Your first metric tree in 10 minutes

We are going to build a metric tree for a fictional B2B SaaS company called Relaybox. Relaybox sells a team collaboration tool on a per-seat, monthly subscription. They have around 400 paying accounts, a small sales team, and a product-led growth motion where users can sign up for a free trial and upgrade themselves.

This profile is deliberately ordinary. If you work at a SaaS company, a lot of this will feel familiar. If you work in a different industry, the principles transfer. The point of a metric tree is to model how your specific business creates value, so the structure will look different for an e-commerce company or a marketplace, but the process of building it is the same.

Before we touch any tools, we need to answer one question: what is the single most important metric for Relaybox? Everything else in the tree will exist to explain and influence that number.

Every metric tree starts with a single metric at the top. This is your North Star: the one number that best captures the value your business delivers. For Relaybox, monthly recurring revenue (MRR) is the clear choice. It reflects how much value customers are willing to pay for every month, it is measurable on a regular cadence, and teams across the company can influence it.

You might be tempted to start with something broader, like customer satisfaction or market share. Resist that urge. A good North Star is specific enough to decompose and measure, but broad enough that multiple teams contribute to it. MRR fits that test. Sales, marketing, product, and customer success all have a hand in growing and retaining it.

Do not overthink this step. If you are unsure, pick the number your CEO would check first every Monday morning. For most B2B SaaS companies, that is MRR or annual recurring revenue (ARR). You can always adjust later once the tree reveals whether your choice decomposes cleanly.

Now we ask: what are the two to four factors that directly determine MRR? For a per-seat SaaS subscription like Relaybox, there is a clean mathematical decomposition: MRR equals the number of paying accounts multiplied by the average revenue per account. This gives us two first-level branches.

This decomposition is multiplicative. If either branch grows, MRR grows. If one doubles and the other stays flat, MRR doubles. That mathematical clarity is what makes metric trees powerful. You are not just drawing boxes and arrows; you are building an equation that describes how your business actually works.

Some businesses will have an additive first level instead. A company with three distinct product lines might decompose total revenue as the sum of revenue from each product. The structure depends on your business model. The key is that the first level should be exhaustive (it fully explains the parent) and that the branches should be meaningfully different from each other.

With two first-level drivers in place, we now decompose each one further. Start with the left branch: Paying Accounts. How does Relaybox gain and lose accounts? The number of paying accounts at any point in time is a function of three flows: new accounts acquired, existing accounts retained, and accounts lost to churn. So Paying Accounts decomposes into New Accounts, Retained Accounts, and Churned Accounts (which has a negative relationship to the parent).

New Accounts can be broken down further by acquisition channel. Relaybox gets customers through three routes: organic sign-ups from their website, paid acquisition campaigns, and referrals from existing customers. Each of these is a lever that a specific team or person controls.

Now take the right branch: Average Revenue per Account. For Relaybox this is driven by the number of seats per account (since they charge per seat) and the plan tier that each account is on. Larger teams on higher-tier plans pay more. Both of these are influenced by the product and customer success teams through onboarding, feature adoption, and expansion selling.

Notice how the tree is already starting to tell a story. If MRR drops, there are really only a handful of possible explanations: fewer new accounts, more churn, fewer seats per account, or a shift toward lower-priced plans. Instead of guessing, you can trace the tree downward and find out exactly which branch moved.

This is three levels deep, and for many teams that is enough to start. You do not need to model every metric in your business on day one. The goal is to capture the major cause-and-effect relationships so that everyone shares the same mental model of how the business works.

The tree we have built so far is mostly made up of lagging indicators. Paying Accounts and MRR tell you what has already happened. To make the tree truly actionable, we need to push one level deeper and add the leading indicators that predict what will happen next.

For the New Accounts branch, the leading indicators are the metrics that sit earlier in the funnel. Organic Sign-ups is driven by website visitors and trial conversion rate. Paid Acquisition is driven by ad spend and cost per acquisition. Referrals are driven by the number of active referral invites sent by existing customers.

For Churned Accounts, a powerful leading indicator is product engagement. At Relaybox, accounts that log in fewer than three times per week in their first month churn at twice the rate of active accounts. That engagement metric becomes a critical early warning signal in the tree. Customer support ticket volume and NPS scores are additional leading indicators that predict churn before it shows up in the numbers.

1. 1

   Website Visitors and Trial Conversion Rate

   These feed into Organic Sign-ups. Marketing owns website traffic; Product owns the trial-to-paid conversion experience. When sign-ups dip, you can immediately see whether the problem is fewer visitors or a lower conversion rate.

2. 2

   Ad Spend and Cost per Acquisition

   These feed into Paid Acquisition. The growth team controls how much is spent and which channels receive budget. Rising CPA signals that channels are saturating or that targeting needs to be refined.

3. 3

   Weekly Active Usage

   This feeds into Retained Accounts (and inversely into Churned Accounts). The product team influences engagement through onboarding flows, feature adoption, and in-app guidance. A drop in weekly active usage is the earliest signal of future churn.

4. 4

   Expansion Conversations

   This feeds into Seats per Account and Plan Tier Mix. The customer success team identifies accounts that are ready to add more seats or upgrade to a higher plan. Tracking the number of expansion conversations gives you a leading read on revenue per account growth.

A metric tree without owners is just a diagram. The final step is to assign a named person to every metric in the tree. This is what turns structure into accountability. At Relaybox, the assignments might look like this: the VP of Revenue owns MRR, the Head of Marketing owns New Accounts and its sub-branches, the Head of Product owns weekly active usage and trial conversion rate, and the Head of Customer Success owns Retained Accounts and expansion metrics.

Ownership does not mean that person is solely responsible for moving the number. It means they are the first person who investigates when the metric changes, the person who decides what action to take, and the person who reports back on whether that action worked. They are the steward of that part of the tree.

Once ownership is in place, the tree becomes a living operating model. Each week, owners check their metrics. When something moves unexpectedly, they trace the tree to understand why, log an action, and follow up. Over time this builds a habit of structured investigation that replaces ad-hoc firefighting with systematic learning.

You now have a complete metric tree. It took us five steps: choose the North Star, decompose into first-level drivers, expand each branch, add leading indicators, and assign ownership. The entire process, done with a small group around a whiteboard or in a shared document, can genuinely be completed in ten minutes for a first draft.

The first draft is not the goal. The goal is to start using it. Pin the tree somewhere visible. Refer to it in your weekly stand-ups and planning sessions. When someone asks "why did this number change?", point them to the tree. Within a few weeks, the tree will become the shared language your team uses to talk about the business, and that shared understanding is where the real value lives.

Use this checklist to make sure your first metric tree covers the essentials. It is not a gate-keeping exercise. If you can tick most of these boxes, you have a solid foundation to build on.

1. 1

   Single North Star at the top

   Your tree has one root metric that represents the most important outcome for the business. If you have two North Stars, you have zero.

2. 2

   Clean first-level decomposition

   The first level fully explains the North Star, either as a mathematical equation (multiplicative or additive) or as the exhaustive set of causal drivers.

3. 3

   Three to four levels of depth

   Deep enough that the bottom-level metrics are things individual teams or people can directly influence. Shallow enough that the tree is navigable without a map.

4. 4

   Mix of leading and lagging indicators

   The top of the tree tells you what happened. The bottom tells you what is about to happen. Both are needed for the tree to drive action, not just reporting.

5. 5

   Every metric has a named owner

   Not a team. A person. If nobody is willing to own a metric, consider whether it belongs in the tree.

6. 6

   The tree is visible and referenced regularly

   A metric tree that lives in a forgotten document is not a metric tree. It is a diagram. Pin it, share it, and refer to it in weekly conversations.