The Brief
The client is an independent children’s dance school in London, with over two decades in business and more than 1,000 students through its programmes each week. Revenue has grown by more than a third year on year, confirmed by the company’s statutory accounts.
The school had done what founder instinct usually dictates: built a website with an agency, run paid social advertising, relied on word of mouth for the long tail. It worked, broadly. But the model had a structural problem. Every enquiry was effectively bought. Nothing came in unless someone was spending.
The engagement: rebuild the school’s digital presence to stop renting acquisition and start owning it.
The Challenge
The existing website was slow, template-built, and had accumulated five years of patches without any underlying architectural logic. On mobile, it was measurably poor: pages loaded slowly, the layout required effort from visitors, and there were no pages built around how parents actually search for children’s activities.
The school served multiple communities across London, but its search presence was limited to the immediate neighbourhood. It ranked meaningfully for its own name and very little else. Seven trackable keyword rankings in total. For a school at this scale, that number should have been in the dozens.
Paid advertising was delivering leads at a cost-per-lead below the industry average. But those leads varied in quality, and the website was doing little to qualify or inform visitors before they enquired. The conversion from enquiry to booked trial was leaving significant headroom.
There was no organic search presence worth measuring, and no presence in the AI answer engines that now shape how a growing share of parents research and shortlist local children’s activities.
The founder described the history plainly: “I spent thousands, I mean thousands, with three different companies, and the most recent one… I paid a fortune and it was just not a great website.”
Our Findings
A structured audit across technical performance, search architecture, content depth, and digital authority produced a consistent picture.
Technical debt was suppressing everything. Slow load times, particularly on mobile, were penalising the site in search rankings and inflating the cost of paid advertising. Landing page quality affects how much advertisers pay per click: a poor landing page experience increases that cost directly. This site was doing the opposite of what it should.
Search architecture was absent. The school served six or more distinct communities, but there were no location-specific pages. There was no content addressing the questions parents type when they are researching and comparing options. The site had seven trackable keyword rankings. Nothing outside the immediate area.
AI engine presence was zero. The major conversational AI tools, which parents increasingly use to find and compare local children’s activities, had no clear reason to cite the school. The site’s structure did not support it.
The reputation was genuinely exceptional. Well over 100 five-star Google reviews. A perfect rating. That foundation was an asset being wasted by weak discovery.
The Strategy
The central question was not what the website should look like. It was what the website should do, and how we would know when it was doing it.
Three decisions shaped the project.
Rebuild, not patch. The technical debt on the existing site was not fixable incrementally. A patched version of a structurally weak platform would have produced marginal improvements at best and left the underlying limitations in place. A rebuild gave the opportunity to resolve the technical foundation properly and build the content and structure correctly from the start.
Design for breadth of discovery, not depth of brand. Most small-business websites are built around one question: “When someone searches for us by name, do they find us?” That is a question with a small addressable audience. The more commercially valuable question is: “When a parent searches for children’s dance classes anywhere in London, do we appear?” Answering that required location-specific pages for each community the school actually served, content addressing the questions parents ask when choosing an activity for their child, and technical signals that communicated relevance to both search engines and AI systems.
Measure for self-sufficiency, not just performance. Most website projects are measured against traffic and rankings. The ambition here was different: to identify the point at which the site became capable of generating enquiries without paid advertising running. That required tracking at the lead level, not just the session level.
The bet, made explicitly at the outset, was that a site built to be genuinely useful, fast, locally specific, and clearly structured, would be cited by search engines and AI systems without requiring ongoing payment to appear. The results would confirm or contradict that within weeks.
Decision Log
The judgement calls behind the build, and the reasoning for each.
| Decision | Why |
|---|---|
| Rebuild the site rather than patch the existing one | Five years of patches had left no architectural logic; incremental fixes would have produced marginal gains and preserved the underlying limitations. |
| Design for breadth of discovery, not depth of brand | Ranking for the school’s own name reaches a small audience; ranking when any parent searches for classes across London is the commercially valuable position. |
| Build a separate page for each community served | The school served six or more communities but ranked in one; location-specific pages claimed reach that already existed but had never been built. |
| Add content for parents in the research phase | Capture parents comparing options before they enquire, and give AI systems substantive material to cite. |
| Measure at the lead level, not the session level | The goal was self-sufficiency: knowing when the site generated enquiries without paid advertising, which traffic and ranking figures cannot show. |
| Build a machine-readable structure for search and AI engines | Both needed a clear understanding of what the school offers, where it operates, who it serves, and its reputation, to cite it as a primary local source. |
The Website
The rebuild produced a materially faster site. Mobile performance scores improved by nearly 20 points on the key pages measured. Session duration increased by around 40%; pages viewed per session rose by a similar margin; desktop bounce rate fell markedly.
Several location-specific pages were built, each targeting one of the communities the school actually served. The majority were ranking in search within a week of going live.
A set of targeted pages was added for parents in the research phase: content addressing the questions parents type when deciding between activity options for their children, including the kind of material that AI systems draw on when answering parent queries. A dedicated page built around high-demand local searches with genuine informational depth was included from day one, rather than added as an afterthought.
The site was built to give both search engines and AI systems a clear, machine-readable understanding of what the school offers, where it operates, who it serves, and what its reputation looks like.
No templates. No off-the-shelf theme. The design, structure, and content were built to fit this school specifically.
AI Visibility
As of a measurement snapshot taken in May 2026, the school had strong presence across the major AI answer engines.
On the leading conversational AI platform, the school appeared in the large majority of tested local queries, ranking first in the majority of those. For several specific local queries covering a range of the school’s services, the school ranked first simultaneously across all four major AI platforms tested.
The school was also cited within Google’s AI-generated answers for local dance and fitness searches, appearing alongside substantially larger national sources, including encyclopaedic reference sites. Its blog posts and programme-specific pages were being cited as sources, not only its homepage.
A separate live check in June 2026 confirmed the school is the named top recommendation on the leading conversational AI platform when parents search for children’s dance classes in the local area.
These are dated findings, not a continuous measurement. AI citation patterns shift with content updates and platform changes. What the data shows is that the authority signals built into the site were sufficient to be recognised as a credible, primary local source within months of launch. That positioning takes time to build and, unlike a paid placement, it does not stop the moment spend stops.
Search Visibility
Measurement note: figures compare the period following the rebuild against the pre-rebuild baseline, drawn from search-console data. They reflect the early months after relaunch and are not presented as steady-state. No claim is made beyond what the data shows.
Within weeks of the site going live, keyword rankings had doubled from their initial count. Organic daily search impressions grew substantially, from a low initial baseline to a peak of around 700 per day.
The site reached the first page of Google for “dance classes near me”, a query generating thousands of searches per month, within weeks of launch. It appeared in the top five positions for multiple location-specific queries across the London area it serves.
The composition of that search traffic changed substantially in the months that followed. By June 2026, the significant majority of search impressions were coming from parents who had not searched for the school by name: parents who did not know it existed, searching for children’s dance options in their area. That is the measure of reach. It is different from brand recognition, and it is commercially far more valuable.
Outcome
In March 2026, the school’s paid advertising was paused.
The expectation, based on prior experience, was that enquiries would stop. That had always been the relationship: pay for ads, receive leads. Stop paying, receive nothing. The school had never received meaningful organic enquiries before the rebuild.
Instead, in April 2026, over a period of less than two weeks, with no paid advertising running at all, the website generated around ten enquiries through organic search. Every one came through the rebuilt site. None came through paid channels.
One figure makes the channel quality point clearly. In the months before the advertising pause, website-sourced enquiries converted to enrolled students at approximately six times the rate of paid social leads. Volume was smaller; quality was substantially higher.
Revenue grew by more than a third year on year, confirmed by the company’s statutory accounts. The website is one contributing factor among several, but it is now a functioning acquisition channel rather than a brochure kept visible by advertising spend.
What We Learned
The most transferable insight from this engagement is also the most honest test of any website: what happens when you turn the ads off?
Most small-business websites fail that test immediately. They have no organic search presence, no content that AI systems recognise as authoritative, and their location reach rarely extends beyond the immediate area. They function as brochures that are visible only while advertising is running. The moment the budget stops, so does the business flow.
Building a site that passes the ads-off test requires different decisions from the start. It requires thinking about discovery before design, architecture before aesthetics, and measurable outcomes before launch. It requires patience, because search engines and AI systems take weeks to months to index and assess new content. And it requires a willingness to measure honestly, tracking leads rather than traffic, to know when the asset has become self-sustaining.
This school now has a website that generates enquiries when no one at the school is doing anything to generate them. That is a materially different kind of asset from what most businesses invest in when they commission a website.
Leadership Takeaways
Most small-business websites are expenses, not assets. A site that only generates enquiries while advertising runs is a conduit, not an asset. The distinction matters: one has residual value; the other ceases to produce the moment you stop paying.
Website quality directly affects paid advertising costs. Slow pages and weak content structures increase cost-per-click, independent of bidding strategy. Better technical performance lowers acquisition costs across every paid channel simultaneously.
AI engine visibility is already happening, not coming. Parents are asking conversational AI tools to recommend local children’s activities right now. Businesses are being cited, or not, in those answers. The window for early local positioning is open but will not stay open indefinitely.
Channel quality outweighs channel volume. Website-sourced enquiries converted to enrolled students at six times the rate of paid social leads. A smaller volume of higher-quality enquiries is a better commercial outcome than raw lead volume, and it costs less to manage.
Local search reach is not automatic for local businesses. A school serving several communities had meaningful search presence in one. Location-specific pages changed that within a week of going live. The reach was there to be claimed; it had not been built.
This is the work behind our SEO services: local visibility a business owns rather than rents. If that is the position you want, tell us about your business.