Skip to the content

Case study

Customers hit an error and leave: making reporting one press

On a client's online tool, a red error message gave people no quick way to tell the team, so most said nothing. Reporting is now one press, with the page carrying the details instead of the person.

Key takeaways

  • A customer who meets an error with no easy way to report it may close the tab, and the fault is still there for the next person.
  • Put the report button on the error message itself, so telling you costs one press and no change of page.
  • The page already knows the error, the page address and the failed requests, so let it carry them instead of asking the person to remember.
  • Keep a report to what the failure showed, with no customer input and no sign-in detail, and say so on the page.
  • Fold repeat reports of the same error into one ticket, or the button buries the team instead of helping it.

Most people who meet a red error message in an online tool close the tab instead of telling anyone, and the business hears about the fault late, second-hand, or never. On this client's tool, reporting a problem meant leaving the page, finding support and writing out what had happened, so most people did not. We put a single Report button on the error message itself, and it was live on the client's tool the same day they asked for it.

What an unreported error costs a business

Every unreported error is a person who tried, failed and may not come back, and a fault that stays in place for the next person.

The team learned of a fault late, through a second-hand message, or not at all. And when a report arrives without detail, someone has to ask which page it was, what was clicked and what the message said before any work can start — by which time the person may no longer remember.

Why the report never comes

Reporting a fault cost the person more effort than simply leaving, and the facts the team needed were gone the moment the message closed.

To report it, someone had to leave what they were doing, open support, write what had happened and guess which details mattered. That is a lot to ask of a person in the middle of something else. A report that arrived without details gave the team little to act on.

What our engineers found about customers not reporting website errors

Almost nothing needs to be asked of the person, because the page already holds what the team needs.

We listed what a useful report requires and what a person can reasonably be asked to give, and the two lists barely overlapped. The error message, the page, the item it was about, the failed requests to the server behind it with their statuses and messages, and the browser were all known at the moment of failure. The page simply kept none of it. The turning point was to stop asking people to describe the error and let the page remember the last few failures itself, so one press could send them.

What the tool does now

Every red error message in the tool carries a Report button, and every signed-in page has a small one in the corner.

  • One press opens a support ticket — one item in the team's support queue — with the error message, the page and the item it was about, and the team is told at once in its own group, as with any ticket.
  • The page keeps the last ten failed requests to the server and page errors of that tab — the address, the status and the message only, never what was sent and never any sign-in detail — and the button sends those with the browser.
  • Words are optional. After sending, the person can add a line, and they can follow the ticket and the team's answer on the Support page.
  • The same error message reported again on the same page within half an hour joins the ticket already open, and reports count against the usual ticket limits, so the button cannot flood the team.

We tested it: a press opens a ticket, a repeat joins the open ticket, and a report from a page with no error message still files by its page. What the report carries is cut to size and never includes another customer's material. The tool's help guide now has a short entry on reporting a problem in one press. We have not measured how many reports it brings or how much faster faults are found, so we make no claim about either.

What you can check yourself

  • Try something that will fail harmlessly on your own website or booking page, such as sending a form with a required field empty, and read what the person is shown. Is there a way to tell you from that very screen?
  • Walk through reporting a problem the way a customer would. If it means leaving the page and typing a description, expect to hear about faults late.
  • Read the last few problem reports you received. Do they name the page, what was clicked and what the error said, or did you have to ask?
  • Ask whoever looks after your site whether your pages keep any record of the errors people meet, and whether that record could travel with a report.
  • Before a report button is added, ask exactly what it would send and whether the page says so in plain words.

Questions people ask

How do I find out when customers hit errors on my website if they never tell me?

You cannot rely on people telling you, so the page has to do the telling. Put a way to report the problem on the error message itself, so it can be sent without leaving what the person was doing. Then ask whoever looks after your site whether the page already records the failures behind that message, because that record is most of what you need to fix it.

Should I ask customers to fill in a form when something goes wrong?

A form is a barrier at the worst moment, when the person is already frustrated. Ask for one press and let the page supply the details, with any words optional afterwards. If you must ask something, keep it to what the page cannot know, such as what they were trying to do.

What information should an error report send to my team?

The error message itself, the address of the page, the item on it the person was working with, the recent failed requests with their statuses and messages, and the browser. Leave out what the person typed, any file they uploaded and anything about their sign-in. Say plainly on the page what a report contains, so nobody has to guess.

Will a report button flood my team with tickets?

It can if every press starts a fresh ticket, because one fault can be met by many people. The usual guard is to fold a repeat of the same error on the same page into the ticket already open, and to count reports against the same limits as any other request. One fault then produces one conversation instead of a pile.

Is it safe to send error details from the page?

It is, when the report is limited to the failure itself: address, status, message, page and browser. The risk comes from bundling everything the page holds, especially sign-in details or customer input. Ask whoever looks after your site to confirm exactly what a report collects before it goes live, and have that sentence shown to whoever presses the button.

From Tech Concierge

Want this handled for your business?

Your own engineer builds and runs the website, the app, the CRM and the AI assistant that answers your customers.

The first 7 days are free.

Read next

Case studyWebsites

Why a client cannot find the work you delivered

A client could not see or open work we had just delivered, because the report arrived as plain text with addresses that could not be tapped. What we changed on our own portal and app, and the checks that catch it before a client does.

4 min read

Case studyWebsites

A website for a second country is not just a translation

Our own platform was built for one country, so the country was written into hundreds of places rather than set anywhere. Here is how we traced it, what we changed, and what an owner can check before serving a second country.

5 min read

We count visits to this blog with our own analytics, never Google's, and share what you read with nobody. May we remember this browser with a cookie? See the privacy policy.