In this second interview about 3X thinking, Antony Marcano and I talked about advanced topics:

  • The fractal nature of 3X
  • How 3X relates to the three investment horizons model
  • Pioneers / settlers / town planners
  • How 3X relates to Agile
  • How companies can apply 3X and try radical, new ideas while protecting their existing brands

But first we do a short recap of the last video: What is 3X - or explore / expand / extract - and how can you use it? Watch the video here or read the transcript below.

David: Welcome back to this video series about 3X thinking, In our second - and for now, last - video about “eXplore - eXpand - eXtract” or, how it’s also called, 3X. I’m here with Antony Marcano who will explain the concept more to me and to you.

So, Hello Antony!

Antony: Hi David!

David: In the last video, we did some talking about the theory, like how 3X thinking was invented by Kent Beck or was first thought up by Kent Beck and he did a lot of thinking around it. And you, too, did a lot of thinking and teaching around it.

And it basically is a thinking tool that tries to help you figure out in which phase of an idea or a product you currently are and which techniques to apply, right?

Antony: Yea, in these terms, yea!

Recap: eXplore, eXpand, eXtract

David: So, it links the success of an idea or a product to it’s payoff. And this link is not linear. And I think we should, as a short recap of the last video, explain what that means and the three phases. Just very quickly.

Antony: Yea, it’s three phases of an idea. The X axis is in terms of success, the Y axis is in terms of payoff.

Success is what signals that your idea is of value to your customer and the payoff is what signals that it’s of value to you, the creator of that thing.

Antony and David discuss how payoff is linked to success of an idea and the three phases explore, expand and extract

For example, for products - products may start as just a very small experiment. It could be nothing more than a quick prototype. You do like a faked screencast of it and put the screencast out there. And people might just see that thing and are “No, I don’t like that”. So you try another approach and another one and people say “That’s a really cool product, when can we buy it”.

Now you’ve tested the market and then you might still go on and do further experiments where you create a very basic version of the product and maybe a few people start using it. And then you start to figure out: How is this product gonna be of a payoff to us? Like, you may not know how to make money yet, so you might be exploring different ways of which it could make money.

And really once you’ve could turn the bend from when you discover an idea that is successful - So, you’re customers want it or some customers want it, might not be your current customers, might be completely different customers that you stumble across and you find out how it can be of a benefit to you, then you enter the expand phase.

Expand is really just about survival. Becaus it may be growing at such a fast rate, you literally have to throw in infrastructure and money and people just to survive the expansion phase. But once this starts to tail off and success slows down and payoff slows down, then you are into extract.

And at that point, it may be more beneficial to look into how can we start automating or simplifying or lowering the operational cost of how we actually got through that expansion phase and to extract.

And it’s the extract area that many large companies think they’re operating in. And they’ll apply the same principles where you’re extracting the value from something to an idea that really should be considered being in the explore phase.

David: Yeah, I think I’ve seen that too and you have said it in the last video, that a lot of companies try to play by the rules of extract but they are actually in explore. And I think I’ve seen that at my clients too, where they define a solution, write down a user story for a solution where they did zero experiments, they write down in very much detail what they want, then they implement it and then they release it to the world. And then maybe it works or not, but they’re already trying to do this defined process and cutting costs and everything that you would expect from someone in the extract phase but they are really trying things.

Antony: Exactly, that’s exactly what I see also. And that’s where I think this model is valuable. To help you recognize when - First, to have a language that you could use to share with other people, where you can ask the question: Are we applying extract thinking to an explore problem here? And then “maybe it is an explore problem”. You can have a conversation around it.

And I’ve actually been leaning on this quite a lot in the last year with my clients. And it’s actually been really, really helpfult. It helps us - I really get a lot of more bureaucratic steps, shall we say, out of the way of the things we’ve identified as “explore”.

And one of my clients, in fact, has benefitted from this model enough that they - as we worked our way through what kinds of problems are explore, expand or extract problems within their world, that resulted in them really thinking “are our teams structured effectively to actually fully exploit this?”

And what they’ve done, for now, is they’ve split their teams in - sliced their teams in both horizontal and vertical. And the verticals take care of looking at what ideas are successful with our consumers. And they’re operating predominatly in the explore space. And what they do is, they explore and expand, and begin expansion.

And the horizontal teams are much more about the technology, innovation around the technology, the platforms, support, the existing features, the platforms that they can put in place to now start supporting the new ideas that are coming out of the vertical teams. And they kind of complete the expansion process and transition it into extract.

So, they’ve actually had a huge, huge benefit from it. In fact, I found out on the other day that in the first six months this year alone they managed more than - because of the rate that they were able to test ideas and then get them out to market, they were able to nearly double revenue in one of their product areas, that’s sort-of been left alone for a while.

And it was given both the attention and the speed of being able to try ideas allowed them to arrive at the things of what’d gonna work for them, and in the first 6 months, they doubled the revenue at that part of the product.

David: That sounds awesome and I think that also illustrates a point that we didn’t really talk about in the last video, which is the fractal nature of this thing. That it’s not “we, as a company, are in extract”, but every product of the company and every feature idea in every product might be in a different phase, right?

The Three Investment Horizons

Antony: Yeah, and so the company, in itself, can be in extract. And some of it’s products might be in extract. And some of them might be expand, and some of them, lots of ideas, hopefully, would be things they’re exploring.

And I think companies of the future, and even companies of today, even if they don’t realize it doing this, the most successful ones are, I think, are doing this.

It ties in somewhat with an investment model called three horizons, which is a much older idea. It comes from the book called “The Alchemy of Growth”. And in the three horizons your first horizon was the horizon… So horizon one is your current business. That’s what’s making you money today, generates today’s cash flow. And horizon two are the growth areas, the things that are expanding. And then horizon three is the place where you’re generating future growth options. So this is where you’ll be trying lots of different experiments to see what’s the next thing that’s gonna grow your business.

The three investment horizons from The Alchemy of Growth

And obviously, these horizons - they’re rolling horizons. As in, there comes a point where products may become irrelevant and it becomes your past. And the thing that was your horizon two becomes your horizon one and the things that were your - some of the things that were your horizon three become your horizon two. And then you need to generate new ideas.

So, this comes from “The Alchemy of Growth”. When I was looking at this I realized that actually, those horizons are really just the “extract” parts of a 3X curve. Because horizon one is where you are extracting today, and you should then be expanding another idea and exploring other ideas. Horizon two is where extract will be once you’ve continued growing an idea. And horizon three is really the extract phase for whatever idea you have explored that you’ve then taken through the growth phase and into horizon three.

Annotations added to the last graph - How the three horizons can be linked to 3X

So I think many - this is the idea of having this sort of three very different perspecitves - is not a new idea.

In terms of “The Alchemy of Growth”, this was told about how we distribute our investment. So gernerally, the advice from the authors of that book were talking about the majority of your money should go into generating today’s cash flow because that’s the lowest risk investment. So, a lesser amount going into your; invested in growth opportunities. And then having a smaller amount of investment in lots of different ideas because you don’t know which one of those ideas is going to become your horizon two and which one is going to become horizon one.

And I think what 3X gives us is a… So, the three horizons is one way of thinking about how to invest - Whether it’s time, money, resources - when I say resources, I don’t mean people - you know, office space, etc. And, but what 3X gives us is a way of thinking about what game are we playing right now? What are the rules for that game? And are we still playing - Are we playing by the rules of the current game or are we still playing by the rules of the old game we were playing?

So, it just gives us a different way of thinking about “how do we apply that investment?”

So, three horizons tells us how much do we invest, 3X is a way of telling us how do we apply that investment.

David: Let’s talk about investment later maybe and stick with how you apply the wrong rules to the phase you’re in. I mean, I think it’s pretty understandable if you’re an established company, like a Telco, and you’re operating a customer portal where people can order the newest phones or whatever…

Then you’re clearly in the extract phase. And then you suddenly have the idea: “What if we gave people online storage, like dropbox, but with their phone contract. Then it’s - I think it’s understandable that they treat it as an extract problem too, even though it’s a completely different feature idea, and they should be exploring, right? So, maybe one thing here is to raise awareness… How can we raise awareness?

Antony: Well, I think where 3X is at the moment, or from Jeffry Moore’s “Crossing the Chasm” technology adoption curve, I think where 3X is at the moment is kind-of in the “innovators” phase. It’s not something that… It’s not wide-spread, it’s not something that everybody’s talking about.

Pioneers, Settlers, Town Plannsers and other ideas

David: Yeah, I think it’s like a handful of people world-wide who are really talking about it, right? Like you, Kent and a few others.

Antony: Yeah, just a handful of people are really “exploring” at the moment, with 3X. And there’s a lot to be done to develop it. I’ve run a few workshops with various people where we’ve kind-of looked at what kind of practices are applicaple in each of these phases? What kinds of people will thrive more or less in one phase or another?

There’s some work out there from Simon Wardley where he’s looked at… he talks about pioneers, settlers and town planners and he’s also now making the connection between 3X and some of his thinking, where different types of people will prefer working more or less in one space than the other.

So, there’s a lot of development to be done. I think really it’s very much the enthusiasts, like yourself and me and a few other people are really trying to delve into this. Like, we know there’s something there. We’re starting to be able to put it into words but we can’t show it yet.

And I think it’s a case of, you know, those of us who are getting value from understanding it are sharing with other people, getting feedback, and are allowing it to grow organically. And there will come a point where we can put this into a format that the early adopters and the early majority then might start taking interest in. And then we’ve got a whole new, different set of problems because we’ll be having to play by expand rules at that point in time.

David: Yeah! One quick question… Could the other thing happen too? That you’re staying in explore too long? Like you’re applying lean startup and MVPs and pivot all the time and you miss the point where you should expand? Or is this harder?

3X and Lean Startup

Antony: No, no, absolutely! I think it would be harder to miss in many regards, because, you know, people will be banging on your door for your product. At that point, they may be also complaining on Twitter because they’ve been trying to log in to your product for the last half an hour and they can’t because you’re servers can’t handle the capacity.

Those are going to be pretty clear signs that you’re likely to have an expand problem. And you should be operating in that space. I think, and again, these are phases and you transition, it’s not like…

David: You don’t flip a switch like that…

Antony: No! And the payoff also… What is the payoff to you will be different as well. So, for an early-stage startup, the payoff might just be investment. That might be the payoff they’re looking for, they might be looking to grow their user base so they have an asset that attracts investment but they may not know how that product’s going to make money yet.

But it doesn’t mean that they’re not in expand, because they’ve got lots and lots of users trying to use their product, because they know what the payoff is: The payoff is more investment.

And then, at some point, where it becomes less about investment and more about “how does this company make money in a self-sufficient way?”, then again you’re back into explore. You may be extracting from your investors at that point, but now you’re having to explore: “How do we make this product generate revenue?”, so you’re back at the beginning of the cycle again. “Maybe we try a little bit with advertising or promoted hash tags?” or whatever it is that makes the product valuable.

Then you’re exploring well, how can this product make money now for our investors. So, it is very, very fluid, very, very fractal. Any company, like you said can be in extract, you have several ideas in expand and hopefully many, many more ideas going through the explore cycle.

Investemen and Your Return on Invest

David: I think we don’t have much time left and I wanted to talk to you about investment. One thing I remember from a talk from Kent Beck - and I’m not sure if I remember it correctly - was that you optimize your return on investment differently in different phases.

Like, when you explore, you want, for every unit of money or time invested, you want to learn as much as possible and do as many experiments as possible. While in expand, you want to optimize for growth and for developing features as fast as possible. And in extract, you want to optimize for lowering the costs or increasing the payoff or the return on time invested.

Antony: Profitability in whatever is the thing that you’re interested in, not necessary financial profit. At that point, it really depends on what your payoff ist.

But yeah, I mean in expand it’s perhaps largely more about…

David: Not dying…

Antony: Yes, survival! Like, the examples that he’s given me is much more about survival.

I could think of a really good example: So, one of my clients is ITV, which is a UK television company. And several years ago, their online streaming service was something they started building up. Demand for it started increasing quite radically, the skillset and technology they had available to them at the time resulted in them building all the services that backed the streaming service in a particular technology which I won’t name.

And then they could throw more and more money at the infrastructure in order to cope with the continuous and rapid growth on the viewers trying to stream live TV shows and captured TV shows.

And then, what happened as the user-based has matured, they realized that infrastructure was costing a lot of money so they actually decided, well it could be wise for us to invest in a less expensive platform.

And then they moved over to…

Fractal Nature: Extract can Mean Exploring

David: OK, so now they were in extract?

Antony: Now they were in extract, and now they had the time and the space to think: Oh, this is very expensive! Before, it was like: Can we survive until tomorrow? And now it’s: Oh, we’ve got time to think about this now, becaus everything… we understand the problem, we understand the solution and… How can we make this actually a little bit less expensive.

So now they started looking into how they could replace all the services with different technology that can be in the cloud, running at much lower costs. So they went about the process of starting that, but again, it was unintentional at the time, but when we reflect on it, they also went through the explore, expand, extract with how they applied that. They explored different technologies, they tried a few different things, they found the one that was working the best, that expanded to support all of the services and then the old technology was gone.

And now they’re in the extract phase for the new platform that they have. And they’re equally applying this now to many, many other things including… Now conciously, this is being conciously applied by teams at ITV now, where they are saying, you know, are we exploring here in which case we need how can we try this expriment as quickly as possible? And then when they do land on an idea, the team that understands that idea kind-of carries it into the expansion phase. Until eventually, it becomes business-as-usual, therefore in extract.

3X and Agile Practices

David: Maybe just one more question, if you have another few minutes, I think I can see how agile would apply in the explore phase… Like, the agile techniques that we’re using, like quick experiments, small steps, and also throwing away things, a backlog that can live and grow and shrink, …

And I can also see how to use the same things in the extract phase… Like where you do small steps, but iterative and incremental development…

But, can we also link the agile techniques to the expand phase? Or did I get anything wrong in the first two?

Antony: I think there some aspects of… Of course, agile is an umbrella term for many, many different methodologies, made of many, many different practices. So, it’s hard to really say “does agile apply here or there?”

I think there are elements of agile, you know, methods and practices, and some will apply better in explore and not in extract. And some will apply better in expand and not in explore and so on.

So, I think it’s hard to say, again, there’s so many different methods and practices that we call agile.

I think it’s safe to say that understanding, for example, in explore, if the bottleneck for testing an idea, for example, is writing tests… For whatever reason, if in the context of the team that you have at that time, if they want to try an idea out and you want to try as many ideas as possible, and as long as the things that need to work are protected by tests, you want to try this idea, you shouldn’t need tests. You just need a way of limiting the impact of that thing on your users. So, if it doesn’t work very well, only a tiny, tiny percentage of your users ever saw it.

And if writing tests is slowing that team down, if that’s the bottle neck, then, if it was me, I’d be OK with not writing those tests. Because, if I’m gonna try a hundred different ideas, maybe even a thousand different ideas, and only one of them is going to survive, I’m not going to be worried about… And a feature idea takes me maybe a day to try, I’m not going to be worried about whether the quality of the code of that idea - as long as we retain the discipline that, if it is an idea that we’re going to keep, and it’s not one that we throw away, that we then…

In fact, expand, will push you to the point where you have to make sure that you keep things working and start to improve the quality - the internal quality of the code in order to cope with the rate of expansion.

It’s Only Technical Dept if You Keep It

So, one of the things I say is that, anything you do up to the point of where you’re transitioning from explore to expand, at that point, it’s not really technical dept. It’s only dept if you keep it. If you actually delete that code, then it’s fine.

David: It would be wasted money if you did the best thing possible and then throw it away because it’s not going to be successful, right?

Antony: Yeah!

Although it depends on the team. I have worked with people who… and to an extent I feel I’m in the category that… I find writing tests helps me with my thinking anyway, so I would not stop writing tests.

So, that’s why I’m saying… In the context of a team where maybe that part is slowing them down, as long as they’ve got protection around everything else that needs to work, you don’t necessarily need that with each of your experiments.

Exploring in Legacy Code

If you have optimized - If you have an existing product, let’s say in extract, and you want to be able to explore quickly, then one of the things that you may need to do, is optimize your product to allow you to explore quickly. So that you can drop in a completely isolated piece of code that adds new functionality, that if it’s not something you’re going to keep, you could just delete that code, and it doesn’t affect anything else.

This is the kind of thinking - it starts to make you think “Oh, well, yes, it’s optimized to keep our costs down in operating it… But what if we want to try new feature ideas on this platform? Can we do that quickly?” And then you start exploring ways of changing the internal structure of your code so that you can make it easy to start doing experiments. At which point, then you can start doing experiments more quickly.

David: Yeah, I think I know a couple of big systems that are not optimized for that at all…

Antony: Absolutely, and that’s the case, like if you’ve got a lot of legacy code already, then exploring in that space is really, really difficult. One of the first things you might want to explore is how to be able to experiment fast and safe.

And this is again takes us back to the fractal idea. So, maybe your feature ideas… Or you have a set of ideas you want to try, but it’s in a product that’s already in extract, so, you know, you’ve got to change how you go about doing those… How you explore in an extract product.

Exploring While Protecting Your Brand

Whereas, if you have no product, and you’re just trying to explore a product space, then you might be in an extract company, now you’ve got to think about your brand. Like, “We’ve got to protect our brand, and if we do something… If we do a product idea that fails, is it going to have a bad effect on our brand? How can we manage that? How can we make it quick and easy for us to experiment with ideas?”

And as an example of that, one of the bits of advice that I’ve given in the past, is: Consider creating a different brand that’s intentionally there for the enthusiasts.

And example of how that could be applied… I point people at the history behind and O2 and giffgaff. O2, as a mobile operator, was significantly involved in helping giffgaff get started. And giffgaff is a network that is literally run by the users of the network. They provide support to each other - If someone doesn’t know how to do something on their phone, they ask a question and someone in the community answers it.

So, giffgaff then becomes a place where, if they wanted to - I’m not saying they do this, I’m just saying if they wanted to - O2 can experiment with new, whacky features or ideas. Like, you know, cloud storage with your network plan and things like that. They can experiment with ideas with giffgaff, see if they work, and then if they work well, learn from that and that sell it to the O2 customers later.

You know, this is the kind of thing that you could do with that brand separation. I think they’re much more separate now, I don’t think there’s a close conneciton between them, but I just use that as an example to say, you know, you can create a brand that is actually intentionally targetting the enthusiasts. So, your extract brand can be insulated from the things that you want to experiment with that might not work and might cause bad PR. So, you put them in a different place wher actually trying and failing is welcomed by the customer base.

David: Yeah, Dynatrace did the same thing when they moved from a traditional product to a cloud product.

They created the cloud product under a completely different name, it was called “Ruxit” back then. And when the cloud product became successful, when the… I think in or after the expand phase, they re-named it back to “Dynatrace”. And now they had Dynatrace cloud and the other product. And now I think it’s only Dynatrace.

Antony: Both of those stories give us an example of how fractal it is and how the interplay between the zoom level above the one your at the moment is a part of the context of the zoom level that you’re currently applying this thinking to.

It can even be at the industry-levels: There are whole new industries that we don’t even know are going to exist yet. The people are experimenting with cool and interesting ideas. That maybe only inthusiasts know about at the moment. And those ideas will eventually… They’ll figure out how to make that both a success from their customers point of view and a payoff for them.

And then they’ll be expanding, they’ll be extracting. And then they’ll become extract brands and they’ll have to change how they go about things. You know, it’s a continuous cycle.

So, a whole new industry could be born from an idea that someone is experimenting with today. Then you’ve got an industry that might be in extract in the future that, within which there are several different companies.

Any one of these three stages within each of these companies. They can have products that can be in one of each of these stages, and then the product’s features can be in each of these stages. So, it’s fractal.


David: So, yes, thank you Antony! That was very interesting! I think the important things we talked about in this video was that…

3X or explore, expand, extract gives us a way to tell tell us when the rules have changed. Because you have to use different techniques and behave differntly when a product or idea or feature or whatever you do is in one of those phases.

And also that you have to treat investment and what you expect from this investment differently.

And I think the second important thing we talked about today was how this whole idea is fractal. That you might work on something right now that’s in the extract phase but you’re also working on the next thing. And then on things further in the future which will be in different phases.

Or that you have a product or a company that’s in the extract phase but it works on ideas that are in expand or explore.

Did I miss anything?

Antony: I think that’s about everything. And, you know, the main thing is I think to say about also 3X itself is… It’s kind-of still very much in explore. But maybe videos like this might start pulling it in a direction of turning into an expand idea.

David: Yes, I hope so! So, Antony, where can people find you on the internet if they want to learn more?

Antony: Yeah, if they just google me, that would be a great place to start. Antony, without an “h”, Antony Marcano. Find me on Twitter by my full name, on Twitter. I’m also on LinkedIn but I don’t look at that very often.

Find me via email from my consultancy’s web site, Drop us an email there. And check out the RiverGlide blog which is and there you’ll find articles about 3X and lots of other things.

David: And I will also add some links to this video and my name is David Tanzer, I am @dtanzer on Twitter and on GitHub and you can also find this video and more stuff about me and my consulting business on

So, subscribe to my YouTube channel and see you next time! Thanks, Antony, for being here!

Antony: My pleasure! Bye!

David: Bye!

Read / watch all parts of “#3Xthinking” here: