How to choose your next growth engine

Growth Strategy2026-10-017 min readAtlas Strategy Group

When a company decides to grow beyond its current trajectory, the options are almost always the same five: a new customer segment, a new geography, a new product, a higher price, or a new channel. All five look plausible in a presentation. The choice between them is where most growth strategies quietly go wrong — not by picking a bad engine, but by picking the engine the founder finds exciting rather than the one the evidence supports.

Five criteria that actually separate engines

• Proximity. How close is the engine to what already works — same customers, same capabilities, same economics? Growth in the adjacent cell of the grid is cheaper, faster and more reversible than growth two cells away. Distance multiplies every cost.
• Cost per unit of proven growth. Not «budget required» — the money needed to learn whether the engine works at all. Some engines are cheap to falsify and expensive to scale; others are the reverse. You want to know which pair you are holding.
• Time to first honest evidence. How many months until reality casts its first vote? An engine that takes a year to produce its first real signal absorbs a year of capital and confidence before anyone can change course.
• Reversibility. An engine you can exit at the cost of a post-mortem is a different risk class from one that requires leases, hires and long contracts before the first customer appears.
• Concentration. What does betting on this engine do to the balance of the business? Sometimes the right choice is the second-best engine, because the company cannot afford a single point of failure in the only one it has.

The discipline of writing it down

The exercise that separates a decision from a preference is unglamorous: all candidate engines on one page, the five criteria as columns, filled in by the people who will execute — before anyone has fallen in love. The numbers will be rough. The argument that follows will still be real, because it is written by the same people who will spend the money. Companies skip this page for exactly the reason it works: on paper, the exciting engine often loses to the boring one, and nobody wants that argument at the stage where it is still cheap.

The cheapest growth is usually adjacent to something that already works. The most expensive is adjacent to something the founder wants to be.

Sequencing beats selection

The strongest versions of this decision are rarely «engine A forever». They are portfolios with an order: the near engine funded first to pay for the experiment with the far one; each engine's next tranche released by the evidence its own last tranche produced. This turns a one-time bet into a system that learns — and it is the difference between a company that grows on decisions and a company that grows on momentum.

Choosing the engine is a large part of what the growth strategy practice is hired to do — and the input that makes the choice honest is the diagnosis of where growth already comes from.