The Lottie story - what happens when you scale a directory wrong

Lottie was the UK's biggest care home directory. Series A funded, name-checked by The Telegraph, Tatler and the FT, valued at around £40 million at peak. Their model was clean. Aggregate every care home in the country, build a city page for each location, monetise via lead generation to the homes.

In the 18 months between January 2024 and June 2025, Lottie lost 87 percent of its organic Google traffic. Their February 2024 monthly visit number was around 1.1 million. Their June 2025 number was around 145,000. Recoveries since have been partial. The business is not dead, but it is a different business from the one Series A investors funded.

I find this case interesting because the same dynamics apply to every UK service directory, including this one. We have built GoodHound deliberately not in the Lottie pattern. Below is what they did, what Google did to them, and what we are doing differently.

What Lottie built

The Lottie model in 2022-2023 was templated city and region pages at scale. Care home in Manchester. Care home in Leeds. Care home in Stockport. Care home in Stockport BR1. Care home in Stockport BR2. Hundreds of thousands of pages, each one a thin variation on the previous, with a list of homes pulled from a single underlying database.

The pages worked initially. Google indexed them all. They ranked for long-tail searches with low competition. Traffic compounded. By 2023, Lottie was the default answer for care home queries in dozens of UK regions.

The pages had three structural problems that did not matter while traffic was rising.

First, low content depth per page. A typical Lottie city page in 2023 had a templated intro, a list of homes pulled from the database, and a generic FAQ at the bottom. The substantive content was minimal. The content that did exist was largely the same on every city page, with the city name swapped in.

Second, weak verification. Lottie listed every home in the country, regardless of CQC rating, regardless of whether the home was currently accepting residents, regardless of whether the listing had been updated in the last 12 months. The directory traded breadth for trust. They had 14,000+ homes listed at peak. Most of the listings were stale.

Third, no real editorial moat. The articles section was thin and generic - "How to choose a care home", "What to expect from respite care", and so on. The kind of content any competitor could write in a weekend.

What happened in 2024

Google ran two major core updates in 2024 (March and August), plus a "helpful content update" that bled across the year. The combined effect was a savage downgrade for the pattern Lottie had built.

What the updates were targeting was clear from Google's own statements. Pages that exist primarily to capture search traffic rather than serve the user. Pages with thin content that is templated at scale. Pages where the user clicks through and finds substantially the same content as the previous result, just with different city names. Pages that promote a thin directory over the substantive information the user is actually looking for.

Lottie had all of those structural traits. The collapse was the algorithmic answer.

Their traffic loss was concentrated on the templated city and regional pages. Their CMS-driven editorial articles lost less, but were not pulling enough weight to offset. Their best-performing pages going into 2024 were also their thinnest pages in terms of unique substantive content.

The business response was visible in the company's hiring and communications. Editorial roles were added through the second half of 2024. Page consolidation happened. Some thin pages were removed entirely. The recovery has been partial - traffic stabilised at around 15 percent of peak by mid-2025 and has slowly climbed from there.

What this means for any UK service directory

Three lessons.

Templating without depth gets punished

A city page that exists only because the city name was substituted into a template is the kind of page Google is now actively downgrading. Every city page on a service directory has to earn its keep with content that is genuinely specific to that city. Not "dog walkers in Bristol love the green space at Ashton Court" in a template that swaps Bristol for Liverpool. Actually different content per location.

This is harder to write. It is also what users genuinely want.

Verification has to be a moat, not a marketing claim

Lottie listed homes regardless of whether the home was accepting residents, regardless of CQC ratings, regardless of whether the data was current. The directory was a search index, not a verified service.

The opposite approach - listing only verified, currently active, licensed providers and surfacing the licence data visibly - is more work but more defensible. It also matches what users want. People looking for a care home want to know which homes have current good CQC ratings. People looking for dog daycare want to know which operators hold current council licences.

Editorial has to be real

Lottie's articles section was thin and generic because the business had not invested in it. When the templated traffic collapsed, there was no editorial layer to fall back on.

A directory that publishes substantive, opinionated, specific editorial - posts that say something real about the category, that take a position, that engage with how the category actually works - has a layer that does not collapse when one Google update fires. The editorial layer also drives the kind of links and citations that protect the rest of the site from algorithmic instability.

How we are building GoodHound differently

Three structural choices.

Licensed-only directory, not aggregated

We do not list every dog care business in the UK. We list businesses we have matched to a real council licence under the DEFRA 2018 regulations. That cuts our database to something like a quarter of what an aggregated directory would have. It also means every listing has a verifiable trust signal that no other UK directory shows.

You can see this on every business page. The council, the licence number, the star rating, the issuing authority, the most recent inspection date. Not our scoring. Theirs.

City pages with genuinely specific content

Every city hub on the site has substantive editorial content specific to that city. Local parks dogs are walked in. Local council licensing quirks. Local price norms. Local operators of note. That work was real - it is what the editorial team has been writing across early 2026.

The city pages are not templated. They are written. That costs more. It is the only thing that survives the next Google update.

A blog that takes a position

This post is one of seven we have published this spring. The others cover the DEFRA licence framework, regional dog walker pricing, vaccination protocols at daycare, home boarding vs kennels, how to spot a bad daycare on a visit, and why we do not take commission.

Each one says something specific about the category. Each one takes a position. Each one is the kind of content you would not write at a directory that was running the Lottie playbook because it does not directly drive bookings.

It does drive trust. Trust is the moat.

The honest summary

Lottie are not a cautionary tale because they did something wicked. They built the directory in the pattern that Series A investors funded in 2022 and they scaled it the way the playbook said to scale it. The Google updates of 2024 are not a comment on Lottie specifically. They are a comment on the playbook.

If you are building a UK service directory in 2026, the playbook has changed. Verification is the gate. Editorial is the moat. Templated pages without unique content are a liability now, not an asset. The bigger directory does not automatically win.

We are betting on that. So far it is working.