Why do reports and the Backoffice show different results?
Two numbers only disagree when they measure the same thing. Check the population, filters, cycle, your visibility, and when each report was generated before treating a difference as a data error.
Start here
Write down the exact report or view on each side, the cycle or programme, and the filters applied.
Check which users each source includes. This is the most common cause of a difference (see below).
Apply the same filters on both sides: groups, teams, status, dates.
Check your own visibility. As an administrator you only see the users inside your organisational scope, so another administrator can get a different total from the same report.
Check when each file was generated, then request a new export if a change happened after that.
If the totals still differ, pick one user who appears in one source and not in the other, and compare that user's record. That user usually explains the rule.
Which users does each evaluation source include?
Evaluation dashboard export in the platform: lists only the users who already have evaluation results. People who have not been evaluated yet do not appear, so a percentage calculated from this file can look higher or lower than in the Backoffice.
Export final report in the Backoffice: if this action is available in your Backoffice, use it when you need the whole population of the cycle. It is not available for every organisation.
Backoffice completion percentages: are calculated over the users of the cycle that you can see.
When is report data generated?
Backoffice reports and exports: each request is queued and generated with current data, usually within a few minutes. If a change is missing, request a new export after the change instead of reopening an old file.
Some Learn analytics reports (play time and quiz performance, for example): read a copy of the data that is refreshed every 30 to 45 minutes. A quiz answered a few minutes ago may not appear yet.
People Intelligence dashboards: are refreshed once a day, early in the morning (Lisbon time). Some behavioural indicators only change once a week. A change made today normally appears the next day.
Different words, different numbers
Submitted, finished, final and shared are different stages of an evaluation. A report that counts submitted evaluations and a view that counts shared results will not match until every evaluation is shared. Compare the same stage on both sides.
What not to do
Do not change users, evaluations, or eligibility just to make two reports match. The right number depends on the question you are answering.
When should I contact Support?
Contact Support when the population, filters, cycle, stage, and your visibility are aligned, a new export was generated after the last change, and the totals still differ. Include the names of both reports, the cycle, the filters, the totals on each side, when each file was generated, and the identifier of one user who appears differently.
Related articles
Cycle and Period Dashboards, Reports, and Export: where to find the evaluation dashboards and exports.
Which report and percentage should an administrator use?: choose the right report for your question.
How do I interpret evaluation statuses and actions in the Backoffice?: what each evaluation stage means.