START HERE

Write the question first.

Before choosing a chart, state the decision the measurement will support. For a contact-page change, that might be whether qualified enquiries increased among visitors who reached that page. For an organic-search change, it might be whether relevant pages received more non-branded clicks. These are different questions and need different evidence.

Use our website change measurement planner to record one count, its source, and two date windows. It calculates a daily average and observed change, then exports the inputs and calculation as a worksheet. It does not estimate traffic, invent a baseline, or claim to isolate the effect of the redesign.

PRESERVE THE BASELINE

Save evidence before the launch.

Choose the metric and audience before looking at the after result. Keep the report or export that supports each total. Write down filters, reporting timezone, and the export date so another person can reconstruct the same comparison. If the tracking definition changes at launch, treat that as a measurement break and document it. A missing baseline stays missing.

  • Name one primary count: for example qualified enquiries, completed bookings, or organic clicks.
  • Define the audience or page group, including country, channel, and any exclusions.
  • Save the source report or export for each window with its filters and timezone.
  • Record the release date and the routes that changed.
  • Choose when to check back and what decision the result will inform.
  • List concurrent changes: campaigns, offers, staffing, seasonality, or tracking.

MAKE WINDOWS COMPARABLE

Compare the same population.

Use non-overlapping before and after windows. Match the audience, report filters, and metric definition. Allow time for the outcome you care about to occur: an enquiry and a qualified sales opportunity may be recorded on different days. If a launch was gradual, record that interval rather than assigning every visit to a single release date.

Our planner includes both boundary dates. For a window with total T over D days, the daily average is T divided by D. The observed percentage change is (after daily average minus before daily average) divided by before daily average, multiplied by 100. If the baseline is zero, that percentage is undefined; report the absolute difference instead.

Daily normalization handles different window lengths. It does not adjust for weekends, holidays, a different traffic mix, or missing events. A seven-day window and a thirty-day window can still describe very different business conditions. Prefer comparable business cycles and retain the limitations.

KEEP THE DENOMINATOR

A count is not a conversion rate.

More enquiries may arrive because more people visited, because a higher proportion enquired, or both. A per-day enquiry count is useful for operating volume. A visitor conversion rate needs a defined eligible audience as its denominator. Do not enter a percentage into a count worksheet and describe its output as a conversion-rate improvement.

When evaluating website conversion improvements, keep the count and the eligible visitor total together in the source report. Review lead quality as well as quantity. A form can attract more submissions while producing fewer useful conversations.

WRITE THE RECEIPT

Make the result reproducible.

A useful result statement names the metric, source, audience, dates, totals, and method. It also says what else changed. Keep the preview and landed change alongside that statement so the team can trace the implementation without confusing delivery with impact.

Use this structure: “For [audience], [source] recorded [before total] from [dates] and [after total] from [dates]. Using daily averages, the observed change was [result]. During this interval, [other changes or limitations].” Fill every field from evidence. If a field is unknown, say so.

Before-and-after observation alone does not establish causation. A controlled experiment may help answer a causal question when the design, sample, and instrumentation support it. For an observational redesign review, describe the association and avoid attributing the whole difference to your work.

DECIDE WHAT HAPPENS NEXT

Use the result to choose the next step.

Check the measurement for missing data and changed definitions before deciding whether to keep, revise, or investigate the change. A small sample or an unusual week may justify a later check rather than a confident verdict. Record that next date and the evidence you need.

If you are still scoping the work, our website redesign cost guide helps frame the investment. Pair that scope with a baseline and a measurement plan before launch. Then prepare your worksheet and keep it with the project record.

FAQ

Common questions.

What if I have no baseline?

Do not manufacture one. Record the first reliable window, document when tracking became usable, and plan a later comparison. Delivery and measurement can have different completion dates.

Can I use different window lengths?

The planner divides each total by its inclusive day count. This makes the arithmetic comparable per day, but you still need to assess differences in weekdays, seasonality, traffic, and tracking.

Does a positive result prove the redesign worked?

It establishes an observed increase for the specified metric and audience. It does not by itself isolate the redesign from other changes.

Where are my worksheet inputs stored?

The calculator works in your browser and does not submit its fields to our server. Download the worksheet if you want a copy; refreshing the page clears the form.