Scaling on Amazon in 2026 Is a Margin Game, Not a Traffic Game

Hey everyone,

Welcome back for another bite to chew on.

Amazon is the biggest revenue line most DTC brands have. It is also the channel where the old playbook has quietly stopped working. Launch, run PPC, chase rank, repeat: that loop now feeds a fee stack and a CPC auction that eat the growth it produces.

2026 made the shift impossible to ignore. Fulfillment fees went up again in January. The low-inventory fee now bites at the variant level. And Rufus, Amazon's AI shopping assistant, is answering hundreds of millions of queries a day, which means a growing share of your customers never see a search results page at all.

The brands still scaling on Amazon are not the ones spending the most. They are the ones playing a structurally different game: SKU-level economics first, demand built off the platform, and ad accounts run like portfolios instead of slot machines. Here is what that looks like in practice.

On the Menu:

Let's get into it.

Hey everyone, Ron here.

A lot of you ask me who I trust when it comes to scaling on Amazon, especially when it comes to profitability, advertising, competitor positioning, and knowing what is actually happening inside the account.

That’s why I’m excited to share that IG PPC is my exclusive Amazon partner.

IG PPC is a hands-on Amazon growth partner built for serious Amazon brands, sellers, CPG companies, and aggregators that want to grow without guessing.

Their team supports brands across Amazon strategy, full-service account management, PPC, organic growth, keyword opportunities, competitor pressure, ranking gaps, and profitability.

They are not just looking at ACoS in a vacuum. They look at the full picture:

How your account is performing, where your spend is going, where competitors are winning, and what needs to change to scale profitably on Amazon.

For brands in my network, IG PPC is offering a free Amazon audit plus a free competitor analysis.

You’ll get a clear breakdown of what’s working, what’s holding you back, where competitors are beating you, and the biggest opportunities to improve performance on Amazon.

This is a great fit if you are already selling on Amazon and want sharper support around full-service Amazon management, PPC, profitability, ranking, and category growth.

Click here to grab the free audit and competitor analysis.

Every unit you sell on Amazon now passes through a gauntlet: the referral fee, the FBA fulfillment fee, inbound placement fees, storage fees, and the low-inventory-level fee. Base fulfillment fees rose again on January 15, 2026. None of these are headline numbers on their own. Stacked, they routinely take 40 to 50 percent of revenue before you spend a dollar on ads.

Two changes this year matter most. The low-inventory-level fee now applies at the FNSKU level, not the parent ASIN. One understocked variant generates fees on every sale of that variant, even when your total inventory looks healthy. The threshold sits at 35 days of supply.

The second is quieter: inbound placement fees drop to zero if you split shipments to five or more fulfillment center destinations. Brands that treat inbounding as an afterthought pay the fee. Brands that treat it as a lever do not.

Most brands look at blended account margin and decide they can afford to scale. Blended margin is a lie of averages. It hides the three SKUs quietly losing money on every ad-attributed sale behind the two SKUs carrying the account.

The fix is unglamorous: a per-SKU P&L that starts with sale price and subtracts referral fee, fulfillment fee, landed cost, storage, and the ad cost per unit at your current conversion rate. What is left is the number that decides whether a SKU deserves spend. Some SKUs earn scale. Some earn maintenance. Some earn a price change or a quiet exit.

Rank without margin is not an asset. It is an obligation to keep funding a loss.

The fastest way to change your Amazon P&L is not a better campaign. It is a better unit of sale. A single-unit listing with a $22 price and a $6 fulfillment fee has structurally worse economics than a three-pack at $58 with nearly the same fee.

Multipacks raise AOV against a mostly fixed fee base. Bundles differentiate against copycats competing on the single-unit price. And Subscribe & Save converts your best customers into a recurring revenue base that costs nothing to reacquire.

The operators scaling in 2026 design pack architecture before they design campaigns. The ad math only works when the unit economics underneath it do.

What you can do: Build a per-SKU contribution margin sheet this week, including ad cost per unit, and sort it. Scale spend only on SKUs that stay profitable after every fee, and fix or exit the rest.

Amazon search is no longer one algorithm. You are now optimizing for three layers at once: the classic A9/A10 keyword engine, the COSMO knowledge graph that interprets intent, and Rufus, the conversational layer sitting on top. Rufus is fully rolled out and handles hundreds of millions of queries a day.

That changes the job of a listing. Keyword-stuffed bullets tell Amazon what your product is. COSMO and Rufus want to know who it is for, when it gets used, and what problem it solves. A shopper asking "what should I bring to a week of trade shows" never types a keyword you bid on. Rufus answers with products whose listings carry that context.

Listings written like answers get recommended. Listings written like keyword inventories get skipped.

A10 weighs off-Amazon signals more heavily than its predecessor did, and the weight on converting external traffic has grown year over year. Brands sending outside demand to their listings are climbing competitive keywords faster than brands relying on PPC alone.

Amazon pays you to do this. The Brand Referral Bonus rebates roughly 10 percent of sales driven from external traffic, ranging from about 6.5 to 11.2 percent depending on category. That rebate applies against your referral fees, which means your creator and affiliate traffic is cheaper than it looks on paper.

One warning: the boost only comes from traffic that converts. Blasting incentivized clicks or junk traffic at a listing tells the algorithm your product disappoints people. Volume without intent actively hurts you.

The signals that compound on Amazon in 2026 are the ones that prove customers stay happy: review velocity, return rate, and repeat purchase behavior. These feed both the ranking algorithms and Rufus, which draws on review content when it recommends products.

This is where the margin work and the discovery work connect. A SKU with healthy contribution margin can afford post-purchase inserts, Subscribe & Save discounts, and the Vine enrollments that seed early reviews. A SKU scaled on thin margin cannot fund its own flywheel.

The brands that win treat reviews as an operations metric, not a marketing one. Fix the product issues your one-star reviews repeat, and the algorithm notices before your competitors do.

What you can do: Rewrite your top three listings to answer the who, when, and why questions a shopper would ask out loud, then stand up one external traffic channel with Brand Referral Bonus attribution turned on.

The easiest ROAS in your account is the spend defending your own brand name. It is also the least honest. A customer searching your brand name already chose you. Paying a CPC to complete that sale repurchases demand you already own.

Some branded defense is rational when competitors conquest your terms aggressively. But most accounts overweight it because it makes the blended ROAS look great, and blended ROAS is what gets reported upward.

The test is simple: pause a branded campaign in one region or match type for two weeks and watch what organic recaptures. Most brands find the true incremental value is a fraction of what the dashboard claims.

Sponsored Products hits a ceiling in every account. Past that ceiling, more budget buys higher CPCs on the same keywords, not new customers. The graduation path is Amazon DSP for reaching shoppers off the search results page, and Amazon Marketing Cloud for measuring what any of it actually adds.

AMC is the part most brands skip because it is operationally annoying. It is also the only place you can see path-to-purchase data across campaigns, run overlap analysis, and answer the question your CFO keeps asking: which spend is incremental and which is theater.

Scaling teams in 2026 run the same discipline sharp paid social teams adopted years ago: controlled tests, holdouts, and decisions made on lift instead of platform-reported attribution.

A scaled Amazon ad account holds three kinds of campaigns, and they should never be judged on one blended number. Launch campaigns buy data and rank for new SKUs, and they are allowed to lose money for a defined window. Harvest campaigns run proven keywords on proven SKUs, and they are held to strict efficiency targets. Defense campaigns protect shelf space, and they are judged on share, not ROAS.

Blend them into one TACOS number and every conversation gets dumber. The launch spend looks like waste. The harvest spend hides it. Nobody can say which lever to pull.

Split the account by intent tier, give each tier its own target, and review them separately. That is the whole trick. It is also the difference between an account that scales on purpose and one that drifts.

What you can do: Split your TACOS reporting into launch, harvest, and defense tiers this month, and run one branded-spend holdout test. Kill or cut whatever fails to prove incrementality.