Run a Prime Day Post-Mortem: Turn Last Year's Data Into This Year's Wins

Most Prime Day prep starts from a blank page every year, as if the last event never happened. That's a waste, because the single best predictor of how your products will perform this Prime Day is how they performed at the last one. The sellers who consistently win the event treat each one as a controlled experiment and run a real post-mortem afterward — not a vibe-check, an actual look at the numbers. Do that once and your next event plan more or less writes itself: which SKUs to stock deep, which discounts to repeat or kill, and which products to leave out of it entirely. Here's the debrief worth running while the data is still fresh.

Separate real winners from revenue mirages

The first and most important pass is to sort what actually worked from what only looked like it worked. A SKU that did huge revenue on the day can still be a loser if it only moved because you slashed the price to the bone. Go product by product and ask the question that matters: did this lift in units *and* hold its margin, or did it lift only because the discount gave the profit away? The high-revenue, low-or-negative-margin product isn't a Prime Day success — it's a margin donation dressed up as one, and repeating it next year just means donating again at a bigger scale.

The five questions every post-mortem should answer

Pull last event's performance and answer these for each meaningful SKU. They map almost one-to-one onto next year's plan:

• Which products multiplied their normal velocity, and by how much? These are your stock-deep, spend-here candidates for the next event.
• Which sold out early? You under-stocked them and left money on the table — buy deeper this time, because you can't recover a sold-out surge.
• Which barely moved despite the traffic? The event doesn't favor them; don't over-commit inventory or ad budget there again.
• Which only sold because the discount was steep? Decide whether a shallower discount still moves them, or whether they don't belong in the event at all.
• Where did elevated ad costs outrun the return? Clicks get pricier during the event; flag the SKUs where the ad math stopped working so you redirect budget next time.

Turn the answers into a stock-and-spend plan

Once those questions are answered, your inventory and advertising plan is mostly a sorting exercise. The high-lift, healthy-margin products get deep stock and the lion's share of ad budget, because running out of those mid-event or under-funding them is the most expensive mistake you can make. The products that barely moved get normal inventory and little to no event budget — the traffic won't rescue them. The discount-dependent products get a decision: shrink the discount and see if the volume holds, or pull them from the event so they're not bleeding margin under the spotlight. You're not guessing what the market will do; you're acting on what your own market already did.

Account for the margin squeeze the event creates

Prime Day compresses margin from two sides at once — the discounts you run and the higher ad costs you pay to compete for spiking traffic. A clean post-mortem looks at *net* outcomes with both folded in, not gross revenue. The goal of the debrief isn't to celebrate the biggest sales day; it's to find the most *profitable* plan for next time, and those two numbers routinely point in different directions. If your review only tells you what sold, you've learned half of what the data was willing to give you. The other half — what actually made money after the discount and the pricier clicks — is the part that makes next year better instead of just busier.

Frequently asked questions

When should I run the post-mortem?

As soon as the event data settles — within a couple of weeks, while the numbers and your memory of what you tried are both fresh. Waiting until next year's prep means reconstructing decisions from incomplete recollection, which is exactly the blank-page problem the debrief is meant to solve.

What if my margins were too thin to tell winners from donations?

That's itself a finding: it means your discount depth was set without a clear margin floor. For the next event, set each SKU's maximum discount from its real net margin so there's a hard line below which you won't go, and the winner-versus-donation question answers itself in the moment instead of in the post-mortem.

Do category benchmarks matter, or just my own data?

Your own data wins every time it exists, because it reflects your exact products, pricing, and audience. Category patterns are a useful fallback for brand-new SKUs with no event history, but for anything you've sold through a prior Prime Day, trust your numbers over generic benchmarks.