A website redesign needs a measurement plan before the new layouts are approved. Otherwise, a change in reported conversions may reflect a new tracking rule, a different traffic mix or a broken form rather than a better customer experience. The useful question is specific: which journeys should improve, what evidence would demonstrate that improvement, and can the business compare the old and new versions fairly?
Choose outcomes that survive a visual redesign
Write down the commercial outcome for each main journey. A service business might care about enquiries accepted by sales, while an online shop may prioritise completed purchases and subsequent cancellations. Button clicks, video plays and form starts can help explain behaviour, but they should not quietly replace the outcome the business actually needs.
Define terms jointly with the people handling the result. An enquiry might be technically valid yet outside the service area or unrelated to the offer. Agree what counts as accepted, rejected, duplicate and awaiting review. Keep these definitions stable across launch so a change in qualification policy does not masquerade as a redesign improvement.
Write a measurement contract for each event
For every event, record its name, trigger, permitted attributes and expected destination. Include the exact success condition. For example, a successful enquiry should correspond to a confirmed submission accepted by the receiving system, rather than a press of the submit button. A visitor can press that button while required information is missing or the server is unavailable.
Describe duplicate handling explicitly. Refreshing a confirmation page, navigating back or retrying a slow request should not manufacture additional successful enquiries. Give the implementation team test cases for those behaviours. Keep event attributes limited to the information required for analysis; free-text messages and contact details do not belong in general website analytics.
| Journey stage | Evidence to capture | Interpretation limit |
|---|---|---|
| Service evaluation | Relevant landing page and journey type | A page view does not establish purchase intent. |
| Form interaction | Start and validation error categories | A start is not a submitted enquiry. |
| Submission | Confirmed successful receipt | Receipt does not establish commercial fit. |
| Sales review | Accepted or rejected outcome | Unreviewed leads must remain identifiable. |
Capture a baseline with enough context
Save the existing event definitions, form fields, campaign conventions and reporting filters. Export a baseline covering a period that represents the business's trading pattern. Record outages, offers, unusual campaigns and seasonal demand. There is no universally correct comparison window: a business receiving occasional complex enquiries has different evidence needs from a busy online shop.
Break the baseline down by device, acquisition channel, landing-page group and journey. Keep the breakdown manageable enough to investigate. A higher overall conversion rate can result from attracting more returning customers while the mobile experience deteriorates for new visitors. Looking at comparable groups makes that problem easier to see.
Separate experience changes from visibility changes
A redesign may alter both how people arrive and what happens afterwards. Monitor search landing pages alongside on-site completion. If a high-intent page disappears, the remaining traffic may convert differently even when the form is unchanged. Preserve a mapping between old and new page groups so reports remain interpretable after navigation and URL changes.
When URLs change, follow Google's official site-move guidance for mapping, redirects and monitoring. Treat migration stability as a separate workstream within the redesign. A successful event implementation does not establish that search engines can still reach the content that previously introduced customers to the business.
Test measurement as part of acceptance
Run complete journeys on the devices and browsers your audience uses. Check valid submissions, invalid entries, server errors and interrupted connections. Confirm that the business receives each successful test and that reporting records the intended event once. Label test records consistently so they can be excluded from commercial evaluation without hiding real failures.
Test the site's actual consent states as well. Record what measurement is expected when analytics permission is unavailable, and verify that implementation follows the configured choices. Changes in consent behaviour can change observed totals. Explain those limitations in the reporting notes instead of treating every unobserved interaction as a lost customer.
Give one person responsibility for reconciling the browser event, receiving system and reporting destination. Otherwise, the designer may see a confirmation message, the developer may see a successful request, and the sales team may still receive nothing. Each layer needs its own evidence before the journey is accepted.
Set decision rules before seeing results
Agree which failures require immediate repair and which questions need more observation. A broken submission path is an operational defect. A small shift in conversion rate may be noise, especially with few enquiries. Record sample sizes and commercial outcomes rather than celebrating percentage changes without their underlying counts.
Where practical, a properly designed controlled experiment offers stronger evidence about a particular change than a simple before-and-after comparison. If the entire site must switch at once, describe the result as observational. Account for campaign spend, pricing, availability and sales follow-up changes. A launch annotation alone cannot remove those competing explanations.
Turn reporting into a review routine
Arrange an early technical review and a later commercial review matched to the sales cycle. The first establishes whether journeys and measurement work. The second considers whether accepted enquiries or completed orders have changed. Keep unresolved leads visible so a slow review process is not mistaken for poor website performance.
Visuosofts' conversion rate optimisation service is relevant when the redesign needs a clear evaluation framework. For service websites, pair this plan with B2B website lead qualification forms. To discuss your existing measurement and intended outcomes, contact Visuosofts with the journeys that matter most.