Retargeting has become the automatic answer to cart abandonment: a banner that follows whoever left, for three weeks, wherever they go. It works, but it treats a symptom, not a cause. And most of the real causes have nothing to do with how persistent the ads are afterward.
The people who have studied this the longest are at the Baymard Institute, which has spent years measuring real checkout behavior across hundreds of sites. Their average across 50 studies puts the abandonment rate at 70.22%. Their findings tell a different story from the one in most marketing blog posts: most abandonment gets decided before the last click, in design choices someone else made months earlier.
Which abandonment causes can you actually fix?
First: costs that only show up at the end. This is the single most cited cause: 40% of abandonments once you set aside people who were just browsing, according to Baymard's latest survey. Shipping, tax, payment fees: if the real total only appears on the last screen, people feel misled, even when nobody meant to mislead them. Fix it by showing an estimated total in the cart itself, not only in the final screen.
Second: the forced account. Around a quarter of abandonment: in Baymard's 2022 survey, 24% of US online shoppers had dropped at least one cart only because an account was required. The reason is the same every time: someone who wants to buy one thing does not want to sign up for a service. Guest checkout should be the default, not an option hidden under the login button.
Third: a checkout that is too long or confusing. Every extra field is one more place someone can stop and never come back. The practical rule is blunt but true: if a field is not needed to ship the order or take the payment, it should not be there. In 2024 the average checkout Baymard measured had 11.3 form fields. Most sites need 8.
Fourth: low trust in payment security. Missing logos, a site that looks dated, no visible HTTPS: small signals like these are enough to make someone with a card already in hand start doubting.
Fifth: a missing payment method, or a declined card. Someone who had a specific payment method in mind and cannot find it almost never switches to another one. They close the page.
Sixth: delivery that is too slow, or no faster option offered. Not everyone wants next-day delivery, but everyone wants to know the timing before the last step, not discover it there.
Checkout does not convert because it persuades. It converts because it gives no reason to stop.
The seventh cause, the one that cannot be fixed
A huge share of abandoned carts belongs to people who were only browsing: in the same Baymard survey, 42% of US online shoppers said they had abandoned a cart because they were "just browsing / not ready to buy". Comparing prices, saving something for later, waiting for payday. They never intended to buy on that visit, and no checkout on earth would have stopped them there.
It is the part almost no blog post says out loud, because it ruins the promise in the headline. An agency that promises to "eliminate cart abandonment" is selling a number that will never exist, because part of that number was never a problem to begin with. We learned this working across different online stores: the first useful step is not reducing abandonment, it is splitting it apart. Separating carts with valid contact details from anonymous ones completely changes what is worth fixing, and which half is worth spending money on.
- AICart segmentation — separate carts that left an email from anonymous ones, and people who saw checkout from those who stopped before it
- NOIReal checkout diagnosis — try to buy something the way a customer would, not just read the dashboard
- AITechnical fixes — visible total upfront, guest checkout, local payment methods
- NOIDecide what actually deserves retargeting — only carts with real intent, not all traffic
- AITargeted recovery sequence — email or messages only to people who left contact details, with a real reason to come back
Segmentation and the recovery sequence are the two steps an AI agent handles well without constant supervision, in the sense described in our guide to AI agents for marketing: it reads the carts every day and routes the ones that left contact details. Deciding which fixes to make on checkout itself stays a job for people looking at the site with a critical eye.
What should you check before changing anything?
| Metric | What it actually tells you |
|---|---|
| Overall abandonment rate | Almost nothing on its own: it mixes browsers with real customers |
| Abandonment on carts with a captured email | The closest thing to a problem worth fixing |
| Drop-off between checkout screens | Where people who actually wanted to buy are getting lost |
| Recovery rate on post-abandonment emails | Whether the recovery sequence is working, or just existing |
The third row is what a GA4 funnel exploration shows you: how many people make it from one checkout screen to the next. The left column is what almost everyone checks first. The right column is what decides whether a change is actually worth making. An abandonment rate that drops from 75% to 70% says nothing on its own: you need to know whether it fell because checkout works better, or only because more traffic arrived from people who already intended to buy.
How long before you see an improvement
Technical fixes — a visible total, guest checkout, one more payment method — show up almost right away, usually within the first few weeks after launch, because they act on a behavior that happens many times a day. A post-abandonment email sequence takes longer to judge fairly, because you need a minimum volume of carts with a captured contact before you can tell whether it converts or just sits in the inbox. The method stays the same as everywhere else on this blog: you measure, you do not assume. We go deeper on that loop in our guide to vibe marketing, where the produce-measure-fix cycle is the same one, just pointed at a checkout instead of a campaign.
Frequently asked questions
What is the single most common cause of cart abandonment?
According to Baymard Institute measurements, it is discovering extra costs only at the very last step of checkout: shipping, tax, or fees not shown earlier. It gets fixed by showing an estimated total in the cart page itself, not only at the end.
Can cart abandonment be eliminated entirely?
No, and anyone who promises that is selling an impossible outcome. A large share of abandonment comes from people who were only comparing or browsing with no real intent to buy at that moment. It is not a technical problem, it is normal behavior.
Do I always need guest checkout, with no forced sign-up?
In the large majority of cases, yes. Forcing account creation before payment is one of the most consistently measured causes of abandonment, and the account can almost always be offered afterward, once the order is confirmed.
Does retargeting actually work to recover abandoned carts?
It works, but only on people who had real intent, and it should target the contact details you already captured, not all traffic that ever saw a product. Applied to everyone, it wastes budget on people who never intended to buy anyway.
How do I know if my checkout has a real problem?
Try it yourself, start to finish, with a real card, on desktop and on a phone. Looking only at the abandonment percentage on a dashboard does not show you where the path actually breaks: you need to see it through the same eyes as someone buying.
Losing more carts than expected?
Seven questions, and we tell you if the problem is checkout or just how the numbers are being read. Founders answer within 48 hours.
ApplyKeep reading
Founder of As Growth, an AI-first growth partner. He runs growth and strategy across partner businesses: OroEtic, Giojello, Equity Plus and others. Apply if you want to work with us.