Paid acquisition
Make better use of visitors already being acquired through Meta.
An ongoing, evidence-led CRO case study
CRO Case Study: How Do You Improve a Website Without an Analytics Baseline?
Written by Neil Webley, founder of FreeCROTool and a CRO and A/B testing specialist with more than ten years of experimentation experience · Published
This project starts before the first A/B test. An established touring entertainment business already attracts paid and organic visitors, but has no useful website analytics baseline from which to judge what happens after they arrive.
The task is to understand what already works, establish reliable measurement and investigate the visitor journey before claiming that any change has improved it.
The commercial challenge: make the website work harder
The anonymous client is an established UK touring live-entertainment business with three distinct productions. The same lead performer appears in each, but every show has its own identity, venues, towns, dates and schedule.
Meta advertising is already commercially important and actively measured. The measurement gap starts after a person clicks an advert and reaches the website. Organic search can also introduce visitors to a production.
Organic acquisition
Carefully expand relevant discovery through search.
Conversion
Help visitors identify a relevant performance and continue towards tickets.
Resilience
Reduce dependence on any single third-party acquisition platform.
Make the paid click count
When Meta is successfully generating visitors, the website should make good use of traffic the business has paid to acquire. That means helping people find the relevant production, performance, date, venue and ticket route without assuming the current journey is ineffective.
Reduce channel dependency
The owner has experienced occasional temporary Meta restrictions apparently associated with automated interpretation of a production's branding. This is not a judgement on the branding, Meta's decisions or intellectual-property rights.
It does show why stronger organic acquisition and a stronger owned website matter. The objective is diversification and resilience, not replacing a successful channel.
What happens after the paid click?
A useful visitor journey is likely to be: advert, search or other discovery → production → relevant performance → date, location and venue → ticket provider → ticket purchase.
At present, a visitor may need to work to answer basic questions: which show is this, when and where is it next performing, is it near me, and how do I buy tickets? Important performance information and ticket calls to action can sit below the fold, with much schedule information contained in a third-party iframe.
That is an observation, not evidence that the website loses ticket sales. We do not know that yet.
An imperfect website is not a website where nothing works
The website is built in Wix, is graphics-led, has relatively little text on some important pages, uses an iframe for event information and has a homepage video that appears poorly positioned on desktop. Strong imagery is not inherently a problem for an entertainment business. The question is whether important information and commercial actions are sufficiently easy to discover alongside it.
The early search evidence changes the diagnosis. This is not an SEO-disaster story and it is not a reason to rebuild the site blindly.
Do not mistake an imperfect website for a website where nothing is working.
Changing URLs, structure, redirects or production pages without understanding their current role could damage valuable signals.
Working principle
Preserve and expand.
First identify what Google already understands, which pages rank, which queries they answer, how domains resolve and which real-world links point into the site. Then investigate justified improvements.
What the first Search Console sample revealed
Google Search Console had been connected for only about five days. The following is an initial observation from roughly three meaningful days, not normal performance and not a long-term trend.
| Initial observation | Figure |
|---|---|
| Search impressions | 249 |
| Search clicks | 30 |
| CTR | Approximately 12% |
| UK performance | 208 impressions, 29 clicks, average position 4.37 |
| Mobile performance | 207 impressions, 27 clicks, average position 4.84 |
Approximately 83% of impressions in this small sample were mobile. That makes mobile ticket discovery, embeds, scrolling, touch interaction and CTA visibility important areas to inspect. It does not, on its own, prove that the mobile experience is poor.
Production pages
The strongest-performing production page recorded 177 impressions, 19 clicks and average position 4.38. Two other production pages recorded 48 and 23 impressions respectively, with average positions 6.33 and 5.87.
Other visible pages
The homepage recorded 20 impressions and four clicks at position 7.25. The tour dates page recorded eight impressions, one click and position 2.75.
Query pattern
A branded search for the strongest-performing production averaged position 3.07. Searches combining that production name with tour and tour dates averaged positions 2.65 and 2.0 respectively.
The initial pattern suggests that Google already understands the productions, branded/show searches and touring intent surprisingly well. It is too small a sample to treat as a settled conclusion. For the broader method, see how to use Google Search Console for CRO.
The opportunity may be to expand the search footprint
Several existing production-name rankings are already strong. The larger opportunity may be to make genuinely useful information available around performances, venues, towns, cities, regions and dates.
This does not mean generating hundreds of thin location pages or inventing keyword demand. It means investigating whether visitors would benefit from accurate, useful event information directly on the website, and whether that provides search engines with additional context.
Backlinks and domains: audit before changing
Ahrefs shows a mixed profile. An interesting side effect of operating a touring business is that it appears to have generated some useful link-building activity without evidence of a deliberate link-building campaign.
Venues, ticket and event platforms, local organisations, councils, entertainment websites and local media all have genuine contextual reasons to mention and link to individual productions. Those links may contribute to Google's existing understanding of the site and its authority, but correlation does not establish causation.
Low-quality or irrelevant links should be considered separately, not averaged together with relevant real-world links and labelled strong.
Relevant real-world links
Venue, event-platform, local-organisation and entertainment links are contextually plausible for a touring business.
Questionable links
Irrelevant directories, scraper-style sites and obvious spam are not proof of useful authority. There is no supplied evidence that a disavow is required.
Redirect setup
Production-specific domains should be audited for direct 301s, relevant destinations, chains and consistent canonicalisation before any change.
Why we are not A/B testing yet
At the start of the project, we know considerably more about Meta campaign activity, initial search visibility and third-party SEO data than we do about visitor behaviour after arrival.
There is no useful GA4 behavioural or conversion baseline. We cannot responsibly claim a new ticket CTA, next-show message, map or location-aware experience improved performance until we can measure the relevant outcome.
1
Measurement
Implement appropriate analytics, define conversions, validate events and document reliability.
2
Obvious fixes
Repair confirmed broken functionality, tracking or serious usability defects without pretending they need an A/B test.
3
SEO and content
Improve genuinely useful event information, context and internal linking carefully, then monitor.
4
CRO hypotheses
Test uncertain ideas such as CTA prominence, performance discovery or layout where appropriate.
Experimentation is for uncertainty. A broken tracking implementation should be fixed. A plausible alternative ticket-discovery journey may need an experiment. Read how to tell whether a website change actually worked before relying on a simple before-and-after comparison.
What should you do before your first CRO experiment?
For this project, the first work is not a design refresh. It is a short, approximately two-week benchmarking period in which we establish what success means and collect evidence.
- Understand the commercial objective and acquisition sources.
- Define primary and micro-conversions that represent meaningful progress.
- Validate analytics and document the measurement model.
- Inspect landing pages, visitor journeys, devices and ticket interactions.
- Review existing organic visibility, URLs, redirects and real-world links.
- Separate confirmed defects from uncertain ideas worth testing.
- Record the evidence, prioritise and write focused hypotheses.
Privacy, consent and trustworthy measurement
The site does not appear to have a dedicated cookie-consent banner, but Wix and other components may already set cookies or browser storage. We do not yet know what technologies are operating or which require consent.
Before adding analytics or marketing measurement, the implementation needs an audit: what sets storage, what it does, what categories apply, what additional technology is introduced and what privacy configuration is appropriate. This is not a legal conclusion about the current site. It is part of implementing measurement responsibly.
GEO baseline
The initial GEO/Generative AI sample recorded 26 impressions across the homepage and the three production pages.
It is useful to preserve as a baseline, but too small to demonstrate growth, decline or future potential.
What happens next
The project documentation now records the baseline, observations, research, recommendations, changes, hypotheses, experiments and results. CRO is a repeatable evidence and decision-making process, not a collection of random website tweaks.
Future updates may cover GA4 implementation, conversion definitions, baseline behaviour, appropriate cookie and consent work, SEO/content changes, performance-listing changes, organic and GEO visibility, ticket-CTA hypotheses and experiments. A neutral or negative result would still be useful evidence.
Start with evidence, not a redesign
SEO can help people discover a production, venue or performance. CRO then helps them answer: is this the show I want, is it near me, when is it, where is it and how do I buy a ticket?
The aim is to make the website a stronger owned acquisition and conversion asset while preserving the value it has already accumulated.
Related guides
Useful next reading for an evidence-led approach to website improvement.
Why is my website not converting?
Diagnose traffic, measurement, clarity, friction and technical problems before changing a page.
A/B testing without developers
Explore how a focused experiment can be implemented once there is a measurable question to answer.
How to interpret A/B test results
Learn how to read an eventual experiment with appropriate caution.