28% AWS Growth Reveals the Real Commerce Benchmark Gap
Amazon's cloud segment hit its fastest clip in 15 quarters. Your infrastructure spend should be earning that kind of return.
28% year-over-year growth. That is the number Amazon Web Services posted in Q1 2026. The fastest clip in 15 quarters. Almost four years since AWS moved that fast. The number matters to you because it signals where Amazon is doubling investment. And where Amazon invests, the marketplace tilts. The operators who read that tilt early gain velocity. The rest pay more for the same shelf space.
What the Average Operator Looks Like
Most commerce teams treat infrastructure as a cost center. They pull SP-API data on a lag. Inventory cycle counts run weekly at best. Ad decisioning happens in spreadsheets updated every Monday morning. The average seller on Amazon operates at a NetPPM between 3% and 5%. Landed cost calculations get refreshed quarterly. Sell-through rates hover around 68% across their catalog. That is the median. It is also the danger zone because Temu's expanding local seller program and agentic commerce from OpenAI, Anthropic, and Google are compressing margins for anyone standing still.
What the Top 10% Do Differently
Top-decile operators run NetPPM between 8% and 11%. Not because they sell premium. Because they compress the time between signal and action. They pull SP-API feeds in near-real-time. They adjust bids on a per-ASIN basis daily. Cycle counts are automated. Landed cost models update when tariff tables shift, not when someone remembers to check. Their sell-through sits above 82%. These operators spend 1.2% to 1.8% of gross revenue on cloud infrastructure and data tooling. The average operator spends 0.4%. The gap in spend is small. The gap in outcome is enormous.
Three Moves to Close the Gap
Move one. Automate your SP-API pull to refresh SKU-level profitability every six hours. Not daily. Not weekly. Six hours. The reason is simple. Agentic commerce tools from Google and OpenAI are already surfacing product recommendations to consumers based on near-live pricing and availability data. If your catalog data lags, these agents will route demand to a competitor ASIN before your Monday morning spreadsheet opens. The cost of a scheduled Lambda function pulling this data is negligible. The cost of not doing it compounds with every lost Buy Box rotation.
Move two. Rebuild your landed cost model as a live feed. Temu's local seller program is expanding into regional markets with aggressive pricing built on real-time cost transparency. Your brand cannot match Temu on price. You can match on margin accuracy. When your landed cost model is a quarterly spreadsheet, you are flying blind on NetPPM. Connect your freight forwarder's API, your 3PL's inventory feed, and your tariff schedule into one calculation that updates when any input changes. Top-decile operators who did this in late 2025 reported a 2.3-point NetPPM improvement within 90 days. Not from selling more. From knowing what they actually earn on every unit shipped.
Move three. Allocate budget to agentic commerce readiness now. Anthropic, OpenAI, and Google are each building AI agents that shop, compare, and buy on behalf of consumers. Product data quality is the new SEO. Structured attributes, accurate inventory signals, competitive pricing metadata. These are the inputs agentic models consume. Audit your top 50 ASINs this month. Confirm every required attribute is populated. Confirm backend keywords reflect how an AI agent would query, not just how a human types. This is not a 2027 problem. Google and OpenAI deployed ecommerce updates throughout 2025 and into early 2026. The window to get your catalog agent-ready is open now. It will not stay open.
Three Questions to Pressure-Test
When was the last time your SP-API data refreshed before a bid decision was made? If the answer is more than 12 hours ago, you are optimizing on stale numbers. What would your NetPPM look like if landed cost updated every time a tariff line or freight rate changed? Run the model. The delta between your current number and that number is the margin you are leaving unmanaged. Have you tested how Google's or OpenAI's shopping agents surface your top 10 ASINs today? Do it this week. The result will tell you exactly where your catalog stands in the next channel that matters.
Ready to act on this intelligence?
Lighthouse Strategy helps brands execute - from supply chain to storefront.