Growth Decomposition
Breaking a change in revenue into its underlying drivers - price, volume, mix, and segment - so you find the real cause.
What it is
Growth decomposition is the practice of taking a top-line number that moved - 'revenue fell 8%' - and splitting it into the pieces that actually drove the move. Revenue is never a single lever; it is quantity times price, summed across products, segments, and channels. So a decline can come from selling fewer units, charging less per unit, shifting toward a cheaper mix of products, losing a customer segment, or any combination. The method is a disciplined way to keep dividing the number until you reach a driver you can act on. Its power is diagnostic: it turns a vague 'sales are down' into a specific 'we lost volume in one region while price held', which points straight at a cause.
When to use it
Use it whenever a case gives you a change in a headline metric and asks why, or asks you to grow one. Profit-decline, revenue-drop, and 'grow this business' cases almost always reward decomposition early. It is also the honest way to size a growth opportunity: you cannot grow revenue without knowing which lever you are pulling. It is less useful when the question is qualitative (should we enter a market at all) or when you already know the driver. Do not decompose for its own sake - stop dividing once you reach a driver specific enough to act on.
How to apply it
Start from the identity: Revenue = Volume x Price, and both can be broken further (volume = customers x purchase frequency x units per purchase; and the whole thing splits by segment, product, geography, channel). Pick the split that matters for this business and follow the numbers. Ask which branch changed and which held steady - the branch that moved is your lead. Then keep drilling only down the branch that changed: if total volume fell but price held, do not waste time on price; ask which segment or product lost the volume. The discipline is to isolate one driver at a time and to let the data, not a hunch, choose which branch to open next.
Common mistakes
The most common error is stopping too early - blaming 'lower sales' without asking whether it was price or volume, which are completely different problems. The opposite error is decomposing everything into a sprawling tree with no hypothesis about where the answer lives. People also double-count (mixing a segment split with a product split so the pieces overlap) and forget that mix shift can move revenue even when nothing about price or unit volume changed. And a decomposition that never lands on an actionable driver has not finished its job.
Example
A subscription meal-kit company reports flat revenue and wants to know why growth stopped. You decompose: revenue = subscribers x average revenue per subscriber. Subscribers actually grew 15%, so the flatness must live in revenue per subscriber, which fell. Drilling there: price per box was unchanged, but average boxes per subscriber per month dropped - existing customers are ordering less often. That single branch reframes the whole problem from 'we have a growth problem' to 'we have a retention-of-usage problem among existing subscribers', which points at engagement and menu variety rather than acquisition spend. The decomposition, not intuition, located the real lever.
Practice it
Practice on profit- and revenue-change cases by building the driver tree out loud and letting each number tell you which branch to open next, stopping only when you reach a lever the company can pull. The skill to build is discipline - one clean split at a time - and knowing when the tree has gone deep enough.