Growth Accounting: Decomposing User Growth Into Its Component Forces

Your daily active users climbed 8% this month. That single number tells you almost nothing about what happened underneath it. Growth accounting decomposes any change in an active-user count into four component forces — new, retained, resurrected, and churned users — so you can see whether growth is real, borrowed, or masking a retention problem entirely.

Quick Answer: Growth accounting splits net user growth into New + Resurrected − Churned = Net Change against a Retained base. If churn is quietly rising while new users mask it, your top-line number is lying to you — and growth accounting is how you catch it.

What Is Growth Accounting, and What Does It Actually Measure?

Growth accounting is a decomposition method that breaks a period-over-period change in active users into the discrete behavioral segments that produced it. Instead of reporting "DAU is up," you report how many users are new, how many stuck around, how many came back from dormancy, and how many left. The net number is just the sum of those four forces.

The concept was popularized inside Facebook's early growth team, led by Chamath Palihapitiya and later documented publicly by growth executive Alex Schultz, who built the practice of tracking user cohorts as flows rather than as a single gauge. Andrew Chen, the former Uber growth lead now a general partner at Andreessen Horowitz, has written extensively about the framework as one of the most underused tools in a product team's analytics stack.

Origin note: growth accounting wasn't born as an academic framework — it came out of a growth team needing to explain, in a weekly review, exactly which lever moved the number.

The core insight is simple: a single aggregate metric can go up for reasons that should worry you. A DAU chart trending upward could mean your product is winning. It could also mean a paid acquisition campaign is dumping in low-intent users who churn within a week, while your core retained base quietly shrinks underneath them.

This is exactly the kind of blind spot that a broader data-driven decision-making practice is supposed to catch — growth accounting is one of the sharpest tools for doing it, because it refuses to let a single number stand in for the underlying behavior.

Why Aggregate Metrics Hide the Story

Consider two products with identical 100,000 DAU this month, up from 95,000 last month. One gained 8,000 new users and lost 3,000 existing ones. The other gained 5,500 new users, resurrected 1,500 dormant ones, and lost only 2,000. Same net number, completely different health.

Without decomposition:

• You cannot tell if your retention curve is improving or degrading.
• You cannot tell if acquisition spend is buying durable users or one-time tourists.
• You cannot separate a marketing win from a product win.
• You cannot forecast next quarter's DAU, because you don't know the underlying churn rate driving it.

This is precisely why growth accounting pairs so naturally with disciplined anti-vanity-metrics thinking — DAU alone is a vanity metric the moment you stop asking what's inside it.

The Four Forces Behind Every Change in Active Users

Every change in your active-user count is the sum of exactly four mutually exclusive segments, measured against a fixed prior period. Get these four buckets right and the math of growth accounting becomes almost mechanical — the hard part is defining the boundaries precisely and consistently.

New, Retained, Resurrected, and Churned, Defined Precisely

The precision matters more than it looks. A resurrected user is not a new user — they have history, a reason they left, and (often) a reason they came back. Lumping them into "new" in a dashboard erases the most interesting cohort you have: people who churned and then reconsidered.

1. New — first-ever activity in the current period, full stop.
2. Retained — active in period n-1 and period n. This is your foundation.
3. Resurrected — active in period n, inactive in period n-1, but active at some earlier point.
4. Churned — active in period n-1, inactive in period n. This is the segment most dashboards under-emphasize.

The relationship is a simple identity:

DAU (this period) = Retained + New + Resurrected
Net Change = New + Resurrected − Churned

A Worked Example

Below is a hypothetical monthly snapshot for a mid-size B2B SaaS product, illustrating how identical net growth can come from very different underlying mixes.

Both scenarios net to zero change. Scenario A is acquisition-dependent — cut the marketing spend that produces those 4,000 new users and the product shrinks immediately. Scenario B is winning back dormant users at three times the rate, which usually signals a product change (a re-engagement email, a fixed bug, a new feature) that's pulling lapsed users back in. The net number is identical. The story is not.

How to Build a Growth Accounting Model From Your Own Data

Building a growth accounting model requires only an events table with a user ID and a timestamp: tag each active user, each period, into one of the four buckets by comparing their activity to the prior period, then chart the buckets as a stacked bar. No specialized tool is required, though platforms like Amplitude, Mixpanel, and Heap expose this as a built-in report.

The mechanical steps:

1. Define "active." Pick one clear behavioral threshold (a login, a core action, a session) and use it consistently — inconsistency here invalidates every downstream number.
2. Choose your period. Weekly is standard for most consumer products; monthly suits lower-frequency B2B tools. Daily/weekly/monthly active user definitions should match your product's natural usage cadence.
3. Tag each user, each period, against the four categories using their activity in the current and prior period (and their full history, for resurrection).
4. Sum each bucket and chart it as a stacked visualization over time — this is the single most useful chart most growth teams don't build.
5. Layer in cohort analysis to see whether newer cohorts retain better or worse than older ones — this is where cohort analysis and growth accounting reinforce each other directly.

The Quick Ratio: One Number That Separates Real Growth From Fragile Growth

The Quick Ratio, a heuristic popularized by Facebook's growth team and later formalized in growth curricula like Reforge's, compresses the four buckets into a single fragility score: (New + Resurrected) / Churned. A ratio above roughly 2 suggests you're adding users faster than you're losing them by a healthy margin; a ratio drifting toward 1 means growth and churn are converging — the earliest warning sign of a plateau.

The Quick Ratio is a compass, not a verdict — it tells you which direction to point your investigation, not what you'll find there. A dropping ratio driven by rising churn (a retention problem) needs a completely different response than one driven by falling new-user volume (an acquisition problem).

What Growth Accounting Reveals That Vanity Metrics Hide

Growth accounting exists to surface two specific failure patterns that a single DAU or MAU number actively conceals: flat metrics masking two offsetting crises, and rising metrics masking an eroding core. Both patterns are common enough that most growth teams eventually get burned by one of them before adopting decomposition as a standing practice.

Case Pattern: Flat DAU That's Actually Two Crises

A flat top-line number reads as "stable, no action needed." Decomposed, it can reveal that both acquisition and retention are deteriorating simultaneously, and a one-time spike in resurrected users (say, a win-back email campaign) is the only thing holding the aggregate steady. Once that campaign's effect fades, the flat line becomes a falling one — with no warning, because nobody was watching the components.

Case Pattern: Growing DAU That's Actually a Retention Problem

A rising DAU line reads as unambiguous good news. Decomposed, it can show that retained users are shrinking as a share of total DAU every month, while new-user volume (often paid) grows just fast enough to overcompensate. This is the classic "leaky bucket" — the moment acquisition costs rise or budget tightens, the underlying churn problem is fully exposed.

Both patterns connect directly to why users churn in the first place, which is a customer journey question as much as an analytics one — the moments where users disengage usually map to specific friction points along that journey. It's also worth asking whether churned or resurrected users ever had the job the product claims to do; a jobs-to-be-done lens on your churned cohort often explains why the numbers moved, not just that they moved.

Reichheld's retention research at Bain & Company — the origin of the Net Promoter framework — has long argued that even modest improvements in retention compound into outsized gains in long-run value, precisely because retained users are cheaper to keep than new ones are to acquire. Growth accounting is the mechanism that lets you see whether your retention rate is actually moving before that compounding shows up (or fails to).

Connecting Growth Accounting to Your North Star

Growth accounting works best when the "active" definition you decompose is the same one anchoring your broader measurement strategy. If your team has already done the work to pick a single north star metric, growth accounting is the natural next layer — it explains the movement in that metric, not just its current value.

The Assumption Behind Every Growth Accounting Mistake

Nearly every growth accounting error traces back to the same root cause: an unexamined assumption baked into the model — about what counts as "active," which period boundary to use, or why a segment moved — that nobody wrote down when it was made. Six months later, in a retro, the team is reconstructing that assumption from memory, and memory is a bad analytics tool.

Common unlogged assumptions that quietly break a growth accounting model:

• "Active" was redefined mid-quarter (a new feature added a low-effort action that counts toward activity) without a corresponding note, so retention appears to jump for reasons that have nothing to do with the product improving.
• A resurrection window was assumed (30 days? 90 days?) without documenting it, so two team members build incompatible charts from the same raw events.
• A churn spike was attributed to a specific cause in a Slack thread, verbally, and never revisited — so when someone asks "why did churn spike in March" nine months later, the answer is gone.
• A Quick Ratio dip was dismissed as seasonal without writing down the reasoning, so nobody can check later whether that guess was actually right.

This is the practical argument for capturing a metric hypothesis or an analytics assumption the moment you form it, rather than trying to reconstruct your own reasoning during a retro. Prodinja's Journals are built for exactly this moment: you can log an assumption — "I think the March churn spike is the pricing change, not seasonality" — with real browser voice capture, so it's timestamped and revisitable later, instead of being reconstructed (imperfectly) from memory weeks after the fact.

That habit doesn't replace rigorous growth accounting. It protects it — because a decomposition model is only as trustworthy as the assumptions baked into its "active" definition and period boundaries, and those assumptions are worthless if nobody can find them again.

Key Takeaways

• Growth accounting decomposes net active-user change into four buckets — new, retained, resurrected, and churned — turning one ambiguous number into four diagnostic ones.
• Identical net growth can hide opposite underlying stories: acquisition-dependent growth and retention-driven resurgence can produce the exact same DAU delta.
• The Quick Ratio ((New + Resurrected) / Churned) compresses the model into a single fragility score — useful as a compass, not a verdict.
• A flat or rising top-line metric can mask a retention crisis that only decomposition, not the aggregate, will surface in time to act on.
• Pair growth accounting with cohort analysis and a defined north star metric so the decomposition explains movement in the number your team actually reports on.
• Every growth accounting model rests on assumptions — the "active" definition, the resurrection window, the attributed cause of a churn spike — and those assumptions need to be logged when made, not reconstructed later.