Amazon Growth Strategy: The Flywheel, and How to Tell If Yours Is Working

What 'Amazon Growth Strategy' Actually Means

Amazon's growth strategy is a system, not a tactic. Jeff Bezos laid it out in the 1997 shareholder letter Amazon still attaches to its annual report: build for the long term, accept being misunderstood, and let customer obsession decide arguments that would otherwise get settled by quarterly targets. That's the whole doctrine. Everything else — Prime, AWS, the marketplace, the ad business — is an application of it.

The mechanism people call the flywheel works like this: lower prices and wider selection pull in more shoppers; more shoppers attract more third-party sellers, who add selection without Amazon carrying the inventory; more selection and more sellers generate the scale that lowers Amazon's per-unit cost structure; that lower cost structure funds the next round of lower prices. Each piece pays for the next one. None of it is new for 2022, or for this year — it's the same loop Bezos described in 1997, running at a much larger scale.

If you're looking for the source document behind most of what gets summarized as "Amazon growth strategy pdf," it's usually that 1997 letter, repackaged. Worth reading directly rather than through a slide deck.

The Flywheel Mapped to What a Brand Actually Controls

None of this explains how an individual brand grows on Amazon, because a brand doesn't control Amazon's pricing algorithm, its logistics network, or its balance sheet. What a brand does control is a miniature version of the same five levers — price competitiveness, catalog breadth, marketplace expansion, cost discipline in media spend, and the customer experience that turns a first purchase into a second one.

The table below maps Amazon's version of each lever to the brand-level equivalent. It's not a metaphor — it's the same structural logic, just running at a scale one company controls instead of one platform.

A Worked Example: What Growth Looks Like When You Measure It Correctly

Numbers make this concrete. Across 30 of reMKTR's Amazon DSP advertisers in July 2026, the portfolio delivered 6.04x return on ad spend — measured across the whole book, not cherry-picked from the best line item. That book ran 78.4 million impressions at a $4.00 CPM and a blended $1.42 cost-per-click (the $0.41 CPC people quote in this category is online-video only, not the whole book), for a blended cost per acquisition of $5.49 across 57,137 attributed purchases.

The number that tells you whether it's actually growth, rather than just spend showing up on an invoice, is this one: 20.1% of those purchases came from a shopper new to the brand. That's the flywheel-in-miniature. If most of what you're crediting to ads is existing customers who were going to buy anyway, you're not growing — you're paying for demand you already had. If a meaningful share is new-to-brand, the loop Amazon describes at corporate scale is happening at yours.

When the Number Is Bad News: Diagnosing a Stalled Growth Strategy

Sometimes the honest read is that growth has stalled, and the instinct is to spend more into the same setup. Check three things before you do that. First, has your new-to-brand share been falling while total attributed sales stay flat or rise? That usually means spend is shifting toward retargeting your own existing buyers, not toward acquisition. Second, are DSP and sponsored ads both taking credit for the same purchase in separate reports? Last-click attribution can't see across channels, so it double-counts routinely — reconciling in Amazon Marketing Cloud is the only way to see the real, de-duplicated number. Third, are CPMs or CPCs climbing faster than conversion rate, which quietly erodes the same margin Amazon's flywheel depends on protecting.

If the diagnosis says the growth isn't real, the fix isn't to scale harder. It's to hold part of the audience out, measure the lift against that holdout, and only reinvest in what the holdout proves — not what the dashboard credits.

The Common Mistake — Including One We've Made

The most common mistake brands make copying Amazon's growth strategy is copying the willingness to sacrifice margin without copying the measurement discipline that makes it safe. Amazon can run categories at thin or negative margin because it can prove, internally, exactly what that spend buys back in customer lifetime value. A brand that scales ad spend on last-click ROAS alone is making the same bet without the same evidence.

We've made this mistake ourselves. Last-click attribution is the default in Amazon's own console, and for years the industry — including us — reported against it because it was what the platform gave you. It cannot prove incrementality and never could; it can only tell you which ad happened to be the last thing a shopper touched before checkout. Holdout tests and matched controls can answer the actual question, which is whether the sale would have happened anyway. Moving to that standard changes which campaigns you'd call "growth" and which ones you'd call "noise."