Each team's daily metric should trace, in one line, to the outcome the business is actually chasing.

A crowded dashboard is not a strategy. We connect the outcome to drivers teams can act on and add guardrails that reveal when one metric improves at another team's expense. You end up with a KPI tree that lets each team trace its day-to-day metric back to the business outcome it is meant to support.

Plan KPI architecture All Measurement Strategy & Analytics Governance tasks

Some of the 500+ brands we've worked with

See all references

We organize metrics as a connected system of outcomes, drivers, diagnostics, and guardrails. Four stages turn strategic intent into a metric system with clear formulas, ownership, targets, and review points.

How we hold ourselves to it

Define the outcome metric — We work with leadership to agree on the one metric that best represents whether the strategy is working.
A tree of drivers teams can actually move — We break the outcome down into drivers that teams can actually move day to day, with a lever attached to each one.
Every metric gets a formula, a source, and an owner — Each metric gets a precise formula, data source, and owner, so two people calculating it independently get the same number.
What happens when someone games this number? — For every metric a team might be tempted to over-optimize, we define a counter-metric that catches the damage before it becomes a bigger problem.

Clarify the strategy
We interview leadership to define what the business is trying to achieve and the timeframe for that outcome. The leadership agrees on the one outcome that counts.
Strategy summary
Build the metric tree
We decompose the outcome into drivers, diagnostics, and guardrails, reviewing each proposed relationship with its owner. The metric owner confirms they can move the driver.
KPI tree
Write metric contracts
We document the exact formula, source, and owner for every metric in the tree. The owner signs off on the formula and data source.
Metric registry
Set targets and review cadence
We use the available history to help set realistic targets and agree on how often the tree should be reviewed. The leadership approves the target and review cadence.
Target and review plan

A metric without a guardrail is an invitation to be gamed.

Automation summarizes the leadership interviews into candidate outcomes, drafts driver relationships from historical data, checks formula consistency across the registry, and projects targets from trend. Ownership is the human part: leadership names the outcome, a metric owner confirms they can actually move a driver, and targets and cadence are approved rather than calculated.

You receive a governed metric system, not another list of numbers without context.

Architecture document
KPI tree
A visual document connecting the outcome metric to its drivers, diagnostics, and guardrails.
Reference document
Metric registry
The exact formula, source, and owner for each metric, enabling reproducible calculations.
Operating guide
Target and review plan
Targets grounded in available history and a cadence for checking whether the tree still reflects reality.

We call it done when: every metric has an owner who agrees they can move it, two people compute it the same way from the registry, and each driver has a guardrail where gaming is possible.

Too many metrics and no shared owner usually means the same argument keeps happening in different meetings.

A good fit when

Teams report many metrics but cannot explain which ones have a meaningful connection to strategy.
Departments calculate the same KPI from different formulas, so growth and conversion numbers start a fresh argument whenever teams compare them.
Targets lack clear baselines or owners, or no one has defined what to monitor when improving one metric harms another.

An architect arranging KPI blocks into a pyramid with a north-star metric at the top

Better handled as other work when

You need to define which events to collect, not which metrics matter for strategy, that's Measurement Strategy & Tracking Plan.
You need dashboards built to display already-agreed KPIs, that's Looker Studio Dashboard Development.

If one of these is closer to your situation, start here instead: All Measurement Strategy & Analytics Governance tasks

We call it done when: leadership has agreed on the one outcome that counts, so the tree has a root instead of a committee.

Zeo designs measurement systems that connect a business decision to governed collection and reporting you can check. The people shown here work on the part of that system this page covers.

Content we've produced on this topic

Videoby Ezgi Gülsen YaylıData Studio ile SEO KPI'larını Ölçümleme - Ezgi Gülsen Yaylı Videoby Ezgi Gülsen YaylıSEO Kampanyalarında Performansın Doğru Ölçümlenmesi - Ezgi Gülsen Yaylı

Google Analytics
checked while writing each contract, to confirm the metric is actually measurable as defined
Notion
carries the metric tree and each metric's written contract, from first draft to the review cadence

Bring your current dashboards and the strategy they are meant to support. We will build the metric tree that connects them with explicit assumptions and ownership.

Plan KPI architecture

How many KPIs should we actually have?

Use only as many KPIs as needed to represent the outcome, its meaningful drivers, and the guardrails that matter. Diagnostic metrics may be numerous, but labeling all of them as KPIs removes the distinction.

What's the difference between this and a tracking plan?

A tracking plan defines the events to collect. KPI architecture defines which metrics matter strategically and how they relate, assuming the necessary data exists or identifying what is missing. Where a tracking plan gives you a name and a definition for an event, this gives every metric its own formula, source, and named owner. The two documents work together, but they answer different questions.

Can you also build dashboards for the KPI tree?

Not within this task. Here we define what the metrics are and what they mean. Looker Studio Dashboard Development covers the reporting layer built on top.

Whose time do you need for this?

We need time with leadership to clarify strategy and with each prospective metric owner to confirm that they can genuinely influence the number.

By team

By industry

By tool

KPI & Goal Architecture

Leadership names the outcome. The teams beneath decide how to reach it.

Clarify the strategy

Build the metric tree

Write metric contracts

Set targets and review cadence

Two people calculating the same metric should land on the same number.

KPI tree

Metric registry

Target and review plan