Your cart abandonment rate will never tell you what to fix.

A number nobody has ever moved

The Baymard Institute aggregates roughly fifty independent studies on cart abandonment. Its average sits at 70.22 percent, and the interesting part is not the level — it is the flatness. The figure has hovered near 70 percent for more than a decade, through the arrival of one-click payment, mobile wallets and ten years of checkout redesigns.

When a measure stays put for a decade across an entire industry, it rarely describes a defect. It describes a property of the behaviour being observed. Yet the percentage keeps showing up in steering committees as a target to bring down, usually paired with a round number nobody can trace back to a specific action.

The metric is not the problem. The level of aggregation it gets read at is.

What the headline number blends together

Baymard's surveys on reasons for abandonment show that a substantial share of abandoned carts — more than four in ten among the US shoppers surveyed — comes from people who say they were simply browsing with no intent to buy. Comparing options, saving items, planning a later purchase: here the cart is a list, not a commitment.

No checkout improvement will convert that segment, and that is good news: roughly half the headline number carries no quality signal at all. Once it is set aside, the remaining reasons become legible — and they are not the ones most teams assume.

The useful reading fits in one sentence: the dominant fixable cause is price surprise, not form friction. That is almost the reverse of the priorities written into most redesign plans, where the effort goes into counting fields and collapsing steps.

A perfectly designed checkout does not rescue a price that changes its mind on the last screen.

Measure moments instead of mapping journeys

The standard response to this is a journey map: workshops, personas, forty documented moments on a very large wall. These exercises produce alignment between teams, which has real value. They rarely produce a decision, because they do not rank what anything actually costs.

A leaner approach is to pick three moments, instrument them properly, and knowingly leave the rest unmeasured. Which three depends on the business model, but the logic holds: instrument wherever someone learns something they did not know, because that is where decisions change.

• The moment total cost first becomes visible, shipping and tax included.
• The moment a constraint appears: account required, delivery window, missing payment method, partial stock.
• The moment an error occurs and the person has to choose between starting over and leaving.

At those three points only one question matters: how many people arrive here, and how many continue? The rest of the journey can stay approximate without consequence.

Decide on money, not on percentages

Once those measures exist, the trade-off changes shape. You stop comparing rates and start comparing amounts: traffic at that moment, times the share who stop, times average order value. That single conversion removes most of the opinion from the room.

It also surfaces an uncomfortable category: fixes that are not digital at all. Shipping fees perceived as excessive are fixed by commercial policy. A delivery window that is too long is fixed by a logistics agreement. Mandatory account creation is fixed by a business decision, usually defended by a team that values the data it collects.

That is the most useful outcome of the exercise: it moves part of the problem outside the digital team, toward the people who can genuinely settle it. A checkout redesign that skips this reallocation will improve the experience without moving revenue — and the headline rate will sit stubbornly at 70 percent.