doing unit economics without even realizing

Joe:

it.

TJ:

awesome.

TJ:

Well, I appreciate that.

TJ:

Thanks.

Mike:

Hey everyone, the real Mike Fuller here.

Mike:

If I were on the phone with Joe and Stacy, I'd say that the true FinOps

Mike:

Nirvana is listening to FinOpsPod.

Mike:

Well, except for all the parts where Joe's ranting, especially about

Mike:

unit economics, that's all rubbish.

Mike:

I wanted to thank Tony Johnson for talking some sense into Joe about unit

Mike:

economics and also a huge thank you to our contributors in the introducing

Mike:

unit economics working group, you can see all their smiling faces at the

Mike:

bottom of the white paper on FinOps.org.

Mike:

If you want to get involved in our working groups, join the foundation.

Mike:

You can find the unit economics working group in slack at the

Mike:

WG unit economics channel.

Mike:

Thanks Stacy.

Mike:

You always bring such high energy and Joe leave the Ozzie accent to me next time.

Mike:

That's all for this episode.

Chapters

Video

More from YouTube