Numerator Day 2 update: Prime Day 2026 average order value drops to $46.89 from $57.12

What's happening

Numerator's Prime Day 2026 live tracker, refreshed June 24 at 4pm ET, now reports an average order size of $46.89, down from $57.12 at the same point in last year's event. That's an order-level number, distinct from the panel-level household-spend figure (down 16% to ~$89) that drove the Day 1 narrative. Roughly half of tracked households placed more than one order on Day 1, so the per-order compression is real even though basket frequency partially absorbed it.

The Adobe Day 1 print of $8.3B in total U.S. online sales (up 5.3% YoY) holds in parallel. More baskets, smaller per-basket. The honest read is that Amazon brought in slightly more orders at materially lower order value.

Why it matters for Amazon brand owners

A ~$10 drop in AOV is the difference between a healthy contribution margin and a break-even order on a lot of FBA categories. Variable order costs (pick, pack, ship, return reserve) do not scale linearly with order value, so a 17% drop in AOV typically translates to a sharper drop in per-order contribution. PPC budgets sized off 2025 AOV are now bidding past break-even on the smaller order they actually win today.

The second-order effect is on bundle and variation strategy. Shoppers are clearly defaulting to the single-pack and the smaller size. SKUs whose contribution math relied on a bundle attach are the most exposed. The brands that finish the event in the black are the ones that re-rank deal allocation by per-SKU contribution at the actual Day 1 to 2 AOV, not the planned AOV.

What to do today

1. 1.

   Recompute break-even ACoS at the Day 1 to 2 AOV

   Use the actual AOV your account is producing this week, not the AOV your annual plan was built on. Cap bids on the SKUs where the recomputed ACoS sits below your current bid in PPC analytics.

2. 2.

   Pause deal allocation on bundle SKUs that lost their attach

   If the 2-pack is selling like a 1-pack at the 1-pack discount, the deal is funding negative contribution. Flag bundles with collapsing attach via the same per-SKU P&L you use for everyday decisions.

3. 3.

   Watch unit velocity, not gross sales, for Day 3 and 4 inventory calls

   Gross sales held up; units are higher than they look in dollars. Plan stock-out risk off units shipped per hour, not dollars per hour, using daily Amazon analytics.

How Nova helps

Nova ships a per-SKU contribution view that updates daily, so the AOV compression Numerator is reporting flows straight through to per-order margin without a manual rebuild. FBA sellers can sort the deal list by current contribution and rebid PPC against the AOV they're actually producing, not the AOV the annual plan assumed.