learn-metric-decomposition-with-phoebe / Leader session 4 of 6
Learn Metric Decomposition with Phoebe · Leader track · Session 4 of 6

The six domains, side by side

You have learned to read a driver tree, spot leading indicators, and trace a drop. Now the question every leader eventually asks: does this method survive contact with a business that is nothing like an online store? A CFO runs finance, a CMO runs brand, a founder runs a SaaS book - do they all decompose the same way? Today we lay six very different businesses on the same table and show that the method is one, even though the shape of the tree changes.

🔵 Leader track Leaders: CxO · VPs · founders No code - thinking mode 45 min
0-3 · Welcome 3-22 · Same method, six shapes 22-42 · Read across the domains 42-45 · Q&A
Part 0

Why put six businesses on one page

Most leaders learn metric thinking inside one industry and quietly assume it does not travel. It does. A driver tree is a way of writing down "this outcome is made of these levers" - and every business, from a marketplace to a government contractor, has that structure. What changes is the shape of the tree: some multiply, some bridge, some funnel. Today you see all three shapes, across six domains, so you can walk into any business unit and ask the same three questions with confidence.

Live - presented in session Self-study - read after class Framework sources covered
★ What you walk out with today One method that transfers across every business you will ever lead: a mental map of the six top-line trees, the ability to tell a multiplicative tree from a bridge from a funnel on sight, and the discipline to pick one top-line metric per unit instead of drowning in a dashboard.
Part 1 · covers the domain metric canon, the three tree shapes

One method, three shapes of tree 9 min live

The method never changes: name the top line, break it into the levers that make it, assign each lever an owner. But the levers do not always combine the same way. In some businesses the top line is a product of rates and volumes. In others it is a bridge - a starting balance plus and minus flows. In others it is a funnel - a wide top narrowing stage by stage. Learn to spot which shape you are looking at and the rest of the diagnosis follows.

1 · Ecommerce GMV = Traffic x Conv x AOV Shape: multiplicative Rate x volume, clean pass-through 2 · Marketing ROAS = Revenue / Spend Shape: funnel to revenue Spend to clicks to leads to won 3 · Branding Reach to intent to sales Shape: brand funnel Share of search leads share 4 · Internet / traffic Ad rev = Impr x Fill x CPM Shape: multiplicative Impr = DAU x sessions x slots 5 · B2B SaaS ARR bridge (flow) Shape: bridge Start + new + exp - churn 6 · B2G Bookings = Pipe x Win x ACV Shape: multiplicative Plus mission-outcome KPIs Same three questions everywhere: what is the top line, what are its levers, who owns each one?
🔍 Click to zoom - six businesses, one method, three shapes of tree
LiveThe three shapes, and how to tell them apart3 min

You do not need to memorise six trees. You need to recognise three shapes and know which businesses wear which. Once you can name the shape, you know how a drop will behave and where to look first.

  • Multiplicative - the top line is a product of rates and volumes. Ecommerce GMV, ad revenue, bookings. A percentage move in one lever passes straight to the top line, so diagnosis is fast: find the lever that moved by the matching percent.
  • Bridge - the top line is a starting balance plus and minus flows. SaaS ARR: you begin the quarter with a number, add new and expansion, subtract contraction and churn, and land on the ending number. You diagnose by asking which flow changed, not which rate.
  • Funnel - a wide top narrows stage by stage to a converted outcome. Marketing spend to won, brand reach to sales. You diagnose by finding the stage where the conversion rate slipped.
Self-studyWhy the shape decides your first move2 min read

The shape is not trivia - it tells you the first cut when a number moves. In a multiplicative tree you compare percentage changes across levers. In a bridge you compare this quarter's flows to last quarter's flows - the churn line, the expansion line. In a funnel you compare stage-conversion rates to their baseline to find the leaky stage.

  • Multiplicative: ask "which lever moved, by what percent?"
  • Bridge: ask "which flow changed - did more churn out, or did less new come in?"
  • Funnel: ask "which stage stopped converting at its usual rate?"
Self-studyThe same drop, told three ways2 min read

Imagine revenue is down 6% in three businesses. In the ecommerce store, "conversion fell 6%, everything else held" - one lever, one owner. In the SaaS book, "new bookings held but churn doubled" - a flow problem, a retention owner. In the brand-led business, "reach held but consideration slipped" - a mid-funnel problem, a messaging owner. Same headline, three completely different rooms to walk into. The tree shape is what routed you to the right room.

Real world

Marketplaces run two trees at once. A ride-hailing or delivery business decomposes both supply (drivers, couriers, listings) and demand (riders, orders) as separate multiplicative trees, then joins them at the marketplace level. Leaders who only watch the demand tree miss supply-side drops entirely - which is why the best operators keep both trees on the same wall.

Part 2 · covers leading vs lagging across domains

Leading and lagging, domain by domain 8 min live

Every domain has a lagging top line the board watches and a leading indicator that moves first. The leader's edge is knowing the leading one for each business you touch - because that is the number that warns you weeks before the board number confirms the bad news. Here is the whole set on one table.

DomainTop line (lagging)A key leading indicator
EcommerceGMV / revenueAdd-to-cart rate, checkout starts
MarketingMarketing-sourced revenueMQL-to-SQL rate, cost per lead
BrandingMarket shareShare of search, share of voice
Internet / trafficAd revenueDAU, sessions per user, fill rate
B2B SaaSARRPipeline coverage, net revenue retention
B2GContract bookingsPipeline coverage, win rate on late-stage
LiveThe leading indicator is where you actually lead3 min

The lagging number is for the board; the leading number is for you. GMV confirms last month happened. Add-to-cart rate is falling right now and gives you weeks to act before it shows up as revenue. The whole point of decomposing a metric is to get down to leaves that lead - so you steer by them and report the root.

  • Report the lagging root - it is the accountable, board-facing number.
  • Manage by the leading leaves - they warn early and are quick to move.
  • Pipeline coverage deserves special mention: in both SaaS and B2G it is the single earliest signal, because a thin pipeline today is a bookings miss two quarters out.
Self-studyShare of search - the brand leader's leading number2 min read

Market share is the slowest number a brand leader watches - it lands quarters late and only after the market has already moved. Share of search (your brand's share of category search volume) tracks it closely and arrives weekly. When share of search dips, brand-driven sales tend to follow. It is the closest thing brand marketing has to the add-to-cart rate: an early, cheap, leading proxy for a slow, expensive, lagging outcome.

Self-studyBoardroom moment - one leader, five business units2 min read

A group operating officer inherited five business units - a marketplace, a media property, a SaaS product, a services arm, and a government-contracts desk. Five industries, five vocabularies, five sets of dashboards that did not talk to each other. She refused to learn five methods. Instead she made every unit head answer the same three questions in the monthly review: what is your one top-line metric, what are the three-to-five drivers under it, and who owns each driver? Within two quarters the reviews got shorter and sharper. The vocabularies still differed, but the thinking was identical - and she could now smell a problem in a business she had never operated.

Real world

This is the transferable skill. Leaders who rise into multi-unit roles cannot be domain experts in everything. What travels is the method: same three questions, applied to whatever tree the unit happens to have. The domain expert knows the answers; the leader knows the questions.

Part 3 · covers Lean Analytics, the one metric that matters

Pick the one top-line, then its drivers 6 min live

The trap on the other side of decomposition is sprawl. Once teams learn to break metrics down, they break down everything, and the dashboard grows to two hundred tiles nobody reads. Lean Analytics gives the antidote: at any moment, one metric matters most. Pick the one top-line for the unit, put three to five drivers under it, and let everything else be a leaf you consult only when a driver moves.

The ONE top-line what this unit is judged on Driver 1 has an owner Driver 2 has an owner Driver 3 has an owner Driver 4 optional Driver 5 optional Three to five drivers, each with an owner. More than five and nobody watches any of them.
🔍 Click to zoom - one top-line, three to five owned drivers, and no more
LiveThree to five, not thirty3 min

Attention is the scarce resource in any review. A tree with five drivers gets watched; a tree with thirty gets ignored. Discipline is choosing the small number of levers that explain most of the movement and pushing the rest down into leaves you consult only on demand. If everything is a headline metric, nothing is.

  • One top-line per unit - the number the unit is judged on this quarter.
  • Three to five drivers - the levers that explain most of its movement.
  • Everything else is a leaf - real, tracked, but consulted only when a driver above it moves.
Self-studyThe one metric changes as the business changes2 min read

Lean Analytics is clear that the one metric that matters is not permanent. An early SaaS company watches activation; a scaling one watches net revenue retention; a mature one watches margin. The tree stays; the metric at its root is promoted as the business matures. The skill is noticing when your one metric has stopped being the constraint and a different driver has taken its place.

Real world

Dashboard sprawl is a leadership failure, not a tooling one. When a review has forty charts, it is not because the tool made too many - it is because no leader was willing to say which one number matters most this quarter. Picking the one top-line is an act of leadership, and it is uncomfortable precisely because it makes you accountable for a single number.

Homework

Try it yourself - this week ◐ 20-30 min total

Framework sources

Frameworks this session draws on

This session leans on the frameworks that make cross-domain metric thinking durable. This page draws on:

Lean Analytics (Croll & Yoskovitz) - the one metric that matters, per business modelPart 3 · pick one top-line, resist dashboard sprawl
Domain metric canon - GMV, ROAS, brand funnel, ad revenue, ARR bridge, bookingsParts 1 & 2 · six top-line trees and their shapes
Amplitude North Star Playbook - one output, a few inputs that move itPart 3 · one top-line with three to five drivers; full treatment in b3
Check yourself

Three questions before you go 🎯 ◐ 90 seconds

1 · A SaaS ARR number is a "bridge" tree. What does that mean for diagnosis?

A bridge is a starting balance plus and minus flows: start + new + expansion - contraction - churn. You diagnose it by comparing this period's flows to last period's, not by reading rate pass-through.

2 · Why does a brand leader watch share of search, not just market share?

Market share lands quarters late. Share of search tracks it closely and arrives weekly, giving the brand leader early warning - the same role add-to-cart rate plays for ecommerce.

3 · Your unit's dashboard has thirty headline metrics. What does Lean Analytics tell you to do?

If everything is a headline, nothing is. Choose the one top-line the unit is judged on, keep three to five drivers under it, and consult the rest only when a driver above them moves.

Leader session 4 cheat sheet · pin this

One method, three shapesMultiplicative (product of rates x volumes), bridge (start + flows), funnel (wide top narrowing). Name the shape first.
MultiplicativeEcommerce GMV, ad revenue, bookings. A % move in one lever passes straight to the top line. Find the matching lever.
BridgeSaaS ARR = start + new + expansion - contraction - churn. Diagnose by which flow changed, not which rate.
FunnelMarketing spend to won, brand reach to sales. Diagnose by the stage whose conversion rate slipped.
Lagging vs leadingReport the lagging root to the board. Manage by the leading leaf - add-to-cart, share of search, pipeline coverage.
The same three questionsWhat is the top line, what are its three-to-five drivers, who owns each? Ask them in every unit you touch.
One top-line per unitLean Analytics: pick the one metric that matters now. Three to five drivers under it. Everything else is a leaf.
Running skillThe method transfers across every domain. Next: a5, run your tree through a weekly business review.