Quick answer
A user error is a mistake made by the person using a system, such as a wrong value or button. When many people make the same error at the same place, it signals a design problem, which you see in error rates and retries.
What user error means
A user error is a mistake made by the person using a system: a mistyped value, a wrong button, a misread screen. It is the common phrase for "the operator did it, not the machine". The phrase is also a verdict. It says whose fault the problem is before anyone has looked at the design.
Human-factors and medical-device standards avoid the verdict. They use use error, defined as an act or omission that produces a different result from the one intended or expected, and they note that many such errors come from poor interface design rather than carelessness (Wikipedia: Use error, citing IEC 62366). Tech-support slang such as "PEBKAC" (problem exists between keyboard and chair) is the verdict version.
The working position of this page: the error is real, but if it keeps happening at the same place, the cause is in the design, and the evidence is in your numbers.
Slips, mistakes and what they need
Nielsen Norman Group separates two kinds of error, and they need different fixes (NN/g):
| Kind | What happens | Example | Typical fix |
|---|---|---|---|
| Slip | The goal is right, the execution is wrong; the person is on autopilot | Typing "i" instead of "o"; picking the wrong date in a picker | Constraints, suggestions, good defaults, forgiving input formats |
| Mistake | The execution is right, the goal is wrong; the mental model is wrong | Treating a warning as a different warning and acting on it carefully | Clearer labels, better feedback, simpler concepts |
A third group is worth naming even though it is not an interface problem: violations, where a person knowingly breaks a rule. Counting them with slips and mistakes hides what you can fix.
How to tell user error from design error
Use a simple rule: one person making an error is an incident; many people making the same error at the same place is a design signal.
Check these three signs:
- Concentration. The errors cluster at one field, step or button rather than spread evenly.
- Repetition. The same person retries, often several times, before succeeding or leaving.
- Persistence. The rate does not fall with experience. Returning users make the same error as new ones.
If all three are present, treat it as design. If errors are rare, scattered and fall over time, it is closer to ordinary slips and you accept some rate.
A worked example: a checkout form
Sample data: 2,000 attempts to submit a five-field checkout form. A field error means the form rejected the field and showed a message.
| Field | Failed attempts | Error rate (failed ÷ 2,000) | Multiple of median rate (6.0%) |
|---|---|---|---|
| Phone number | 380 | 19.0% | 3.2 |
| Password | 240 | 12.0% | 2.0 |
| Postcode | 120 | 6.0% | 1.0 |
| 60 | 3.0% | 0.5 | |
| Name | 10 | 0.5% | 0.1 |
| Total | 810 |
Check: 380 + 240 + 120 + 60 + 10 = 810. The median of the five rates is 6.0%, and the phone field at 19.0% is 19.0 ÷ 6.0 = 3.2 times that.
The decision rule used here: flag any field whose error rate is at least three times the median. Phone qualifies; password at 2.0 times is worth a look but below the line. These are not the same cause: the phone field rejects numbers with spaces or a country code (a format problem), while password failures are rules the person cannot see until they fail.
Fix: accept spaces, dashes and a plus sign, and show password rules before typing. Suppose a re-run of 2,000 attempts shows phone failures at 70 (3.5%) and password failures at 110 (5.5%). Phone, at 3.5%, is now below the 6.0% median line and no longer flagged. The same people were using the form; the form changed.
Retries and how they differ from errors
An error counts a rejection. A retry counts what the person did next. Two forms can have the same error rate and very different outcomes:
- If most people fix the field and finish, the error costs seconds.
- If many leave after the second rejection, the same error costs conversions.
Track both: errors per attempt (above) and the share of sessions that end on an error step. A page with a modest error rate but a high exit rate right after the error message is the one to fix first.
A quick illustration with sample data: two forms both reject 380 attempts, a 19% error rate. On form one, 300 of the 380 people fix the field and submit (300 ÷ 380 = 79%). On form two only 140 do (140 ÷ 380 = 37%). Same error rate, but form two loses 240 people at the rejection, so its error message or field is the urgent fix.
Common mistakes in handling user errors
- Blaming the user in the copy. "Invalid input" says the person was wrong. "Enter the number without letters, for example 415 555 0100" says what to do.
- Showing the error late. Messages shown only after submit cost a retry; inline checks cost less.
- Counting only server rejections. Errors the page prevents by constraints never show up, which is the point; measure outcomes too.
- Lumping every error together. A phone-format rejection and a payment decline have different causes.
- Treating a rising error rate as a people problem. Check releases and device mix first. See error rate and frustration signals.
For the wider picture of what makes people give up, see the blog on user friction. NN/g's guidance on error messages is a good checklist for the copy.
Looking at user errors in TapCub
Use the checkout example and see what TapCub can and cannot show.
- Send the validation error as an event. TapCub does not read what people type into forms. Its form-submit tracking (switch on with
data-forms="true") records only the form id, name and action. To count field errors, calltapcub.track('form_field_error', { field: 'phone' })from your validation code, with a snake_case name and short property values. Do not send what the person typed. - Count it. In Web analytics › Events or All-platform analytics › Behavior › Event analysis, count the event and break it down by the
fieldproperty to get the per-field table above. - Tell it apart from script errors. Monitoring › Errors shows JavaScript exceptions (error count, rate against pageviews, affected sessions). A rejected phone number is not a script error, so it only appears if you send the event in step 1.
- See the behaviour. Open All-platform analytics › Behavior › Heatmaps for repeated clicks on a button and session replay (off by default) for a few failing attempts.
TapCub does not fix the form or write the error messages; it shows where the errors concentrate so you know which field to change.
Sources
- Nielsen Norman Group, Preventing User Errors: Avoiding Unconscious Slips.
- Wikipedia, Use error, covering the IEC 62366 definition and the reasoning for the term.
- Nielsen Norman Group, Error-Message Guidelines.
Was this explanation clear?
We use your answers to fix and expand this page.
Couldn’t send. Please try again in a minute.
Thanks for the feedback. We’ll use it to improve this page.You’ve already sent feedback on this page. Thanks.

Technical Engineer
Focuses on tracking code, SDKs, and API integrations, with practical guidance on implementation checks and troubleshooting.