Average Order Value Optimization: The Metric That Changes Your Paid Media Math

Average order value is the number most e-commerce brands optimize last. It's also the one that changes everything when you finally pay attention to it.

Here's the uncomfortable truth: if your blended AOV is £65 and your target CPA is £40, you're running on thin margins before you account for returns, fulfilment, and customer acquisition overhead. But if you lift AOV to £85 without changing conversion rate or traffic volume, your ROAS improves, your CPA headroom grows, and suddenly channels that looked unprofitable on paper become viable. The math is that direct. Yet most paid media teams spend their sprints testing audiences and creatives while the AOV problem sits untouched in a Looker dashboard.

Why AOV Gets Ignored

The honest reason is organisational. Paid media teams are measured on CPAs and ROAS — metrics driven by clicks, impressions, and spend efficiency. AOV is downstream of product page experience, merchandising decisions, and sometimes pricing strategy. When different teams own these levers, it falls through the gap.

There's also a measurement problem. Most attribution tools report AOV as a flat average, which obscures the variance. A brand doing £1M/month in revenue might have 20% of transactions driving £200+ AOVs while the rest cluster around £45. Knowing which customer segments, traffic sources, and product categories index toward high-AOV purchases is the starting point — and most brands don't have that view.

The Levers That Actually Move AOV

Not all AOV tactics are equal. The ones that consistently work without damaging conversion rate:

• Product bundling with a clear value narrative: Don't just put products together — anchor the bundle to a specific use case or outcome. "The full skincare routine" outperforms "buy 3, save 15%" in most beauty verticals because it sells an outcome, not a discount.
• Post-add-to-cart upsells: Offered after the commitment decision, these convert at 20–30% higher rates than pre-cart recommendations because the shopper's intent is already locked in.
• Free shipping thresholds set just above median AOV: If your median order is £62, a free shipping threshold at £75 nudges shoppers to add one more item without creating friction for the majority who are already above it.
• Curated cross-sells on the product page: "Complete the look" in fashion, "frequently bought together" in home — category-specific framing matters. Generic recommendation engines consistently underperform hand-curated pairings in controlled tests.

What doesn't work: mandatory product add-ons, opaque bundle pricing, or anything that forces the shopper to do mental arithmetic to understand the value. If it adds cognitive load, it kills the transaction before it starts.

Connecting AOV to Media Efficiency

A 15% AOV lift, without any change in traffic or conversion rate, is the equivalent of getting 15% more revenue from every pound of media spend you've already committed.

This is why average order value optimization deserves to sit inside a growth and measurement conversation, not just a merchandising one. When you know your AOV by traffic source, you can make smarter channel allocation decisions. A content affiliate that drives 70% higher AOV than your social retargeting channel might look worse on a last-click CPA basis — but its actual revenue contribution is substantially stronger.

The same logic applies to incrementality. High-AOV transactions are often driven by customers who were already in-market and close to purchase — meaning the marginal impact of your media on that conversion is lower than it appears. Testing AOV as a variable inside your measurement framework (not just total revenue) surfaces these nuances and changes how you read channel performance across the funnel.

Where to Start

Pull your last 90 days of transaction data and segment AOV by three variables: traffic source, product category, and device. In most cases, you'll find one or two patterns that explain the majority of the variance. High-AOV transactions tend to cluster around specific entry pages, specific referral sources, or specific product combinations — those are the threads worth pulling first.

From there, run one isolated AOV experiment per four-week cycle. Bundling, threshold shipping, and post-cart upsells each deserve their own clean test — mixing them makes the attribution read impossible. Measure revenue per session, not just AOV in isolation, because a tactic that lifts AOV by 20% while depressing conversion rate by 25% is a net negative regardless of how good the average basket looks.

The brands that compound growth aren't just acquiring more customers — they're extracting more value from the customers they already have. AOV is the most underused lever in that equation, and it doesn't cost an extra penny in media spend to start pulling it.