Most checkout abandonment is created before the checkout page opens. A delivery cost that appears late, an item that turns out to be unavailable, a payment method the customer does not have.
This guide covers how to read the funnel by step and device rather than in total, what the declined payment log usually reveals, and why one change at a time is the only way to know what worked.
Key takeaways
- Read the funnel by step and device: The overall abandonment figure hides the one combination that is failing far worse than the rest.
- Most of it happens before checkout: A total that stops matching the first price is the most common cause and the cheapest to fix.
- Read the decline log: Many failures are configuration rather than design, and two fixes usually cover most of the loss.
- Change one thing at a time: A redesign that moves five things at once produces a number nobody can attribute.
Look for the business signal behind the technology
The signal is in the funnel by step and by device, not in the overall abandonment figure. One step on one device usually accounts for a disproportionate share, and it is rarely the step the team expected when they started looking.
Not all abandonment is a problem to fix. People compare prices, save baskets for payday and shop from two devices, and a basket left open overnight is often a purchase that completes the next morning. Separate those from the customers who tried and failed.
Read the declined payment log for a month and group the causes. Many are configuration rather than design: a card type, a currency, an address check set too strictly. Those are fixable in settings and measurable within days.
Use four lenses to understand the opportunity
Pricing
Find the point at which the total stops matching the price the customer first saw. Delivery, tax or a surcharge appearing late is the most common cause of abandonment, and moving it earlier costs nothing but a layout change.
Inventory
Decide what a customer is told when an item is unavailable and whether they find out before or after entering their details. Discovering an out-of-stock item on the payment page is the worst possible moment and the easiest one to avoid.
Payment
Check which methods your customers actually use, what fails most often, and what the error message says to them. Many declines are configuration rather than design, and a generic failure message turns a fixable problem into a lost order.
Recovery
Work out what happens to a basket after a failed attempt and how the customer gets back to where they were. Recovery is where a shop either collects the order or teaches the customer to buy somewhere with fewer steps.
Assess the current state without buying software first
Pull step-level drop-off split by device and payment method before changing anything. The overall figure hides everything useful, and the split usually points at one combination that is failing far worse than the rest.
Decide what useful measurement looks like
Measure completed orders rather than step conversion alone. Moving delivery cost earlier will reduce conversion at the product page and increase it at checkout, and only the completed total tells you whether the trade worked.
Watch returning customers separately from new ones. A saved card hides problems that first-time buyers hit every day, and an average across both will understate how bad the first purchase experience is.
When software is likely to help
Development is justified where the cause is structural: a pricing rule the platform cannot express, an inventory check that happens too late, a payment method your customers use that is not offered. Those need building and they pay back quickly.
Where the cause is a surcharge revealed at the last step or an error message nobody has read, the fix is configuration and copy. Rebuilding the checkout to solve a problem that lives in a settings page is an expensive way to avoid reading the logs.
For an incentive-led approach to the same customers, worth weighing against fixing friction, see When Does a Loyalty App Make Sense for Your Business?.
Start with a bounded improvement
Change one thing, usually showing the full price earlier, and watch step-level drop-off for a full month. A single change with a clean before and after teaches more than a redesign that moves five things at once and cannot be attributed.
Read the opportunity from the outside in
Start from the moment the customer decides to buy rather than from the checkout page. What did they expect the price to be, what did they assume about delivery, and where did the expectation break? Then follow it inward through pricing, stock and payment until you find where the surprise is created.
Name the smallest change that would show up in completed orders: one cost shown earlier, one payment method added, one error message rewritten. That is a boundary you can measure in a month, which a checkout redesign is not.
Risks to make visible
| Lens | Common risk | Early response |
|---|---|---|
| Pricing | The total changes late | Move every unavoidable cost to the first place the customer sees the item. |
| Inventory | Stock is checked after the details | Tell the customer an item is unavailable before they start typing. |
| Payment | Declines are treated as customer error | Group the failure codes from one month and fix the two largest causes. |
| Recovery | A failed attempt loses the basket | Check what a customer sees after a decline and how they get back. |
A 30-day validation plan: Checkout friction
Days 1 to 5: read the funnel by step and by device. Pull drop-off for each checkout step separately, split by phone and desktop and by payment method. The overall abandonment figure hides everything useful. One step on one device is usually responsible for a disproportionate share, and it is rarely the step the team expected. Look at the same figures for returning customers separately, because a saved card hides problems that first-time buyers hit every day.
Days 6 to 15: read the failed payment log. Go through the declines and the errors from the last month and group them by cause. Many are configuration rather than design: a currency, a card type, an address check set too strictly. Fixing two of those is usually a settings change and it is measurable within days. Check what the customer actually sees when each one happens, since a generic failure message turns a fixable decline into an abandoned basket.
Days 16 to 30: show the full price earlier. Move delivery cost and any surcharge to the point the customer first sees the item, and watch the step-level drop-off again. If abandonment falls at the later step and rises slightly at the earlier one, that is the trade working as intended, and the completed orders are the number that matters. Leave the change in place for a full month, as the first week after a price presentation change is usually the least representative one.
A practical worksheet: Checkout friction
Most checkout abandonment is caused before the checkout screen opens. Work backwards from your own funnel data and complete a row for each place a customer learns something they did not expect.
| Decision area | What to record |
|---|---|
| Pricing | The point at which the total stops matching the price the customer first saw, usually delivery, tax, or a surcharge |
| Inventory | What a customer is told when an item is unavailable, and whether they find out before or after entering their details |
| Payment | Which methods your customers actually use, what fails most often, and what the error message says to them |
| Recovery | What happens to a basket after a failed attempt, and how the customer gets back to where they were |
| Measurement | Drop-off per step rather than overall abandonment, split by device and by payment method |
Read the failed payment logs before redesigning anything. In most shops a small number of causes account for the bulk of lost orders, and they are frequently fixable in configuration rather than in the interface.
Related guides
- If the budget conversation comes first, start with E-Commerce App Development: A Practical Planning Guide.
- There is a longer treatment of the same subject in Mobile Loyalty App Development: Features, Cost Drivers, and Roadmap.
Should Checkout Friction: Causes, Metrics, and Fixes become a platform project with integrations in scope, Discuss Your Commerce Platform.
Ali Boran Gazel