Somewhere right now, a founder is being told their abandoned cart flow is underperforming because it recovers eight percent and "the industry average is twelve." Nobody in that room can say which businesses made up that average, what they sold, what they charged, or where their traffic came from. The number gets treated like a law of physics anyway, and a quarter of work gets pointed at closing a gap that may not exist.
Here is the problem with averages in ecommerce. An average is a single number standing in for an enormous range of businesses that share almost nothing operationally. A furniture brand with a four hundred dollar average order value and a supplement brand at twenty-eight dollars are not doing the same thing when a customer abandons. The furniture buyer is running a considered purchase over days or weeks, comparing across three tabs, waiting for a partner to weigh in, checking whether the sofa fits through a doorway. The supplement buyer got distracted. One of those carts recovers with information and reassurance over a longer window. The other recovers with a reminder inside an hour. Pooling them produces a number that describes neither.
Traffic mix distorts it further. A brand whose sessions come mostly from branded search and email has visitors who already decided to buy and were interrupted. A brand buying cold prospecting traffic has carts built by people who were browsing, not shopping. Same event name in the analytics platform, completely different intent behind it. Recovery rate is downstream of intent, and intent is downstream of where the traffic came from, which is why two brands running identical flows can post recovery rates that differ by a factor of three without either of them doing anything wrong.
The only benchmark that carries real information is your own. What did this flow do last quarter, on comparable traffic, at a comparable discount posture, with comparable product mix? That comparison controls for everything an industry average cannot. If you want the broader landscape for context, the published research on abandoned cart email benchmarks is genuinely useful for understanding which metrics deserve attention and roughly how often people abandon in the first place. Baymard Institute's work puts cart abandonment around seventy percent across the board, which is worth internalizing as a fact about human behavior rather than a target to beat. Use outside numbers to decide what to measure. Use your own history to decide whether you are improving.
Now the uncomfortable part. There is a fast, reliable way to make cart flow numbers look excellent, and most brands reach for it without thinking. Lead with a discount. Put fifteen percent off in the first email and recovery rate climbs immediately. It works because it is supposed to work, and it keeps working right up until it becomes the reason your customers behave the way they do.
Think about what you have actually built. Your most engaged buyers, the ones who visit often and know your catalog, learn quickly that adding to cart and waiting produces a code. They are not gaming you. They are responding correctly to the incentive you designed. You bought a metric and sold a margin, and the trade compounds, because the same people would have bought at full price and now never will. The flow looks like it is generating revenue. Some meaningful share of that revenue was already yours.
If you want a cleaner read, change what you compare. Track revenue per recipient rather than recovery rate, because it prices in discount depth instead of hiding it. Track full-price recovery separately from discounted recovery. Watch what happens to first-email performance when the incentive moves later in the sequence rather than earlier.
Then make the split most brands never make. Cart abandoners and checkout abandoners are not the same audience. Someone who left at the cart is still deciding whether they want the thing. Someone who reached the shipping screen and stopped has usually hit a specific friction point: a shipping cost that appeared late, a delivery window that missed an occasion, a payment method you do not accept, a returns policy they could not find. Those are objections, not hesitation, and they are answerable with copy rather than money. Splitting them lets your abandoned cart flows speak to an actual reason instead of shouting a generic reminder at two groups with nothing in common.
Start with one question about your own data. What percentage of your cart flow revenue arrived with a discount code attached, and what would that revenue have been if you had never sent one?
