Practical framework / 5 MIN READ
Compare the calendar before comparing the campaign
Align operating days, seasonal context and reporting maturity before interpreting a period change.
AI-assisted research and analysis for the SNS Creative Fieldbook
Sources checked 2026-11-06
A before-and-after chart combines marketing activity with everything else that changed over time. Better calendar comparisons make a trend more interpretable, although they do not create a causal experiment.
What you will leave with: Build a comparison note that explains calendar alignment, unusual events and remaining confounders.
Use this when
- Month-on-month results swing without a clear operational explanation.
- A campaign ran across a holiday or a change in opening hours.
This does not establish campaign incrementality or replace a randomized comparison.
Choose a comparison that matches the business rhythm
Identify the cadence that matters: weekdays for business enquiries, weekends for leisure bookings, pay cycles for subscriptions or opening days for stores. Two calendar months can contain different numbers of valuable trading days. A raw total can therefore move even when performance on comparable days does not.
Keep totals because they matter operationally, but supplement them with a suitable exposure basis such as eligible operating days. Do not divide by days automatically if capacity, hours or customer opportunity varies sharply. Explain why the selected comparison reflects the question the team is trying to answer.
Map trend and recurring patterns
NIST identifies trend, seasonal variation and dependence across observations as features of time-series data. In a marketing review, inspect more than one preceding period before declaring the latest movement exceptional. A useful chart shows enough history to make recurring rhythms visible.
Compare relevant prior-year dates where available, but align movable holidays and campaign timing rather than assuming matching calendar dates mean matching conditions. A prior year can also contain unusual disruptions. Note those explicitly instead of treating historical data as a perfectly stable control group.
Annotate operational and measurement changes
Create a short timeline of price changes, stockouts, store closures, sales capacity, website releases and tracking revisions. Add the campaign’s actual delivery dates, not just the launch announcement. These annotations provide candidate explanations rather than automatic causal attribution.
Check whether both periods have equally mature data. The latest week may still be receiving delayed orders, CRM outcomes or refunds. Use consistent definitions and extraction dates, and label provisional periods. A cleaner calendar comparison is still misleading if one side includes incomplete outcomes or a different metric version.
Write the conclusion at the supported strength
Separate description from explanation: orders rose, orders rose after a launch, and the launch caused orders to rise are different claims. A matched-day comparison can reduce obvious distortions while leaving competitor actions, weather or unmeasured demand changes unresolved.
Use the analysis to choose the next investigation. If the movement is concentrated on closed-store days, repair the reporting interpretation. If it persists across comparable periods, examine plausible business causes or design a stronger evaluation. The goal is to avoid unnecessary reactions while identifying a question that better evidence can answer.
Make the reasoning concrete.
A hypothetical B2B service records 420 enquiries in one month and 400 in the next. The first month had 21 staffed working days and the second had 20. Both yield 20 enquiries per staffed day in this simplified example.
Read raw totals only
Opening: Call the decline evidence that the new campaign weakened demand.
Hypothesis: Total workload did fall.
Risk: The number of staffed days also changed.
Show both views
Opening: Report total enquiries and enquiries per staffed day.
Hypothesis: The operational comparison becomes clearer.
Risk: Day-level customer mix and campaign causality are still unresolved.
State that the monthly total fell while the simplified daily rate was unchanged. Do not infer campaign effectiveness from those totals alone.
Take the questions into your next brief.
Use these prompts in your own notes. Nothing is submitted from this page.
Cadence
Which days or cycles represent comparable opportunity?
Alignment
Which holidays, hours and seasonal events differ?
Timeline
What pricing, capacity, inventory or tracking changes occurred?
Maturity
Are both periods equally complete?
Claim
Is the conclusion descriptive, diagnostic or genuinely causal?
Watch for the wrong conclusion.
- Comparing months with different operating-day counts without comment.
- Using last year as an unquestioned control.
- Ignoring movable holidays.
- Calling a corrected calendar comparison proof of incremental impact.
Your next action
Add a five-line calendar and operations note to the next performance report before explaining any change as a campaign effect.
Sources & scope
Frameworks, worksheets and hypothetical examples are original editorial synthesis. Linked primary sources support only the stated concepts. Source checking is not an independent audit of effectiveness. No named human author or reviewer is claimed.
- NIST/SEMATECH: Introduction to Time Series Analysis →Primary statistical handbook · Supports: Identifies autocorrelation, trend and seasonal variation as structures in time-series data. Does not establish that a particular marketing change caused an outcome.