Back to all articlesReports and dashboards

Choose a Weekly Reporting Window That Readers Can Compare Reliably

Alactic Editorial9 min read

Prepared with AI assistance by Alactic Editorial. Examples identified as hypothetical are illustrative, not customer results.

A dashboard says enquiries are down, but its current figure covers Monday through Wednesday while its comparison covers seven days. Another chart says activity is up because its “weekly” number includes a different set of dates. Neither result tells readers whether performance actually changed.

Choose a completed calendar week for settled weekly comparisons, a rolling seven-day window for recent operating conditions, and week-to-date for progress within the current week. Then display the exact dates, time zone, completeness status, and comparison period beside the number. Do not let the word “weekly” stand in for those decisions.

The practical test is simple: can a reader identify what each number includes and explain why the two periods are comparable without opening another report?

1. Start with the decision the reader needs to make

The right window depends on the question, not on which date filter is easiest to select.

  • “How did the last finished week compare with the one before?” Use completed calendar weeks with a fixed start day.
  • “What has been happening recently?” Use a rolling window with a declared cutoff and refresh rule.
  • “How are we doing so far this week?” Use week-to-date and compare it with the same elapsed portion of the previous week.

A Monday-to-Sunday week is a reasonable reporting convention, but it is not the only one. If the business uses Sunday-to-Saturday, document that choice and keep it consistent across reports intended to be compared. A calendar week is a defined recurring block, not necessarily the default week in a particular tool.

You can offer all three views without mixing them. Give each a distinct name and purpose. For example, use “Last completed week” in a performance summary, “Last seven completed days” in an operational view, and “This week through yesterday” in a progress card. These labels reveal more than three cards all called “Weekly enquiries.”

2. Use this filled hypothetical decision table

Fictional worked example: A business counts new enquiries by the time each enquiry was created. The report is viewed on Thursday, 11 June 2026 at 09:00 UTC. Its declared data cutoff is 11 June at 00:00 UTC, so the latest included date is 10 June. Its reporting week runs Monday through Sunday.

The decision table below is presented as three filled rows. Every date, count, and result in this example is hypothetical.

Row A — Completed calendar week

  • Purpose: Compare the most recent finished reporting week with the preceding finished week.
  • Displayed label: Last completed week · 1–7 Jun 2026 · UTC.
  • Start: 1 June at 00:00 UTC.
  • End: 8 June at 00:00 UTC, exclusive; the final included date is 7 June.
  • Completeness: All seven reporting days have ended. Data is available through the period end, subject to later corrections.
  • Comparison: 25–31 May 2026, using the same metric definition and filters.
  • Hypothetical result: 140 enquiries versus 125, an increase of 12%.
  • Suitable use: A weekly performance summary that should not change simply because another day has passed.
  • Misleading comparison: Comparing these seven days with the current Monday-to-Wednesday total.

Row B — Rolling seven completed days

  • Purpose: Show recent activity without waiting for the current calendar week to end.
  • Displayed label: Last 7 completed days · 4–10 Jun 2026 · UTC.
  • Start: 4 June at 00:00 UTC.
  • End: 11 June at 00:00 UTC, exclusive.
  • Completeness: Seven ended reporting dates; the window advances daily after the declared data cutoff is available.
  • Comparison: The preceding non-overlapping seven days, 28 May–3 June 2026.
  • Hypothetical result: 154 enquiries versus 140, an increase of 10%.
  • Suitable use: Monitoring recent workload or activity.
  • Misleading comparison: Treating the change from yesterday’s rolling total as an independent weekly performance change.

Row C — Week-to-date through yesterday

  • Purpose: Check progress during the current reporting week.
  • Displayed label: Week to date · 8–10 Jun 2026 · 3 of 7 days · UTC.
  • Start: 8 June at 00:00 UTC.
  • End: 11 June at 00:00 UTC, exclusive.
  • Completeness: Incomplete week; three complete reporting days included.
  • Comparison: Monday–Wednesday of the previous week, 1–3 June 2026.
  • Hypothetical result: 66 enquiries versus 52, an increase of approximately 26.9%.
  • Suitable use: Assessing whether current progress differs from the same stage last week.
  • Misleading comparison: Comparing 66 with the previous full week’s 140 and describing the difference as a performance decline.

These results can all be valid at once. They answer different questions. Their apparent disagreement becomes a problem only when the labels hide those differences.

3. Separate an ended period from complete data

A period can have finished while its records are still arriving. Conversely, a current week can contain fully loaded days while remaining an incomplete week. A single “complete” badge cannot communicate both conditions reliably.

Use two separate statements where this distinction matters:

  • Period coverage: Seven of seven reporting days, or three of seven reporting days.
  • Data availability: Available through a stated cutoff, delayed, or provisional because a known source is still updating.

Do not infer data availability solely from the dashboard refresh time. “Refreshed at 09:00” says when the report was rebuilt; it does not necessarily say that all events through 09:00 are included.

For the hypothetical example, a useful note is: “Period ended 7 Jun; data available through period end; refreshed 11 Jun at 09:00 UTC.” If one input is delayed, replace the reassuring note with the known limitation. Suppress the percentage change if the delay makes the comparison materially unequal.

Also decide whether past values can be revised. A live report may incorporate late records into an earlier week. A saved report may preserve the value published at the time. Either design can be useful, but readers should know which they are seeing. If published totals change, mark the revision rather than silently leaving conflicting versions in circulation.

4. Define boundaries once, including the time zone

Use a start-inclusive, end-exclusive rule in the reporting specification: include events at or after the start and before the end. This avoids having to express an artificial final instant such as 23:59:59.

The reader-facing label can still show inclusive dates. “1–7 June” is easier to scan than a technical timestamp interval. Keep the precise boundaries in the report definition or supporting note.

Record these choices together:

  1. Reporting time zone: The zone used to assign events to dates.
  2. Week start: The weekday and local time at which the reporting week begins.
  3. Event timestamp: For example, enquiry creation time rather than its latest update time.
  4. Cutoff convention: Through yesterday, through a specified hour, or another explicit boundary.
  5. Refresh rule: When the displayed interval advances and what happens if data is late.

A rolling window especially needs precision. “Seven days” might mean seven completed local dates or the most recent 168 hours. Those are different designs, particularly around daylight-saving changes. For a day-based business report, completed reporting dates are often easier to explain. For an elapsed-time operating measure, a trailing-hour window may fit better. Name the actual choice rather than treating the two as interchangeable.

5. Match the comparison before calculating the change

Equal duration is necessary for many volume comparisons, but it is not sufficient. Check that the periods also use the same metric definition, filters, event timestamp, and data-availability standard.

For completed calendar weeks, compare with the immediately preceding completed week unless the report explicitly names another baseline. For rolling seven-day totals, a preceding non-overlapping seven-day window provides a straightforward comparison. For week-to-date, match both weekday coverage and cutoff.

If week-to-date includes part of Thursday, its comparison should include the corresponding part of the previous Thursday. Comparing a live Thursday afternoon total with a previous Monday-to-Wednesday total gives the current period extra time. A simpler alternative is to exclude the current day from both periods and say “through yesterday.”

Percentages and averages also need care. A three-day conversion rate can be calculated, but that does not automatically make it a reliable substitute for a full-week rate. Show the underlying counts or denominator so readers can see how much activity supports the result. Do not average daily percentages when the intended measure is a rate across all included records; calculate that rate from the relevant period totals.

If the previous value is zero, do not display a normal percentage increase. State the two counts and mark the percentage comparison as not applicable.

6. Explain what rolling windows hide

Rolling windows are useful for recent conditions, but adjacent daily snapshots overlap heavily. A seven-day window ending today shares six dates with the seven-day window ending yesterday. Their difference comes from the date entering the window and the date leaving it.

Hypothetical illustration: Yesterday’s rolling total was 150. The date leaving the window contributed 10 enquiries; the date entering contributed 24. Today’s rolling total becomes 164. The rise of 14 describes that replacement, not seven new days of improvement.

This matters when readers interpret a rising rolling chart as a succession of independent weekly gains. Keep the rolling view for the recent-conditions question, and use non-overlapping periods when the reader needs distinct blocks of performance.

Calendar weeks have a different weakness: they can feel stale late in the current week. Do not solve that by quietly extending the last completed week. Add a separately labelled week-to-date view instead. You preserve a stable comparison while making current activity visible.

7. Repair ambiguous reports without erasing their history

If existing reports mix windows, begin with the most-used headline metrics rather than rebuilding every chart at once.

  1. Inspect the actual query or export. Establish the dates, zone, timestamp field, filters, and cutoff behind each “weekly” figure. Do not rely on its title.
  2. Assign a purpose. Decide whether each figure is a completed-period comparison, a recent-conditions measure, or a progress check.
  3. Correct the label and baseline together. Renaming a card does not repair an unequal comparison.
  4. Recalculate affected changes. If the historical inputs cannot be reconstructed, mark the old comparison unavailable rather than estimating it without disclosure.
  5. Explain the break. Record the effective date and old versus new convention so readers do not mistake a reporting change for a business change.

Where exports or screenshots circulate, place the period information beside the metric, not only in a dashboard-wide filter. A cropped card should still reveal whether it represents a finished week or progress so far.

If two teams need different weeks, retain both only when their purposes justify it. Name each convention explicitly and avoid combining their totals as though the periods align.

8. Run a small comparison audit before the next publication

Choose three prominent weekly metrics and apply the filled decision-table structure to each. For every one, write down the purpose, exact start and end, time zone, period coverage, data cutoff, and comparison interval. Then inspect the published label against that specification.

Use a measurable acceptance check: all three metrics must show identifiable dates and completeness, and every displayed change must use the declared comparison period. Count missing labels, mismatched intervals, and unsupported completeness claims separately. Correct those defects before expanding the exercise to less-used charts.

Finally, ask a reader to explain one result using only the visible report. If they cannot tell whether “up this week” means a finished week, seven recent days, or progress through yesterday, the reporting window is not yet clear enough.

Reporting windowsDashboardsData qualityWeekly reporting

Put the next conversation to work.

Explore Alactic Communications for WhatsApp conversations, campaigns and AI-assisted workflows. Choose the product and plan that fit your team.

Explore Communications