Skip to article content

A WooCommerce checkout is failing: what to review before you change anything

If a customer cannot complete a WooCommerce purchase, start with the step and message they see. Learn what to review with your team without duplicating charges or changing real orders.

What is actually happening before you change anything

When the message arrives that a customer «cannot pay on the site», the first step is to understand what that person is seeing, not to open the admin and disable plugins. In WooCommerce (the shop plugin for WordPress) a purchase goes through several stages, and the complaint reaches you compressed into one sentence that does not say which one broke.

Before touching anything, ask for concrete information: which step stalled, what message appeared, which browser or device and roughly when. A screenshot if possible. That note saves hours of blind testing.

Before repeating a payment, ask the team to check whether an order or payment transaction already exists. Reproduce the fault in an isolated copy using the payment provider’s test mode. A test account or private window alone does not prevent real charges, emails or stock changes.

And separate three complaints that often come mixed, because each points to a different place:

  • «It will not let me pay». Cannot complete the form or the button does not respond.
  • «The confirmation never arrived». They paid, but have not received the email.
  • «My order is not in the panel». The gateway charged, but the order was not recorded in WooCommerce.

Document date, URL, reproduced steps and screenshots before changing anything. That record is the base you will decide on later, and it stops you tripping over your own changes.

Cart, checkout and confirmation are not the same step

Separate what the customer sees from what the store and payment provider record. Not every gateway follows exactly the same sequence:

  • Cart and checkout form. The customer selects products, address, shipping and payment. Record where progress stops and the message shown.
  • WooCommerce order. An order can exist before payment is confirmed. Its presence does not prove it has been paid: check its status and notes.
  • Authorisation and payment. The gateway processes payment and may request extra verification. Depending on the integration, this happens on another screen or within the page.
  • Result notification. The store receives the result using that gateway’s mechanism, such as a notification or webhook: an automatic call between services. The browser’s return and this notification are different things.
  • Customer confirmation. The screen and email should reflect the actual state. A missing email does not necessarily mean payment failed.

WooCommerce’s order status documentation explains the record. Compare it with the payment provider’s transaction before repeating a test or updating the customer.

Look at what has moved on the site in the last few days

Many checkouts that «suddenly stop working» have a prosaic explanation: someone moved something recently. Before diving into logs, review the recent history:

  • Were WordPress, WooCommerce, the theme or any major plugin updated? Gateway and shipping plugin updates are especially sensitive.
  • Did the SSL certificate expire or renew? Did you change domain or server?
  • Has a caching layer, a CDN (a network that distributes content) or a WAF (application firewall) been added? Without fine-tuning for WooCommerce, these layers interfere with the cart and the checkout.
  • Has a new external integration been connected (CRM, ERP, shipping, stock sync, marketing)? A change that looks unrelated may be the real culprit.

If the trigger looks like an update, revisit the sequence we describe to react when WordPress stops working properly after an update; if the candidates are connected services, review how third-party tools can slow down or condition a site.

The rule: do not pile a new change on top of a recent one. If there was an update three days ago, the first hypothesis is not «I install another plugin», but «I undo the change in a test environment and see whether it works again».

Diagnosis: visual fault, code error or payment refusal

With the step pinned down and the recent changes in sight, the next plane is diagnosis: what kind of fault is it? Confusing the family is the fast route to changing things that do not need changing.

Three families of fault and how to tell them apart
FamilyHow to recognise itWhere it is usually diagnosed
Visual or JavaScript faultThe button does not respond, fields overlap, the region dropdown does not load, scrolling locks up on mobile, the total does not update when shipping changes.Browser console (F12), tests in a clean browser and on mobile, review of the theme and plugins that add scripts to the checkout.
Server or code errorBlank screen when pressing «Place order», 500 error, order not saved, email not sent, generic WooCommerce message without detail.WordPress logs (debug.log), server error_log, the Site Health screen, status of the email queue.
Payment or payment-notification issueSpecific message from the gateway («card declined»), redirect that does not return, customer pays but the order stays «pending payment».Gateway dashboard (Stripe, Redsys, PayPal…), transaction log, webhook logs, status of the certificate and the API credentials.

Each family uses different clues and different tools. If you tackle a visual fault by reading gateway logs, you find nothing. If you tackle a gateway refusal by disabling plugins, you add noise without solving it. Ten minutes placing the fault in one of these boxes before acting is the step that makes the difference.

What the logs and the status report are telling you

Logs are the records where WordPress, WooCommerce and the server write what happens. They are more honest than guesswork.

Start with logs available under WooCommerce → Status → Logs and the payment plugin’s logs if provided. If the team needs additional WordPress debugging, it should be set up without showing errors to visitors and with protected log files. The WordPress documentation explains those options.

What to look for:

  • The plugin or file name in the trace. Many lines include a path like /wp-content/plugins/plugin-name/...; that path tells you where to start.
  • The PHP line number. It helps you brief your technical team or the plugin’s support.
  • HTTP codes and the operation’s message. They describe a request’s response, but you need to identify which service answered and read the response body. A 200 does not prove payment was collected; a 500 alone does not identify the responsible provider. MDN’s reference explains the codes.
  • Repetition of the same line. If an error repeats every few seconds, it is often a webhook not being delivered or a scheduled task colliding with the order.

Under WooCommerce → Status you can also obtain the System Status Report, including versions and configuration useful to support. Review its contents and share it privately. WordPress’s Site Health screen may provide other clues, but it does not establish that checkout, payment or email delivery works.

Read them calmly. A log has plenty of noise, and not everything is the cause. Look for messages that match in time with the test and with the plugin you suspect.

Three routes of action depending on what you see

After reviewing the fault and the logs, the team can decide what to change and how to check the result. These are three common options; the choice depends on the confirmed cause and the risks of each change:

Routes of action by diagnosis
RouteWhen it fitsWhat to verify afterwards
Isolate the conflict in a test environmentThe fault is reproducible and the logs point to a plugin or the theme. You have a staging environment (a copy of the site for testing). Never disable plugins one by one in production during trading hours.Disable plugins in groups, switch to a default theme and repeat the purchase. Once you identify the culprit, reactivate the rest and decide whether to replace or report the plugin to its author.
Roll back the recent change from a backupSymptoms appeared right after an update or new activation, and a previous verified backup is available.Test first in an isolated copy. Before restoring the live store, plan how to preserve and reconcile orders, payments and stock changes made after the backup.
Escalate to gateway or hosting supportLogs show HTTP errors from the gateway (401, 403, 500), certificate issues, webhooks not delivered, or server failures your team does not control.Open a ticket with reproducible evidence: URL, exact time, test user, screenshots and a log extract. While waiting, avoid changes that contaminate the diagnosis.

Two criteria: where the evidence points and who controls that stretch. If what breaks is in your plugin, it is yours; if it is in the gateway or the server, you will not fix it. A universal recipe («disable all plugins», «change the gateway», «raise the memory limit») without diagnosis creates more problems than it solves.

When the problem sits in the payment gateway

If the logs lead you to the gateway, separate your side from theirs. The gateway is an external service that charges on your behalf and reports the result. Common causes:

  • An expired certificate or a secure-connection problem. If your site’s SSL or the return endpoint is invalid, the gateway refuses the communication.
  • Expired or rotated API credentials. The keys your plugin uses are refreshed occasionally. If the gateway warned and no one applied it, the channel drops.
  • Confusion between test and live mode. Someone leaves sandbox mode active and real orders are not charged.
  • Webhook not delivered. The background notice does not arrive, usually because a WAF or CDN blocks it, or the public URL changed after a migration. The order stays «pending payment» even though the charge went through.
  • Antifraud policy changes. The gateway tightens criteria or requires strong authentication (3-D Secure) that your plugin was not set up to handle.

Use the gateway’s documented test mode in an isolated copy and check email, inventory and connected services too. If no test mode is available, agree a controlled procedure with the provider and technical team before making any live transactions.

The payment account owner, the team configuring the plugin and the payment provider may need to coordinate, depending on the fault. Supply the transaction reference and relevant messages to support without sending API secrets, full card details or unnecessary personal data.

Fictional example: orders remain pending although the payment provider records successful payments. The team finds a security rule rejecting payment notifications. It reviews the rule for that specific endpoint, retains the provider’s required authenticity checks and reconciles each order with its transaction. This does not justify allowing all traffic from an IP or marking orders paid without verification.

Keep it from quietly breaking again

A checkout that fails silently, with no one noticing until a customer complains, is a pattern worth stopping at the root. Sober routines:

  • A simple alert when orders stop coming in. If your shop usually receives orders every day or week, set up an email notice if an unusually long period passes with no new records.
  • A staging environment. Every update of WooCommerce, the gateway, the theme or critical plugins first goes through a mirror environment where you reproduce a full purchase.
  • A backup before each intervention. Before touching the checkout, take one more and verify it can be restored.
  • A short list of steps to retest after any change. Purchase with a test card, order in the panel, status change, email to the customer, copy to the administrator.
  • Periodic review of the Site Health report. It may highlight version or configuration warnings. It does not replace checkout tests or guarantee that a fault will be detected before it affects a customer.

If the button reacts slowly, the guide to Interaction to Next Paint (INP) explains the browser’s visual response. This metric does not measure all the time taken by the bank, gateway or server to finish the operation; those waits need separate checks.

On a live shop with continuous updates, zero risk does not exist. You do get finding out sooner, having evidence and being able to undo the last change without fighting your own history.

If you want us to walk you through it

A checkout that customers cannot complete is both a technical problem and a revenue problem, and it is easier handled with someone technical looking alongside you. If you prefer not to do it alone, tell us what your customer is seeing: step, browser, message and approximate time. From there we can guide you and, if a review fits, explain the work before starting.

If your intention is a broader review of the shop —response times, external integrations, logs and environment— starting with our performance area usually fits better. You will also find useful the guide to reacting when WordPress stops working properly after an update, a common cause behind a checkout stalling halfway.