How to Measure Whether a Website Redesign Improved Conversions
Measure a redesign with consistent conversion tracking, comparable traffic and business-quality checks. Understand what before-and-after results can establish.
Measure whether a website redesign improved conversions by keeping the outcome definition consistent, validating tracking before and after launch and comparing equivalent audiences. Use a concurrent randomized comparison when practical. Without one, a before-and-after result shows a change in performance but cannot isolate the redesign from other changes.
Niwa supports the redesigned WordPress journey by answering customer questions, completing connected booking and sales workflows, and helping the owner with supported edits. Evaluate those functions against completed enquiries, purchases or appointments. A site built with AI uses it during creation. A site powered by AI uses it for customer and owner tasks after launch.
Save a baseline before launch
Choose the primary outcome and record how it is measured. For enquiries, include qualification criteria and the point where a request is considered successfully submitted. For a store, decide how cancellations and refunds will be handled. For booking, measure a saved reservation rather than a calendar click.
Record counts, rates and the eligible audience for the key journeys. Save breakdowns by device, session traffic source and relevant landing page. Include the operational information needed to judge quality, such as whether the team could serve the enquiries.
Choose a period that reflects normal business activity and note promotions, holidays or stock shortages. There is no universally fair comparison window for every business. An unusual baseline will remain unusual even if it is easy to export.
Keep instrumentation consistent through the redesign
Write down the events and success conditions before developers change the pages. Recheck them on staging and after launch. A new form, checkout or calendar can alter the point at which the same event fires.
Google's analytics metric schema distinguishes session key event rate from raw event counts. Keep the same unit on both sides of the comparison. Adding more events to the key-event list can change an aggregate rate even without improving the original customer task.
Complete a test journey and confirm its real business record. Verify event firing, duplicate handling and delivery to the team. Check the same route on a phone. Exclude test activity from the assessment.
If tracking is missing or substantially different in one period, say that the comparison is limited. A gap should not be filled with estimated conversions presented as observed data.
Prefer a fair comparison over a convenient one
If the old and new journeys can operate safely at the same time, a properly configured experiment can randomly assign eligible visitors and measure a shared outcome. Keep shared backend rules, offers and inventory consistent. Confirm that visitors do not encounter contradictory versions midway through a purchase or booking.
A large redesign may make concurrent testing impractical. In that case, record the launch date and compare equivalent audiences across suitable periods. Note other changes in ads, prices, promotions, service availability and response times.
Use cautious wording: "The measured enquiry rate increased after launch" is a descriptive observation. It becomes a claim about the redesign's effect only when the comparison supports that attribution.
Check whether the traffic mix explains the result
An overall rate can change because the audience changes. Compare like-for-like groups before treating the headline figure as a design result.
Consider an illustrative website with these numbers. Before launch, 1,000 mobile visits produce 20 enquiries and 1,000 desktop visits produce 60. The overall rate is 4%. After launch, 500 mobile visits produce 10 enquiries and 1,500 desktop visits produce 90. The overall rate becomes 5%.
Each device group's rate stayed the same. The overall increase came from the larger share of desktop visits. These fictional numbers explain the comparison problem; they are not a Niwa result.
Do not confuse the source that originally acquired a user with the source of the session you are evaluating. Google's acquisition-scope guidance explains that distinction. Choose the scope that matches your question and keep it consistent.
Evaluate useful outcomes after the initial action
Review qualified enquiries, booked appointments that were actually attended, or completed orders with their relevant costs. A redesign that attracts more unsuitable requests can raise submissions while consuming more team time.
Use a fair follow-up window. Enquiries received yesterday may not have had time to become sales, while older enquiries have. Compare similarly mature groups when judging customer quality.
Record journey errors and operational changes as guardrails. If the redesigned site increased requests but the team took longer to respond, the result concerns both the website and the business handling that followed.
Measure Niwa's role as a separate change where possible
If you add Niwa at the same time as a new design, new prices and a different offer, the total result describes that combined change. It does not establish the assistant's individual contribution.
Configure Niwa with accurate business information and connect the booking or sales workflow you want customers to complete. On WhatsApp, Niwa uses that website context to answer service questions, check availability and complete a booking directly in the conversation. Count the saved appointment. For a website experiment, use the eligible website visitors as your measurement population and record WhatsApp completions separately unless your measurement connects the two.
Use authorized WordPress operations for supported owner edits and keep a change log. Once the redesigned journey is stable, you can plan a more focused evaluation of an individual assistance or content change.
Niwa Business is USD 150 for a custom WordPress website or migration, hosting, maintenance and AI adaptation. Contact Niwa about your redesign with the original journey, the proposed customer task and the records you can use to assess it.
