SEO Experimentation Partnership
Marks and Spencer (M&S) is one of the UK's most iconic retail brands, with a heritage that dates back to 1884. M&S partnered with SearchPilot to ramp up SEO testing and bring SEO to the same level of experimentation excellence already in place in other digital channels throughout the business.
Download the full
customer success story
Our key discovery
Many enterprise brands have a mature SEO setup and a strong culture of experimentation, but they still struggle to prove what SEO is worth.
M&S showed it doesn't have to be that way. When SEO is tested as rigorously as your other channels, its value becomes hard to ignore. In just three months, M&S reached 190% of its annual SEO target, and the multi-year SearchPilot investment was paid for many times over.
The challenge
Due to the competitive nature of online retail, M&S already had a mature SEO setup, with SEO's importance and visibility growing in recent years.
The SEO target was rolled up into the overall e-commerce target, which was ambitious and set with revenue generation in mind.
When asked how he planned to achieve M&S’s ambitious organic search revenue targets, their SEO Manager proposed integrating SearchPilot’s enterprise SEO testing technology into M&S’s leading experimentation team.
The team needed a reliable SEO testing platform to run more tests, help meet the company’s organic search revenue targets, and access the data needed to showcase how SEO translates into revenue.
Our solutions
Advanced SEO test bucketing
SearchPilot collaborated with M&S’s online and technology leadership teams to integrate its application programming interface (API) for advanced test bucketing, which identifies control and variant groups of pages for SEO tests at scale.
Deploying the API enabled M&S to fully leverage advanced bucketing and analysis capabilities. This included the ability to test anything about the HTML, structured data, design, or the layout of product detail pages (PDPs) and product listing pages (PLPs) without requiring internal data science resources for each test.
Complete integration with M&S’s existing processes
To make the experimentation program as successful as possible, SearchPilot needed M&S’s engineering team to build tests quickly. This led the teams to collaborate on an SEO experimentation program that fully integrated with M&S’s existing engineering, quality assurance (QA), and release processes.
Finding and deploying winning changes at scale
The M&S and SearchPilot teams designed a high-cadence testing program that capitalised on the website's large scale. By building a backlog of testing ideas and intelligently segmenting product listing pages and product detail pages into statistically similar subsets, the teams could run a large volume of tests in parallel to find winning changes. Then, thanks to SearchPilot’s intelligent API, winning tests could be immediately rolled out to 100% of the applicable pages.
Professional services collaboration
SearchPilot’s professional services team supported the core M&S team and its agency partners in test ideation and design, measurement, and analysis. When technical and compliance issues affected the reliability of exporting daily analytics data, the teams collaborated on a solution:
Using statistical analysis, they demonstrated that Google Search Console data could supplement missing analytics data with high accuracy. The SearchPilot team then quickly released a new Google Search Console integration to deliver the insights the business needed.
This builds on an existing SearchPilot feature and enables Google Search Console click data to serve as the primary metric for analysing test results (instead of Google Analytics session data). Solving this issue was critical to the success of the whole program.
Making SEO results ‘speak the language’ of business
M&S needed to translate specific channel metrics into business metrics that could be forecasted and analysed. To do this, the internal and external SEO teams collaborated with the analytics team to develop a robust revenue forecasting methodology that everyone could stand behind.
They combined revenue per session with a decay factor that represents how long uplifts from winning tests are expected to last, to account for changes Google and its competitors make. The team could then feel confident that their hard work translated into increased revenue.
