In short
- Around 70% of baskets are abandoned, all sectors combined: that is the average established by a meta-analysis of 49 studies.
- The leading cause is discovering the extra costs at payment: shipping, taxes and fees together account for 40% of abandonments, once you set aside the ones that are just browsing. A separate reason adds to it, being unable to see the total cost upfront, at 12%.
- Email recovery is highly profitable and comes third. It recovers part of the abandonments you failed to prevent, which is not the same as preventing them.
- A large share of abandonments are not abandonments. Comparing prices, saving for later, coming back on another device: these journeys are normal and cannot be “recovered”.
The abandoned basket flow is, rightly, the first one you install when you start with automation. It is simple, it is measurable, and it produces revenue within days. That is exactly why it hides a more profitable problem.
If four in ten of the visitors who genuinely give up leave on discovering the extra costs at payment, sending them an email two hours later amounts to giving them another appointment at the same closed door. For an order shipped outside the Union, the harshest version is customs duty collected on delivery: the customer discovers the bill after paying, and it is the parcel they refuse.
What the figures say, and what they do not
The average abandonment rate sits around 70% across sectors, a figure drawn from a meta-analysis of 49 studies. Once you set aside the baskets that are simply browsing, the leading cause of abandonment is discovering the costs at payment, shipping, taxes and fees combined, at 40%. The separate reason “could not see the total cost upfront” adds another 12%, and it points at the same screen and the same fix.
Two reading precautions, without which these figures mislead.
Not every abandonment is a failure. A basket also serves as a wish list, a comparison tool and a reminder. Part of abandoned baskets correspond to normal journeys, and aiming for a very low rate means fighting how people behave.
Mobile is not the problem. What gets blamed on it points at a checkout that is harder to get through on a small screen: long forms, badly typed fields, costs appearing after three screens. The cause is the same, it simply shows up faster.
The mistake that costs the most Installing recovery before looking at where people leave. The flow will produce revenue, which will validate the decision, and it will kill any appetite for fixing the cause. You end up paying monthly for emails that compensate for a display problem you could fix once.
The three levers, in the order they pay
- Announce shipping costs before the checkout. On the product page and in the basket, not at the payment step. A customer who knows the full cost from the start does not abandon at the last moment: either she continues, or she leaves earlier without having invested mental effort. The free-shipping threshold, if you have one, should be shown with the amount still to reach.
- Shorten and harden the checkout. Ordering without a mandatory account, correctly typed fields on mobile, explicit errors, the payment methods your customers expect. That is the work described in our article on journey friction.
- Recover the rest. This is where the flow earns its place: it handles residual abandonments, not structural ones.
Ten-minute self-check: place an order on your own site, on your phone, noting at which screen the total shipping cost first becomes visible. If it is after the delivery address, you know your priority.
The flow, once you get there
Three messages are enough, and the third is the most debatable.
| Message | When | What it should contain |
|---|---|---|
| The reminder | One to four hours later | The basket, with images, and a one-click return. No discount. Many people were simply interrupted |
| The objection handler | Twenty-four hours later | What actually blocks: delivery times, return conditions, payment methods, availability. Still no discount |
| The incentive | Forty-eight to seventy-two hours | Send to part of your contacts only, and measure. A systematic discount teaches your customers to abandon their basket |
The discount point deserves emphasis, because it is almost always decided the wrong way. A reduction sent to every abandoned basket produces a visible immediate gain and an invisible learning effect: part of your regular customers will end up systematically abandoning their basket to trigger the code. The cost of that habit appears in no dashboard, exactly like the cost of returns.
The same flow can run on WhatsApp, with a difference email does not have: each reminder goes out as a marketing template charged per send, and from 1 October 2026 the replies exchanged with the customer who comes back are charged too beyond 1,000 a month. The revenue recovered is then set against an invoice, not just a click rate.
Two underlying conditions, finally. Recovery assumes an address collected with valid consent, which points back to building a genuinely consenting base. And it assumes abandonment is correctly detected, which is not always the case: a checkout whose events are badly reported produces recovery emails on orders already placed, which is the surest way to lose a customer’s trust.
The options, and what they are worth
| Option | What you get | Who it suits |
|---|---|---|
| Show costs upfront | Addresses the most frequent cause, costs only a display change, and the effect shows on checkout completion | Everyone, first |
| Simplify the mobile checkout | The biggest reserve when mobile carries most of the traffic, which it does almost everywhere | Mobile-majority shops |
| Three-message recovery without blanket discounting | Highly profitable, and no learning effect. The default structure | As soon as a consenting base exists |
| Systematic discount in the first message | Immediate gain, eroded margin, and a durable habit among your best customers | No case in steady state |
Tooling
The technical need is modest and precise: finding a basket again several days later, from another device, without forcing the person to log back in. That is what separates a pretty email from one that converts.
On PrestaShop we deploy the module published by Datafirefly Limited, our agency’s sister company: Save Cart keeps the basket and provides a one-click resume link. One-off purchase · 12 months of updates.
Method beats tooling, and this is the clearest case in the whole series: no module makes shipping costs discovered too late disappear. It catches what the checkout let slip, and it is better off having little to catch.
What this says more broadly
Automation appeals because it produces a measurable result quickly, and that is also what makes it a trap. It measures what it recovers, never what you could have avoided losing.
So the right question in front of a flow is not “how much does it bring in”, but “what problem is it compensating for, and is that problem fixable”. Framed that way, it surfaces ergonomics and display work that costs less and lasts longer. It is how we approach our marketing automation practice, and why we rarely start by installing a flow.
Sources
The figures quoted in this article point to the study that produced them, opened on 11 August 2026. We do not cite a source we have not read.
- Baymard Institute, Cart Abandonment Rate Statistics, reading of 22 September 2025, meta-analysis of 49 studies published between 2006 and 2025, and survey of abandonment reasons among US online shoppers. Read it
FAQ
What abandonment rate should I aim for?
No absolute target makes sense, since part of abandonments correspond to normal use of the basket. Track checkout completion step by step instead, and look for the tallest step: that is the one you fix.
How many recovery emails?
Three is enough in the vast majority of cases: a quick reminder, an objection handler, and an incentive reserved for a segment. Beyond that the marginal gain becomes negligible and the unsubscribe cost does not.
Should the first message carry a discount?
No. Many abandonments are simple interruptions, and in that case the discount is handed to customers who would have come back anyway. Keep it for the last message, on a segment, and measure.
Can I email someone who has not given marketing consent?
Worth putting to your own counsel, and the practical answer is simpler: collect consent at the right moment in the checkout, with wording people understand. A consenting base beats a wide, fragile one.
Can Dotsland help?
Yes. Analysing the checkout step by step to find the real step, fixing how costs are displayed, then building the recovery flow and measuring what it genuinely adds. It is part of our marketing automation practice. Let’s talk, or start by ordering from your own phone.