"CVR is up, sessions are up, but revenue isn't growing the way I expected." When I look across ecommerce dashboards, this disconnect almost always traces back to one neglected metric — and once you find it, it traces back to a second insight: the metric is misleading on its own.
That metric is Average Order Value (AOV). And the trap is that AOV improvements often cancel out with CVR drops, leaving the per-session revenue (RPS) flat or worse.
Here's the short version of how I think about AOV after watching a few too many "AOV up but revenue flat" incidents.
TL;DR
- AOV is one of three factors in Revenue = Sessions × CVR × AOV. Tracking AOV alone hides the cancellation effects with CVR and sessions.
- AOV improvements break into 10 tactic categories — cross-sell, upsell, bundle, free-shipping threshold, membership, quantity discount, price revision, personalization, cart-recovery, post-purchase upsell. Expected impact varies sharply by tactic.
- The "raise free-shipping threshold → AOV up, CVR down" trap is real. Use AOV × CVR = RPS (Revenue per Session) as the decision axis to see whether a tactic actually moved the business forward.
AOV in the Revenue Decomposition Formula
AOV is "average revenue per order" — total revenue divided by number of orders for the same period. Simple definition, but its position in the revenue formula is what makes it tricky.
Revenue = Sessions × CVR × AOV. Three factors, all interacting. A campaign that "raised acquisition" might have lifted sessions while dropping CVR. A free-shipping threshold raise might have lifted AOV while dropping CVR for shoppers who didn't reach the new threshold. Without separating the three, you can't tell which factor moved.
Japan's BtoC ecommerce (physical goods) market reached ¥15.22 trillion (~$100B at current rates) in 2024 with 9.78% EC penetration. The need to decompose revenue grows with market size — and Japan's market is now mature enough that "we're growing because the market is growing" stops being a sufficient explanation.
Calculating AOV: Four Implementation Pitfalls
Shopify's official documentation defines AOV as "gross sales minus discounts, divided by number of orders" — i.e., net sales after discounts, divided by completed orders. This definition exposes four pitfalls when you implement it on a real dashboard:
- Pre- vs post-discount. Heavy coupon use overstates AOV when computed from gross sales.
- Tax-inclusive vs exclusive. Mixed definitions across teams produce conflicting numbers in the same room.
- Shipping included vs excluded. Free-shipping-threshold campaigns become impossible to measure cleanly when shipping is mixed in.
- Refund / cancel timing. In categories with >10% return rates, dashboard AOV and post-close AOV diverge by 20–30%.
Item 4 is the silent killer. A safe approach: track both "AOV at order time" and "AOV at confirmation," and never debate strategy from the order-time view alone.
The global AOV benchmark Shopify cites is approximately $145 across industries. The 2024 US holiday season hit a record $241.4 billion online (up 8.7% YoY) — benchmarks shift every year, so re-base annually for your category and region.
10 Tactics to Increase AOV
AOV improvement consolidates into 10 tactic categories. Shopify's official guide covers 7; WooCommerce adds free-shipping, memberships, and payment plans. Combined, here are the 10 with expected impact ranges:
A few caveats on reading these ranges:
- These are median ranges, not best cases. Cross-sell at +10–30% can land at +0–5% with sloppy recommendation logic, or above +30% with strong personalization.
- McKinsey's January 2025 research on personalized marketing reports +1–2% revenue / +1–3% margin from targeted promotions, with individual companies seeing +5–25% when execution and data infrastructure are mature.
- Running all 10 tactics simultaneously creates interference effects that obscure each one's contribution. The most effective starting set is free-shipping threshold (#4) + cross-sell (#1) + bundle (#3) — low implementation cost, easy to measure, quick to validate the AOV improvement direction.
Why AOV Alone Is Misleading: The CVR/RPS Trap
The classic trap: raise the free-shipping threshold from $50 to $100, AOV goes up, but CVR drops for shoppers who don't reach the new threshold. The result is "AOV up × CVR down," and per-session revenue is actually lower.
The target quadrant is top-right (high AOV × high CVR). Tactics that push you into the top-left (high AOV × low CVR) look like wins on the AOV chart but are net-negative for the business. The decision metric that exposes this is RPS (Revenue per Session):
RPS = Revenue ÷ Sessions = CVR × AOV
RPS expresses "how much revenue per visitor" in a single number. If AOV is up but RPS is down, the tactic failed — full stop. ROAS analysis hides the same trap: the "Revenue" in ROAS = Revenue ÷ Spend decomposes into AOV × CVR × Sessions, so AOV-only thinking blinds you to which factor drove the ROAS movement.
Once a team adopts RPS as the primary AOV-related decision metric, the conversation about "should we raise the threshold" gets shorter. The answer is no longer "AOV went up so it worked." The answer is "what happened to RPS."
Phase-Based Prioritization
The order of tactic deployment depends on business phase:
- Acquisition phase: don't sacrifice CVR. Use passive tactics that don't disrupt purchase intent — cross-sell and post-purchase upsell. A high free-shipping threshold raises the first-purchase barrier and drops CVR.
- Scale phase: deploy active tactics. Free-shipping threshold, bundle, upsell. Accept some CVR drop, A/B test each, judge by net RPS lift.
- Mature phase: individual optimization. Membership and personalization. McKinsey reports +5–25% revenue ranges for AI/gen-AI personalization at the company level — but only when data infrastructure is mature. Personalization on weak data introduces recommendation noise that drops CVR.
Closing — Use AOV as One Lever, Not the Only Lever
AOV is one of three factors in Revenue = Sessions × CVR × AOV. The formula is simple, but discount/tax/shipping/refund treatment branches the implementation in non-obvious ways. Ten improvement tactics exist; the realistic starting set is free-shipping threshold + cross-sell + bundle. AOV × CVR = RPS exposes the true impact of any AOV tactic.
The mental model shift that helps the most: stop asking "did AOV go up?" Start asking "did RPS go up?"
What metrics do you watch when judging an AOV-improvement campaign? Curious whether anyone is using RPS (or a similar per-session revenue metric) as a primary decision axis.