Why strong AI visibility wasn't enough for Sykes Cottages

An SEO-led business facing a changing discovery landscape

Sykes Cottages had built a highly successful organic search business, supported by extensive location based content reflecting the different permutation on how someone may search for a cottage, from region to postcode.

As organic traffic began to decline, Sykes suspected that AI-powered discovery was contributing to the change. The business wanted to understand how prominently it appeared in ChatGPT and identify what it should do to adapt for the LLMs.

Sykes had already started using a GEO measurement platform, Peec, but its existing AI visibility data provided limited strategic guidance.

The business needed to understand where it was being overlooked, which competitive gaps mattered most commercially and what practical changes could improve its position.

The engagement was commissioned by the Group Chief Commercial Officer of Forge Holiday Group, with involvement from the Group Marketing Director and close collaboration with the Group Head of SEO and content.

Measuring a complex holiday market

Holiday accommodation searches rarely involve a single requirement.

A customer might be looking for a dog-friendly cottage in Cornwall, close to the beach, with three bedrooms and a hot tub. Another might need a large property in the Lake District for a family gathering near a pub.

Each requirement can influence which accommodation providers ChatGPT recommends.

Geography adds another layer of complexity. While Sykes competes with major accommodation platforms such as Airbnb and Booking.com nationally, regional specialists can be particularly prominent within certain areas.

So the competitive set North Wales could therefore look very different from its position in Dorset or the Lake District.

There was also a fundamental measurement challenge. ChatGPT can itself recommend an individual holiday cottage, a specialist regional operator or a national accommodation provider. Sykes needed to understand whether it was being recommended as a place to find and book accommodation.

Existing SEO data was utilised but offered limited guidance. Keyword volumes and clicks did not necessarily reflect how customers expressed their holiday requirements to AI assistants.

Sykes needed a measurement framework that reflected its business, its customers and the commercial importance of different opportunities, rather than simply replicating its existing SEO strategy.

Zebora began with a series of workshops to understand Sykes' business, customer needs and commercial priorities.

From this, we developed a bespoke measurement framework comprising 665 prompts across 161 different reporting dimensions, designed to assess how ChatGPT recommended Sykes and its competitors.

Three elements were particularly important.

1. Understanding how customers search

We designed prompts around the different ways customers discover holiday accommodation, accounting for destinations, group sizes, duration, property requirements and customer needs.

We also distinguished between recommendations for individual cottages and recommendations for accommodation providers.

2. Understanding regional competition

We measured performance both at a national level but also at individual regional levels to show how the competitive landscape changed across different holiday destinations.

3. Reflecting commercial priorities

Working with Sykes, we applied commercial weightings to ensure the analysis reflected the relative importance of different opportunities to the business.

The result was a commercially relevant view of where Sykes was performing strongly, where competitors were outperforming it and where improvements should be prioritised.

Strong visibility, but weaker recommendation prominence

The initial audit found that Sykes was already performing strongly in ChatGPT.

Its visibility score was 89/100, sentiment score was 76/100 and it ranked first overall against its closest competitors, including holidaycottages.co.uk and Cottages.com.

However, its average mention position was fourth.

ChatGPT frequently included Sykes in relatively long recommendation lists, typically containing around 10-15 different options. Sykes was present but was not consistently among the top 3 recommendations.

The audit therefore identified recommendation prominence as an important focus and opportunity, even with Sykes' strong overall visibility.

The deeper diagnosis: Brand differentiation

Zebora's brand analysis identified differentiation as a central weakness.

ChatGPT recognised Sykes as an established, credible accommodation provider. However, its representation relied heavily on breadth and scale, attributes that did not clearly distinguish Sykes from other providers, who shared the same.

The audit identified this as a potential factor limiting Sykes' recommendation prominence.

So to investigate further, Zebora examined four important pages on the Sykes website: the homepage, About Us page, a destination page and a specific accommodation type page.

The findings revealed a consistent pattern.

Sykes had the credentials, but wasn't communicating their relevance clearly enough to the LLMs

The website contained substantial evidence of Sykes' credibility, including awards, B Corp certification, longevity and customer reviews.

However, these credentials were often presented as isolated facts rather than connected to tangible reasons for customers to choose Sykes.

A key finding

38% of all Sykes URLs cited by ChatGPT were the homepage.

Yet the homepage's extensive list of accommodation by location and collection types, often with multiple links, competed with its central brand messaging. The pages weren’t providing necessary signals to the LLMs of who Sykes was, what they do and more importantly, what makes them different.

The same pattern appeared on more specific pages.

The pet-friendly page provided limited structured information about dog restrictions, fees and other requirements that could influence recommendations to specific LLM questions.

The location page communicated inventory and geographic coverage effectively but offered less explicit evidence of why Sykes was a particularly relevant provider for that specific destination.

Across the website, Zebora identified opportunities to organise existing content and credentials more effectively, communicate Sykes' strengths and connect them to specific customer needs and requirements.