Do we pause website improvement until analytics is perfect?

No. Pause *unguided redesigns* and *statistical A/B tests* that depend on broken conversion events, while moving straight ahead with fixing reproducible mobile bugs, confusing error messages, hidden delivery rules and missing product specifications that can be verified directly on the screen and against your backend order records.

How can we identify checkout or form friction without relying on GA4?

Combine three sources that don't depend on GA4 tags: a step-by-step walk-through on real mobile and desktop browsers (intentionally triggering form validation errors and checking payment redirects), your customer support and sales desk call logs (which tell you what buyers are having to phone up and ask), and your backend order and CRM delivery logs.

What is the difference between a functional fix and an A/B test?

A functional fix repairs something that is broken, hidden or actively obstructing a buyer (such as an obscured mobile button, a broken address lookup, or missing VAT/delivery clarity). You ship functional fixes directly and verify them. An A/B test compares two viable commercial approaches when the better option isn't obvious and you have both the traffic volume and the clean tracking required to measure the difference.

What should a CRO review do when the conversion data is wrong?

Asked to improve website conversion when GA4 is broken or untrusted? How to separate what you can fix immediately through observation from what needs clean data first.

What should a CRO review do when the conversion data is wrong?

It is one of the most common situations we walk into: a business knows its website should be converting better, wants to start conversion rate optimisation (CRO) work, and then discovers in week one that GA4 is double-counting purchases, missing basket steps, or counting failed form clicks as leads. That leaves the team with an awkward question: **do we freeze all website improvements until analytics is fixed and has collected three months of clean data, or can we start improving the journey now?** The short answer is that you don't have to sit on your hands, provided you know which decisions require statistical funnel data and which problems can be proven directly on the screen in front of you. ## What needs clean analytics vs what you can prove by watching the journey Not every conversion obstacle requires a 50,000-visitor A/B test or a pristine GA4 funnel chart to prove it is hurting your business. If a mobile visitor cannot tap the "Continue to Payment" button because a chat widget sits directly on top of it on an iPhone screen, you don't need three months of clean GA4 data to know that is a bug. Conversely, if someone wants to redesign your entire category page layout bas