A scheduled report should explain which period it covers, when its source data was current and whether the refresh completed fully. Report generation time is not the same as data freshness. Keep the last valid result distinguishable from a newly generated report built on delayed or incomplete inputs.
Let’s plan your report freshness rules.
Bring your workflow, constraints and open questions. We’ll help you define the requirements and work through a practical next step.
Define the reporting contract
Record the population, cutoff, time zone, metric definitions and expected availability. Identify each upstream dependency and its accountable owner. A daily report may rely on a source that arrives less frequently.
For a hypothetical dispatch report, records updated after the cutoff may belong to the next release. Explain that boundary rather than letting users interpret every missing row as a failure.
Use a freshness and completeness register
| Field | Meaning | Misinterpretation to prevent |
|---|---|---|
| Reporting period | Business interval summarised | File creation date mistaken for coverage |
| Source watermark | Latest eligible source change included | Page refresh mistaken for data refresh |
| Extraction time | When data was read | Extraction treated as source completeness |
| Generation result | Whether processing finished | Success hides skipped records |
| Missing inputs | Dependencies not available | Partial report appears final |
| Prior valid version | Last accepted result | Old values shown without age |
Choose terminology users understand. A technical watermark can be translated into a clear statement about included records without exposing internal processing details.
Decide what happens after a failed refresh
Options include withholding the new report, publishing a clearly labelled partial result or retaining the prior valid version. Select the rule from the business use and explain it consistently across email, dashboard and download.
Do not send a normal-looking success email when generation failed. Route the failure to an owner who can investigate and tell consumers what changed.
Test an interrupted source cycle
Delay one input in a controlled environment and verify the report's labels, alerts and distribution behaviour. Restore the source and rerun without creating conflicting copies presented as the same final version.
Validate counts and selected records against the stated cutoff. A spreadsheet opening successfully is not evidence that the report is complete or current.
Scheduled-report questions
Should every report display a last updated time?
Display the timing that answers the user's question, with a clear definition. One ambiguous timestamp can create false confidence.
Can a partial report still be useful?
Sometimes, if its limits are explicit and the business approves its use. Do not quietly treat it as the final population.
Use metric definition guidance, and discuss refresh dependencies with Drema's data engineering team.