Machines & Phases

There are so many different ways to measure a startup. It’s easy to drown in metrics. How do we know what we should focus on?


You can only make as well as you can measure
— Joseph Whitworth


Here are two different ways we can divide and conquer this problem…

Three Machines

Consider the different people working on the team and the responsibilities they have.

There are broadly three specialities: 1

  1. Product - those responsible for developing, maintaining and operating the product or service;
  2. Customer - those responsible for marketing, selling, on-boarding and supporting customers to use the product or service; and
  3. Company - those responsible for recruiting, retaining, running and (not least) funding the team that makes everything else possible.

Three Phases

Or, consider the progression of a startup from early-stage to successful mature business.

Again, there are broadly three stages:2

  1. Build - in the very beginning we need to focus on developing a product that solves a real and valuable problem and attract some initial customers.
  2. Grow - then, our focus shifts to overcoming obscurity and everything we need to do to promote and sell the product to more and more customers.
  3. Profit - finally, our focus shifts to doing all of those things efficiently and ensuring that the amount we spend on each customer is relative to the amount we earn.

Nine Intersections

We can use these two groupings to create a 3x3 matrix, and use that to help us think about the important measures of success for each team at each stage.

What’s at the intersections?

We can work down each column in this framework to consider the progression for each area of the business…

Product Team

The first job is to develop a product or service that solves a problem. Perhaps we stumble across that immediately, but more likely we need to work hard while customer numbers are small to understand if people are actually using what we’re building in the way we anticipated, and adjust accordingly.3 As the number of customers grows we can start to segment, and understand what is attracting our best customers and also what are the gaps for those customers we don’t retain.4 Later, as we get beyond early adopters we need to pay more attention to things like customer churn and the cost of servicing each customer, so we can invest in the right areas.

We progress from measuring active users to measuring delighted users to finally measuring loyal users.

Customer Team

In the early-stages we’re looking for initial customers who are willing to use the product right away, even though it’s only half built. We need to understand how customers move through our sales funnel - e.g. using free trials or a land-and-expand approach. To move beyond one-by-one sales we need to experiment with the different sales channels that are available and understand the costs associated with each. We need to learn what generates growth - perhaps existing customers spread the word for us, but more likely we need to buy it. Maybe there are also opportunities to increase the revenue we earn from existing customers. Later, with the benefit of more time, we can track the lifetime value of different customer cohorts and use that detail to tune our sales process.

We progress from measuring sales conversions to measuring sales margins to finally measuring sales efficiency.

Company Team

At the beginning cash is typically the most critical constraint. Either we need to ensure that the amount we’re spending is relative to the initial revenue we’re generating, or we need to raise external capital to fund the difference through the early-stages. As the team grows we need to focus on team culture (in simple terms, the things we do repeatedly) and make sure we’re attracting the right kind of people to fill the different roles we have and keep them engaged and happy. Eventually we also need to consider how those who have invested time and money in creating the business will get a return.

We progress from measuring cashflow to measuring team engagement to measuring shareholder returns.

Walk → Run → Fly

We can also work across each row in this framework to quickly identify the basket of metrics that are likely to be most important at each stage of the business…

In the early-stage we measure active users, our sales funnel and cashflows.

As growth accelerates we measure net promoters, net new revenue and cost-to-acquire, plus team engagement.

And, as the business matures we measure customer retention and churn, lifetime value (or perhaps more usefully average contract value) and return on investment.

Trying to manage without metrics is like flying blind.

But equally managing entirely by metrics is equally misguided.

It’s common to hear people claim they are led by data. They say they do what the data tells them to do. But it’s never as simple as that in my experience. Data can answer our questions, but we still need to ask them.

It’s also common to manage using incentive schemes based on measurable results. But, whenever we do that we leave ourselves open to people gaming the system or chasing selfish or short-term results at the expense of long-term or collective outcomes. So every measure needs a counter-measure to ensure that the resulting benefits aren’t weighed down by greater associated costs.

Often the biggest and most important decisions we need to make come down to subjective judgement. In those moments we should make sure we have all of the data we can get our hands on to inform our decision. But let’s not pretend that the numbers will make difficult decisions for us.


  1. See: The Three Machines by Brad Feld. ↩︎

  2. See: Lean Analytics by Alistair Croll & Benjamin Yoskovitz, specifically chapter 5 where these three phases are described as:

    • Product Development
    • Customer Growth
    • Unit Economics

    They are themselves expanding on the stages of a startup that Eric Ries described in his book The Lean Startup: make something that is sticky, then make something that is viral and then finally make something that generates revenue↩︎

  3. A common trap at this stage is to be overly influenced by what potential customers say they will do, only to later discover they don’t do that at all once we’ve actually built what they said they wanted. ↩︎

  4. See: Net Promoter Score↩︎