The Three Machines

Lately, I’ve been stewing over increased complexity being generated by companies around their organization approaches. While this activity varies by stage, in many cases the leadership team expands to a large (greater than six) number of people, there become two executive teams (the C-Team and the E-Team), the CEO gets sucked into endless distractions and working “in the company” rather than “on the company”, and I could go on with a 1,000 word rant on the challenges and complexity.

Recently, I saw a structure rolled out by a CEO at a company I’m an investor in that made me pause because of its simplicity and brilliance. I didn’t like the labels the CEO used, but I loved the intellectual approach.

It coincidentally had three categories. Three is my favorite number and has been since I was three years old. While I can carry more than three things around in my head at a time, when there are only three attached to a specific thing I find that it’s second (third?) nature to me and requires no additional processing power to remember and organize my thoughts around three things.

If you recall my post on Three Magic Numbers, this will immediately make sense to you. Or if you’ve ever heard my story about struggling with clinical OCD in my 20s where the number three was one of my key anchor points, you’ll have empathy for my relationship with the number three.

I abstracted the structure I saw from the CEO recently into what I’m currently calling “The Three Machines.” While this can apply to any size company, it’s particularly relevant to a company that is in the market with its first product, or a company that is now scaling rapidly with a set of products.

The three machines are: (1) the Product machine, (2) the Customer machine, and (3) the Company machine.

If you step back and think about all of the activities of a company in the phases I described above, they fit in one of these three machines. However, most leadership teams don’t mirror this. Instead, in a lot of cases, there is a traditional leadership team structure that has a CEO and a bunch of VPs (VP Engineering, VP Product, VP Finance, VP H&R, VP Sales, VP Marketing, VP Customer Care, VP Operations, …) which are often title inflated with CxO titles (CTO, CFO, Chief People Office, CMO, COO, CRO, …) or artificial demarcations between VPs and SVPs (and EVPs.)

Regardless of title structure, the CEO has a span of control that gets wider as the company scales, often with more people being added into the hierarchy at the VP or CxO level. As this continues, and CxOs are added, you end up with the C-team and the E-Team (which includes the non-CxOs). The focus of each person is on a specific functional area (finance, marketing, sales) and traditionally scoped.

In a few cases, big organizational experiments ensue, often after the organization dynamics hit a wall. Holacracy, which is still bouncing around, was a relatively recent trendy one. I disliked holacracy from the first time I heard about it and resisted even experimenting with is, preferring to watch what happened when others tried it. In 2013, Nick Wingfield wrote an often-citied article in the NY Times titled Microsoft Overhauls, the Apple Way that is liked to a now famous graphic of different org charts for Amazon, Google, Facebook, Microsoft, Oracle, and Apple.

I’ve wrestled with hundreds of conversations around this in the past few years. I never have felt satisfied, or even particularly comfortable, until I landed on the three machines recently.

My current hypothesis is that if you are a CEO, focus your organization on the three machines. Product, Customer, and Company. Then, have a direct report own one of them. If you have a sub-scale leadership team (e.g. you are three founders and four other employees), as CEO you can own one, but not more than one. As you get bigger (probably greater than 20 employees), hopefully how you have enough leadership to have one person own each, but recognize that if someone is being ineffective as a leader of one of the machines, you will have to replace them in that role (either by firing them or re-assigning them).

Let’s assume you have enough of a leadership team that you have a key leader who can own each one. Organize the company leadership around each machine. The titles don’t matter, but the hierarchy does. Naturally, you will have a product or engineering leader for Product, you will have a sales, marketing, or operations leader for Customer, and you will have a finance or admin leader for Company.

But, this does not mean that your VP Engineering is your VP Product and Engineering. That rarely works – you want to separate these two functions. But your VP Product, or your VP Engineering, or your CTO could be responsible for the Product machine, with the other VP functions reporting to her. You probably also don’t want to merge your VP Sales and VP Marketing and VP Customer Care function into a VP of Sales, Marketing, and Customer Care. But, if you have a Chief Revenue Officer, you may have done this. While that can work, recognize that it works if the CRO realizes he is in charge of the entire Customer machine.

I’m still in the first few weeks of really building a theory around this so there’s a lot of sloppy thinking on my part so far. For example, I don’t think this necessarily means that the CEO only has three direct reports. But it might. Or, in some cases, at certain scales it might. I haven’t focused on what it means in terms of the overall hierarchy. I haven’t really thought about how multiple different product lines come into play. I don’t know if there needs to be dramatic retitling at the top.

I do, however, have several companies that are very clearly focused on these three machines. Yet, they are at different scale points and have different formal hierarchies. Over the next few months, I’m going to use this lens across every company I’m an investor in as I poke and prod at how it might, can, and should work. And, determine if it’s a valid hypothesis.

Feedback of any type is welcome.


Also published on Medium.

  • Josh Scott

    I’m a big fan of the simplicity of this structure – particularly the way it helps align org structure to the key activities of a company… from experience with a similar structure, I think the interesting rub is in what falls into “Product” vs. what falls into “Customer” likely differs greatly by company and the way you choose to handle this alignment could have meaningful implications both for internal cultural as well as how the offering/customer experience evolves. Specifically, in many cases, a company’s site/app is both the “Product” (or a consequential portion of the product offering) and a “Customer” engagement / marketing channel which needs to be in harmony with the rest of the “Customer Machine”. Would love thoughts on this specifically…

    • All three are linked in many places. @lorangb:disqus long comment above teases toward it – the leaders of the three machines meet with a regular cadence to work across them.

      I think the website/app will almost always be both product and customer. The key is not to view one as owning the website/app, but to have the website/app be an “object” of the business that both act on (I expect that will appeal to anyone who knows C++ …

  • I assume you do not plan to impose this on your companies. Many may feel pressure to adopt it anyway. Are you sure?

    • I don’t impose anything on anyone. And it’s well known by the companies I’m involved in that I like to think out loud through my writing.

  • Rebecca

    I’m feeling similar to Josh below, while I love the simplicity, I have a harder time breaking out some of the functionality and imagine that different choices wou,d have large cascading consequences.

    My biggest question may not be a big one for software, but certainly is for my business around functional support and traditional “operations” — seems to me that we would end up building multiple functional support teams (e.g. Design, analytics) or have messy, you report to multiple teams, challenges. Which, to be fair, exist today, but at a VP level that is controlled by the (5 person) exec team.

  • Sam

    I don’t know. I wonder if focusing on org structure as a lever to solve problems masks what are in reality bad hiring decisions.

    As context, I joined a startup many years ago as the first exec hire after the two co-founders to write our business plan and raise our first serious money. We raised $6M, in hindsight probably 4x what we should have raised, but we figured take the money while it’s there.

    We then hired an exec team with ever-inflated titles. We made lots of mistakes and ended up pretty dysfunctional, and I can totally relate to the C-team vs E-team distractions.

    As I reflect on whether a clear Product, Customer, and Company Machine org structure would’ve helped us, I don’t think so. What fucked us were the multiple bad hiring decisions we made into those executive seats. People who didn’t have the right skill set and mentality for what we needed at that early stage of building a company.

    If you have the wrong people on the bus, it doesn’t matter what seat they’re in.

    • rick gregory

      Those are separate issues, though. No structure will save you from bad decisions. Conversely, good hires could be even more effective in the right setting (and you might not need as many exec-level hires in this structure). Org structure should bring clarity about lines of responsibility… it can’t actually enforce top notch execution

      • Sam

        I agree with you. Separate issues. And I wasn’t very clear above.

        My point was that CEOs should be careful that they’re not jumping to structure as the answer to a question that is fundamentally different in nature. Structure can be seductive for analytical types. It’s concrete. It’s Xs and Os. You can see it in an org chart.

        Much more difficult to think about culture, soft skills, processes like budgeting, feedback and communications, prioritization, and the operating rhythm of the executive team.

        There’s a fantastic org design tool, the 7-S framework that was developed by McKinsey in 1980 and is still highly relevant today. Recommend to any CEO that they tick through the seven categories (one is structure but there are six others) to be thoughtful about the kind of org challenge they are dealing with. Then they can figure out which lever to pull, structure being but one.

        7-S article here for those interested in digging deeper:
        http://bit.ly/2iTSCEN

    • Like my comment to @disqus_wKLeb9VknK:disqus about operating conditions, I’d separate the constructs.

      In my experience, If you have a great structure and terrible people, you will fail. If you have great people and a terrible structure, you will muddle through and likely evolve to a better structure.

  • I cannot recollect where I read a variation of the number 3 – In a company, you need someone to make it, someone to sell it and someone to collect the money. The first two align well with the structure above, the third point here is a gentle reminder about importance of cash flow.

    Continuing with the utility of number three ( 3 is the minimum number of points required to make a plane) -:

    1.) The highest friction typically tends to be between the “Product” and the “Customer” sub-organizations. I wonder if actually these can be merged and I would prefer to call it “Customer”. This is the revenue engine
    2.) In this structure, another role would be of “Company” but the function of this much leaner unit would be to think long-term and not let the current “Customer” drive the entire company. This is the agency establishing the First-Principles in a company and the lodestar
    3.) Finally, I would use a third element similar to the “Company” in your example (and am half-tempted to call it “Grease”!) which enables the smooth functioning of internal non-revenue based operations.

    Oh and a 3-legged stool is more stable than a 4-legged one….

    • We posted this at the exact same time!!!! How funny.

      • Hence proving that great minds not just think alike but also at the same time 😉

    • It must be because three is just such a beautiful number … Good additional analogies – thx.

  • I have always said it is a three legged stool. If you notice a three legged stool never gets tippy even on an uneven floor, as long as the legs are equal. Somebody has to take care of the customers, product, and company. It’s my second rule after finding and keeping the best people. Orchestrate between those three things.

    If the customer area is too long, you promise anything, burn out product, and not build a good company.

    If product is too strong you don’t focus on customers, and then not generate revenue and burn up company

    If company (which really means finance) is too strong you focus inwards and not on customers or product which ruins the company.

    • tgodin

      In my environment (professional services), I tend to think of the three legs of the stool as “clients, employees, and ‘the business.” Conversations about our services (analogous to ‘product’ I suppose) are always in the context of what clients want or what we think they will find valuable. The three-legged stool analogy always works for me.

      • As someone who ran a consulting business for seven years (my first company), I agree. We tried many times to create a product, or productize our consulting, but ultimately we were focused on what our clients needed / asked us for.

    • Good one @philipsugar:disqus. You can see from @lorangb:disqus comment above that he uses this metaphor also.

    • DaveJ

      Actually the legs don’t have to be equal nor the floor flat; it’s just that the seat won’t be level. This means that the CEO, sitting on the stool, will slide toward the shortest (weakest) leg. How about that for metaphor extension?

  • sdso234

    One thing I don’t like about that approach is that it pushes down the org some of the hard trade offs.

    For example – if your customer machine has different pov’s on growth (premium price vs. sales volume), this debate between sales and marketing might best be elevated to the CEO.

    Of course a strong “machine owner” /CRO could adjudicate, but my sense is that a CEO should be highly engaged in these types of debates/disagreements…

  • I”ll be interested to see how your thinking evolves on this.
    IMHO the critical thing for a CEO to recognize is that as organizations evolve so much their organizational form. An organizational structure must be aligned with business needs and all too often it isn’t. It is my suspicion that one of the trickiest balances to get right is that between stability and necessary organization change/revolution. I very much doubt there will ever be a ‘solution’ to organizational structure.

    • Yup. And, as the company grows, you often scale out of your individual leaders. The leader who was great when the company was 20 people might be a disaster when the company is 100 people.

  • Sean Mulvey

    I have been part of successful start-ups and well known public companies. In all these cases I have seen these companies try to define role responsibility and it typically gets murky as people want to over-extend their roles. To this day, I think Bill Belichick of the New England Patriots has the best strategy – “do your job”. I interpret this as don’t worry about titles, perceived roles, experience, etc and if you focus on your specific role the team will win.

    • I think “Do your job” is a operating condition, not a structure. In a holacracy, I think that operating condition would fail miserably. So, I’d assert the the structure matters, but in the absence of an operating condition like “do your job” it’s also easy to imagine that the structure doesn’t matter at all.

  • As the subject of this post, I figured I’d share my thoughts directly 🙂

    1. At FullContact, we call it “GTM” (Go To Market), “EPD” (Engineering, Product & Design) and “G&A” – (General & Administrative). However, I do like the simplicity of the words you used!

    2. In this comment section, folks are talking about a three-legged stool. This is exactly how Rushton McGarr (our CFO) talked about it. I think of this as a three-legged stool of three-legged stools.

    3. Each of these three groups consists of three people who *partner.*
    – Customer Machine – VP Sales, VP Marketing and CEO
    – Product Machine – VP Product, CTO (who runs Engineering & Data), and CEO
    – Company Machine – CFO, Chief of Staff (who runs HR), and CEO

    4. These three groups meet weekly for one hour on Fridays. In these meetings, CFO also attends. A CFO – particularly an operational one like ours – provides a healthy counter-balance to the CEO and and the company benefits tremendously from CFO’s awareness of resource/capital allocation discussions and strategic decisions. The small nature of the group allows for a very focused and productive discussion.

    5. On Mondays, the “Exec” group meets for two hours to decide or discuss anything that has bubbled up from the sub-groups.

    6. I believe the meeting cadence gives us the opportunity to go deep and focus in a particular area if I needed without wasting someone else’s time. For example, if I want to go into the details of some second order Product metrics or features, I’m not boring the Sales or HR folks.

    7. We don’t know if weekly is the right cadence yet, but given the intensity and high pace we’re coming out of the gates in 2017, I decided we needed to meet weekly until we felt it was overkill or unnecessary and dial it back.

    8. We’ve organized meetings, OKRs, metrics, capital allocation and even social outings around this construct.

    9. We currently don’t have one person ‘100% in charge’ of every sub-group, though we do assign DRIs for discreet tasks – like preparing Agendas, Strategies, or OKRs to ensure accountability. Instead, we rely on individuals partnering. I believe the structure is working because we have a strong partnership mentality between the individuals on the team:
    – Our VP Sales literally uses the word ‘partner’ 100+ times a day and he and our VP Marketing have shared goals and incentives. I’ve insisted that they are 100% aligned (as partners should be) on outcomes.
    – Our CTO and VP Product are terrific partners and work at an extremely high intellectual baud rate and share many of the same values on a number of levels.
    – Our Chief of Staff and CFO have a special partnership bond and work extremely well together as a pair.
    – Each of these sub-groups genuinely enjoy each other’s company and we use words like “Love” routinely around here. We also try to be emotionally in-tune with each other and check our egos at the door.

    10. There is some quirkiness at present. We have Professional Services reporting up to the “Customer Machine” and Customer Success reporting into the “Product Machine” – largely because each of those departments service a different type of customer segment. I think over time, we’ll figure out how to align or merge these functions better – but for now, we are simply taking a partnership mentality between these departments.

    11. For this to work at FullContact, the leaders in each group need to be senior enough with the experience able to handle multiple customer segments and product lines and not be overwhelmed with the complexity that introduces.

    On the topic of partnerships, I’ve found the partnership mentality is a theme in my life, and on reflection, I have several partnerships that are always present in my life:

    – Partnership with my wife
    – Partnership with my co-founder and other board members
    – Partnerships with EPD, G&A & GTM
    – Partnership in my venture fund, v1.vc

    All of the partnerships have different relationship dynamics, and balancing partnerships requires thoughtful consideration about what’s best for the partnership, and avoiding self-dealing at all costs.

    I learned a long time ago, that a partnership works when you treat it like a fiduciary relationship. In other words, always make sure you put your partner’s interests ahead of your own! If every partner does that, you’re on the right track.

    • Bart – thanks for the comprehensive additional information. I’m thinking hard about the three partner idea for leadership of each machine (e.g. VP Product + VP Engineering + CEO). I also like the CEO as a partner in each machine.

    • Hi Bart, thanks for taking the time to write in and share.

      A question (and counter point?): I’ve tended to find that every company has some sort of hierarchy between and within the make functions (product, engineering) and sell functions (sales, marketing). Or, in your terms, the product and customer machines.

      Some of this rises due to politics but a lot of it rises due to culture and the company’s core capability. Apple’s design team’s influence is high, same as Google’s engineering team, etc.

      Curious to hear what you think about that.

      • Thanks Rohan, some type of power dynamic can creep in and a company’s historical DNA comes into play. Sometimes it’s because of the personality of the leaders in each area. Often, it’s a result of what the CEO prioritizes (sometimes intentionally, sometimes unintentionally). So, to me, it’s important to simply be mindful of this, and stress a partnership at all times to try to achieve balance in the Force.

        • Makes sense.. even if I can see some very obvious challenges. But, I guess mindfulness is the only way through these sorts of problems.

          Thanks for taking the time to share.

  • John McDermott

    Very similar to a concept I used when working in a small startup. I also wanted three principles in managing the team, so called it ABC
    A= architecture. Still in development phase, so everything to do with designing the product and getting into production
    B=business. Managing the team and owner
    C=customers. Everything about sales and marketing, revenue focus.

    Lot of similarities with your model.

  • The three machines make a lot of sense. I’m just wondering about the interplay between engineering and product, sales and marketing, etc…

    No answer or view, still just thinking about it.

    • I think @lorangb:disqus comment below does a really nice job of expanding this.

  • One of my favorites. It’s the intro to my conference call number.

  • This seem like one of those ideas that is at times happening in practice but lacking a name, e.g. Lean Startup. After reading Bart’s description, it seems his structure is something of the anti-Holocracy, purposefully assigning the key roles coordinating Customers, Product, and Operations to committees rather than individuals.

    Between your post and Bart’s comment, you both agree on a division within the hierarchy across the same three themes, and for most startup software companies there is nothing else of consequence. The messiness from the bigger, public companies diagrammed comes from multiple products.

    The more interesting question here is whether it is better to have a single person responsible for each division or whether that is better done by a team. I’ve seen both fail and both succeed. You’ve seen more instances of more operating startups than most anyone on the planet. What are the most common structures and which are most correlated with success?

    • I don’t know if a person or team is better. I’m leaning toward Bart’s team approach.

  • An email comment from a friend – posting here to keep track of it as I think through this more.

    Read your article on organizational approaches titled The Three Machines. I spend a lot of time in this area coaching CEOs on how to think about their job and the organization. Once a company has reached scale, typically between 20-50 employees depending upon the business I think the org structure should have 6 major components. The number 6 interestingly comes from two different groupings of three that are important. The typical definition of a business sounds something like, an organization that produces and delivers a product or service while balancing the needs of shareholders, employees and customers. For the typical business producing and delivering a product requires three key functions: product, marketing and sales. Once the product is produced and sold then the CEO’s main role is to balance the needs of shareholders, employees and customers in a sustainable manner. Voila another grouping of 3. If you lay these two groupings over each other you end up with 6 unique roles for the typical business. This is a place to start. There are some businesses that have unusual requirements, for example a very heavy legal requirement, that might cause the elevation of another executive. I am happy to go into more detail or teach to your companies if you have any interest.

  • J.J. Thompson

    In our world at Rook, its Build, Acquire, Operate.
    Couldn’t agree more.

    By the way, you should consider a visit out to the Indianapolis area soon. People at Techpoint and Launch Fishers would be good for you to meet with. Kindred spirits separated by the great plains!

    • David Ellenbogen

      J.J. – Could you please expand on your three – Build, Acquire, Operate ?
      – David

      • J.J. Thompson

        In our case

        Build: r&d, product dev
        Acquire: marketing + sales
        Operate: MSS, Advisory, Finance, HR, IT, etc.

        • David Ellenbogen

          Thank you.

  • cavepainting

    Excellent post! If we see a machine as something that produces a clear outcome, then everything that goes into the machine –including VPs and their teams –need to be harmonized for the outcome produced by the machine.

    For example, at the early stage for a SaaS company, the only two machines are capital machine and product machine. The former will need the CEO to lead it, while the latter may need all founders and most employees to work together and will include customer development and early customer-ship. Better to call it a product machine than a customer machine because the primary goal at this stage is to produce a product that works well.

    I guess my point is that machines (how many and what they do) will vary by stage and vertical. Machine is a more visceral word for what used to be called a value stream. I do love the idea that people can be organized more loosely by value streams or machines than in a rigid structure. However, to make it work better than traditional structures needs great communication and clarity from the CEO.

  • cavepainting

    Brad, on another note, the social media vertical bar that appears to the right side of the comments is too broad and spills over to the content on the page. I find that I have to guess the last word on every line on a comment. (I am using chrome).

    • Thanks for the feedback. Getting rid of the bar now.

  • Every team has their own rhythm, structure and style of operation. I have seen everything in between and the kitchen sink that actually works. I am biased to let the founding team figure out the structure that works, the thoughts process and the ultimate structure usually gets owned by the team. I like the thinking Brad, typically that is what companies need to do, Build something your customers want, make sure you are delivering it and hopefully you are doing it in a place, structure that does not kill the team. That being said, not sure I like the machine analogy. I know efficiency is the what is meant by the use of the word Machine, end of the day we are all people, I think focusing on feeling and looking at how tribes organise themselves that has stood the test of time is also a good model. Just jotting down my thoughts but I like the simplicity of this structural view.

  • Michael Shaler

    I prefer the model of Three Activities: Build Stuff, Sell Stuff, and Count Stuff. Not all that different than the machines, but it simplifies the various paths to accountability.

  • Glenn Neal

    This looks like a very interesting discussion (both text and comments). Excuse this very basic question but you wrote in 1st paragraph “… there become two executive teams (the C-Team and the E-Team)”. I’m not sure what you mean by C-Team and E-Team. Can you give me a simple definition of these terms or point me in a direction? Doing a search for ‘E-Team’ in blog brings up every article with ‘E’ in it – obviously that direction didn’t get me where I wanted to go. Thanks! FYI – all answers for me, especially in business, tend to be organic in some way. But I always like adding new info into my mind.

    • C-team is everyone with a “C” title – CEO, CFO, CTO, CRO, COO, …

      E-team is the extended executive team, which includes the C-team and anyone with a VP title (basically – the next level down direct reports to the C’s)

  • Glenn Neal

    With that bit of information and all other comments, much more for me to roll around in my head.Thank you.

  • Gordon Renouf

    Product Customers Company is a close match for this founder’s description of what he focuses on Product People Customers https://www.cbinsights.com/blog/exit-strategy/

  • Great post. I remember an early Steve Jobs internal presentation where he talked centering around 3 things… 1. Great Products. 2. Great Marketing and 3. Great Distribution. It’s a good number 🙂

  • Ben Deda

    Good article Brad. Actually a pretty old military acxiom. A leader is best with three direct reports. Three members under a fire team leader, 3 fire teams in a squad, three squads in a platoon, three line platoons in a company, three line companies in a battalion. Key is how you keep it team if teams vs hierarchical.

  • Chris V. Nicholson

    Our startup is organized in three teams, and their size varies. Product and customer teams are about the same size, and closely intertwined. We have a core engineering team, field engineers in touch with customers, and BD people bringing leads in. In our terminology, operations encompasses finance, admin and HR. Ops is smallest. The size of these teams more or less correlates with their age. We started our trying to make something we thought could sell. Once it was made, we brought on some people to help sell it. And now we have a small group to manage the complexity of employing all those people and tracking their activity. We think of it this way. 16 people total. Raised $3M seed last year.

  • Michael Rodov

    At Market Realist, we have a structure which is close (4 major areas instead of 3) and it works very well. Took years to evolve into a simplified org. Requires company-building around key talent.

  • This reminds me of concepts in two books.

    In The Goal, the author describes a business model in terms of inputs and outputs. The two inputs are customer money and investor money. The two outputs are profit money and waste money. At any time there is a certain amount of money in the business and money is flowing in and out of the 4 channels at various rates.

    In Four Steps to the Epiphany, the author describes a startup approach focused on product development and company development to the detriment of customer development. The author makes a case that using that approach is often a bad bet. He advocates an approach that focuses on customer development as the primary activity to emphasize over the other two. The case the author makes seems to fit well with businesses with relatively more market risk than technology risk.

    • Both excellent and important books.

  • Melissa Cooper

    I’m a fan of the framework.
    At risk of sidelining the conversation… I wonder about the split. Placing product and engineering in one stream and “away” from the customer implies the role of the product manager is predominately managing the build. Or, at least it raises the risk that their focus will be drawn in that direction (as it so often is). To me, the PM needs to keep in daily contact with the customer in order to focus on product evolution. It would be interesting to place what we make and what we sell within the same machine.
    Perhaps we’re too wed to placing product and engineering in the same stream? I’m still ruminating too. I’d love to hear what people think.
    For each organisation, it would come down to their mission and values – I’m sure.

  • sigmaalgebra

    Let’s just review the basics: (1) You have a problem in management, that is, behavior of groups of people, that is, sociology when done as a science (sometimes it is so done, including with a lot of statistics; e.g,, my wife’s Ph.D. at Johns Hopkins was of that kind). (2) You want a solid answer, that is, almost necessarily a scientific answer. Yes, it is possible to do science in sociology. Don’t expect anything like Newton’s second law, but can still do science according to any good definitions of science. (3) In science, usually don’t prove anything but make a hypothesis, test it, and maybe reject it of fail to reject it. If fail to reject with a large, seemingly wide collection of tests, then entertain that can accept the hypothesis. (4) For good results in science, generally the first important step is good problem selection, in particular, to select a hypothesis that can fairly easily test and soon accept. (5) Usually the work of science is to identify variables and define numerical measures for them, how to get the numerical values, and if the measures have validity and reliability. In common terms of statistics, validity is much the same as unbiased, that is, if repeat the measurement many times and average will converge on the actual true value and reliability is essentially variance (or, if you wish, standard deviation) of the measure, that is, intuitively, accuracy. E.g., measuring the diameter of a high quality ball bearing has low variance and high reliability; measuring the diameter of a splat of Jello has high variance and low reliability. You want measures that are unbiased (high on validity) and low on variance (high on reliability). (6) When with the hypothesis, the measures, and the data, test the hypothesis.

    If you have done well on (4) problem selection, then you are nicely ahead of the pack!

    General warning: My brother’s Ph.D was in political science and his Master’s in psychology (my education was not much like that of my wife and brother (my Ph.D. was in applied math, stochastic optimal control with the associated problem selection and formulation, applied math, measurable selection, some algorithms to speed up the work, and some computing), and his summary remark about science about people was

    One can do solid science on the behavior of rats, but the results don’t apply at all well to humans.

    Can do science on humans, and can address problems of two kinds: (A) Problems were can do solid science but where the questions are so simple they are not very interesting for real people and (B) problems that are interesting for people but where doing solid science is too difficult.

    So, with my brother’s wisdom from experience, you should expect that getting solid science out of how to build an organization chart for a company may be an interesting question but where doing solid science is too difficult.

    I’d say that, might be able to get some results that are both solid and interesting if take (4) problem selection seriously, think a lot, use a lot of insight and intuition, and have some bright, clever ideas and some good luck. Don’t rush to settle for less than both solid and interesting (powerful, valuable); instead, be brighter than the nearly all the social science researchers. I.e., be highly exceptional, e.g., as the NBA champions, as in the NBA MVP, the Super Bowl champions and the MVP there, as in the startups you need to get the returns your LPs want :-)!

    For a non-scientific answer, I’d suggest a lot of good, intuitive insight. There I’d use some information that you have that the science would, in the interest of simplicity, ignore. E.g., not all startup organizations should fit the same ideas.

    I’d look for a CEO who, one BoD meeting after another, seemed to understand his company really well and had ideas that quickly seemed to work well. That is, look for a leader with some good insight and ability, a good CEO. Here the CEO would likely be exploiting information specific to his organization that simple, broad science results would ignore.

    Then, to try to return to science, look at, say, the US military and the careers of some successful military leaders, e.g., read the autobiography of Omar Bradley and how, step by step in his career, he was dropped into organizational messes and quickly came out looking good. Do such for a lot of military leaders (the US military has some relatively stable and broadly applied structure that makes finding good patterns easier) and then draw some general lessons as hypotheses. E.g., it was claimed that Patton did well in training and leadership after he was put in charge in North Africa. Well, okay, the movie didn’t say just what he did, so try to find out. Look at what Schwarzkopf did running Gulf War I. E.g., supposedly the USAF had 24 x 7 200 four engine tanker aircraft flying over Saudi Arabia just south of Iraq so that the other planes could refuel. That and the rest of Gulf War I was all put together and made to work right away. So, look into just how the heck that was done. Maybe then try to formulate some promising hypotheses for testing. Look, say, at how Trump organized and led his team through the primaries and then the general election and now the lead up to the inauguration; then see if can see more about how he organizes the White House and the top appointees, handles publicity, the press, Congress, projects like the wall and replacing ObamaCare — should be some good leadership lessons in there somewhere. Maybe then back to some solid social science!

  • JohnLBevan

    Are there any organisation structures focussed around the bottom of the pyramid rather than the top? i.e. 3 machines is primarily focussed on 4 people in the company; what about the rest of the company; what about the majority. Tweaking the top will have some impact, but I best that changing the titles/roles of 3 people has far less impact than finding a structure that facilitates the remaining 99% of your workforce to better get their jobs done & improve communication of relevant information between stakeholders.

    Maybe (in my opinion likely) that’s not going to be one structure for the entire company; but better change organisation method by personality; i.e. if you’ve got people who think the right way and get stuff done, give them more freedom to do that (more freedom typically means an “HR manager”, but not a “Boss”; i.e. they pick up work from a pool (or think up what work’s needed proactively) and just get on with it. Others may need more direction, or may be in roles where freedom is more restricted (e.g. people in a front line role often have to work set shifts to ensure there’s always cover, and have to follow run books for a number of tasks rather than being free to determine their own actions). Those people would need be more likely to need a “Boss” manager since they’re less likely motivated by the work, or need tasks to be assigned and more closely monitored to ensure their work fits in with the rest of the team’s, without requiring them to do that communication/coordination.

  • Sean Jackson

    Brad, the challenge with this framework may be scalability. As an alternative consideration, it may help to organize the company around the primary stakeholders in the business; Customer, Investors, and Employees.

    These three groups represent the primary constituents of any business and while each group may change in importance over time, they will persist regardless of scale.

    And of course, each group could be comprised of three primary focus areas within that group; with cross-over collaboration so that each constituency is represented.

    Further, if we look at the framework from a service leader perspective, we can then define whom each group serves and adjust the roles within each group as the company evolves without sacrificing the operational reporting.

    The Customer group is focused on making & fulfilling the promises the company makes to its customers.
    The Investor group is focused on fulfilling the obligations to the investors, debt holders and equity holders.
    The Employee group is focused on the environmental aspects of running the company.

    So while the mandates for each group remain consistent, the roles and importance of each group will grow and vary over time. As the company grows, the Employee group may grow in importance to address not only compensation, but overall business processes that improve the operation of the business.

    The Investor group may add on additional legal and compliance matters that affect the financial operation of the company.

    And certainly the Customer group will morph as the company evolves its product set, market focus, etc.

    But the framework would hold at scale.

    Just an alternative framework for consideration.

    • It’s good, but I challenge the investors as an equivalent level as customers or employees. I know some investors that approach things that way, but I think if management is optimizing for investors, they’ll make bad decisions that will ultimately lead to lower returns for investors!

      • Sean Jackson

        Investors in this framework is the catch all for parties that have a financial interest in the company, which could include debt holders. Further, each of the three groups will, over time, have their own competing priorities and interest; which is to be expected.

        The CEO in this case manages each of the three groups and sets overall priorities as the company evolves. Certainly in an exit event, all three groups will be impacted but the Investor (financial) group will have a higher priority than say the Customer group (though not always).

        Also, as the company enters into Customer agreements, the contractual terms, etc will also impact the overall financial group since the terms and conditions of any agreements will have a financial impact.

        Basically, each group is represented in the operation of the company (hence their equivalent) but their priority of interest will vary and the role of the CEO (or senior leadership) is to allocate the resources as needed; but at least their interest are represented.

        So maybe a the alternative to the Investor group is really the Financial group.

        Thank you for the article and I appreciate the opportunity to contribute.

  • Dave Wright

    We structured things in 3 similar buckets at SolidFire.
    I had three main groups and three direct reports:
    -Engineering (Chief Product Officer): Builds the product (product, engineering)
    -Sales & Marketing (Chief Sales Officer): Brings the product to market (marketing, sales, channel)
    -Operations (CFO): Runs the backend for the business (finance, legal, HR, facilities, supply chain, IT, support)

    The nice thing is that each group had a clear deliverable:
    -Engineering has a roadmap
    -Sales has a number
    -Operations has all the infrastructure to support the roadmap and the number

    The only real difference compared to what you describe may be where customer support landed. If you define the sales-oriented group as owning the “customer” you would probably put support there. If you define product as owning the customer experience, you might put it there.

    We ultimately decided on putting it in operations, because support wasn’t “our business”. Support doesn’t deliver the roadmap, and it doesn’t deliver the number, but it DOES need to do a great job for us to do both.
    Just like supply chain or legal, it was an organization that was chartered with enabling smooth operations and helping the other groups be successful in their charter.

    This structure ultimately worked really well for us by enforcing clear lines of accountability, and keeping my job as CEO manageable.

    • Raju Radhakrishnan

      Product support is critical for an enhanced life cycle of any customer to stay with the product. For an example if your product is monthly subscription by which if product support performance meeting expectations then customer would stay with us month over month . Bad product support leads to termination of subscriptions in the subsequent months. I would say support plays a vital role in driving sales numbers depending on nature of the business.

      Support is always good to be kept in Operations however we need to keep the customer experience performance review directly with CEO and it has to be next to the Sales #s on the same graph. When Sales team says Sales figures dropped for the current quarter, literally we need to correlate this with customer experience performance of the previous quarter.

  • I love this. Incredibly interesting and relevant for our organization. Thanks!

  • George Woods Baker

    This makes sense organically and intuitively. It also suggests an inverted pyramid which makes sense as well. Clearly these three components keep things simple, provide for strong lines of accountability and maintains focus. Scalability is largely assured as long as the integrity of the triangle is intact.

  • Daniel Barankin

    Your post couldn’t have come at a better time. Love the idea. Will implement it and update you on how it goes.

    • Super – I’m glad! I’d love to hear the feedback.

  • Anji Lopez

    Love this! I think not only at the top but the simplification of what people are responsible for will help the company perform. In IT I’ve seen one way defined as plan,build,run. Or leadership telling technical sales the responsibilities are revenue & deployment. When you simplify responsibilities of structures at large..,helps them focus on what’s important! I was curious that you didn’t have ibm included, it’s mostly similar to the Microsoft model… but really it’s like all of them at different points in the org! Lol

  • dineshtantri

    This seems like a great idea at this level of granularity. I’m curious to know what will the next level of the organization look like? The risk with this model is that we could end up with a number of activity centric teams within each of those machines as opposed to building cross-machine business outcomes based teams. Squads in Spotify for instance would be cross-functional with folks from across the three machines. Having said that this is a great model provided we are deliberate about accountability and owning outcomes. Agile IT Org Design goes into this in considerable detail.

  • Kaila Colbin

    Hi Brad, this is consistent with Ernesto Sirolli’s work — he calls them product, marketing and financial management, or, in even simpler terms, “You have to be able to make it, sell it, and look after the money!” He also says never in the history of the world has a single individual been born who can do all three of those things beautifully.

    His talk from our TEDxChristchurch event is on TED.com and well worth a watch (link below). But something else he talks about — that he doesn’t say in the TED talk — is that those three functions will never get along. The marketing (or, as you say, Customer) person will say, “We need it in blue!” The product person will say, “It can’t be blue — I’ve designed it to be green!” And the financial management person will say, “I don’t care what color it is — how much did you charge for it?” And the marketing person will say, “I gave it away to get in the door…” and round and round it goes.

    The art of entrepreneurship is the art of bringing a team together that can do all three of those things beautifully, and where every member truly appreciates the role of each of the other functions.
    https://www.ted.com/talks/ernesto_sirolli_want_to_help_someone_shut_up_and_listen

  • Sweet! Two suggestions:

    – use the term “Business” instead of “Company” for the third machine. To me, the term Business captures the financial and systems aspect, while the CEO is the one worried about the company as a whole.

    – to lead each machine, I propose the term “Steward” (via Stoos leadership: STewarship of Organic Systems). It avoids the politics of inflated titles, and emphasizes their responsibility is to make their machine serve the company — not vice versa.

    I agree these are best seen as roles rather than jobs. Still, the CEO could theoretically only have three reports:
    – Product Steward
    – Customer Steward
    – Business Steward

    Plus as assistant or Office of the CEO. As a physicist I love 3+1 dimensions. 🙂

  • John McGinnis

    Love this. Essentially the Why (customer), How (the company/business) and What (product). Three is the magic number.

  • Rosey

    Brad, I suggest you also include the thinking of Gino Wickman & Mark Winters in Rocket Fuel — the Visionary (CEO) and Integrator (COO). Some of the best running companies I’ve triaged in are running on EOS — the Entrepreneurial Operating System. The primer for EOS is ‘Traction’ by Gino. It’s good at putting the right people in the right seats (titles aren’t the point) and delivering accountability.

  • Trygve

    In our startup we use a similar organization structure however with dual markets (B2B2C) we have 4 pods and each pod leader is call ‘Champion’ (Product Champion, Business Champion, Client Champion, User Champion). We have communications processes and tools where the Champions are considered team members of the pods that are critical to their success and their input and knowledge is required across the pods. The Client and User Champions are aligned with the Product Pod and vise versa. I am the Chief Knowledge Officer / Co-Founder (CEO) and currently head one of the Pods and will step aside once scaling hits our goal or we have identified the A player to be the Champion (we / I know someone can do this better than I). Strategy, modeling, and investor relation responsibilities are with the CKO. One tool we have that helps ideas flow and be abundant in our company is our “Spaghetti Wall”, a wall where ideas are posted, without judgement. The company reviews all ideas weekly (obviously we are still small enough to do this) and makes decisions on what should be done with them. We’ve had some incredible ideas come from members outside the pod where the concept was implemented. Team members are compensated for ideas that get implemented (compensation is alternative methods: additional vacation time; day care for the month; dinners; vacations; bikes; etc). I believe your (this) concept yields team work and ownership of responsibilities that I have not seen in other entrepreneurial structures I have lead, or been involved with over the last forty-three years.

  • Corey Norman

    Makes me think of the Feasibility, Viability, Desirability framework by IDEO.

    I have encountered a number of organizations in their earlier stages split along the innovation and operations functions. XLR8 by John Kotter is a great book that discusses this division. There are pros and cons to virtually every org design. Defining the engines is a start but keeping them aligned and humming is the important part. What is being done to facilitate the interaction of the three?