Using data integration for better customer targeting
Web Analytics Wednesday is an event Lynchpin organise on alternate months. Each session explores an aspect of data analytics and we are fortunate to entice some great speakers. The March 2016 meetup was no exception with Nick Redding and Wayne Field from William Hill pairing up to present how they combine data science and digital analytics to create a personalised experience for their customers. Next stop for Nick and Wayne is the Adobe Summit and I like to think we got to see a top class act before they hit the big time, for a slightly reduced ticket price 🙂
3.7 million people gambled online last year, 1.9 million with William Hill. William Hill make 400 transactions per second and on their busiest day of the year, Grand National day, they did more business than Amazon on Black Friday. William Hill use customer modelling to understand their highest value customers.
Nick is Head of Optimisation and his team of analysts and optimisers look at the digital data coming in via the William Hill website, while Wayne heads up the Decision Science function. They are united in their strategy and as the presentation went on we were served a healthy mixture of smart thinking, clever data science, humour and Lynchpin’s favourite: common sense.
Big Data is Rubbish (well a lot of it is)
The first point was well made by Wayne: “A Lot of Big Data is Rubbish”. William Hill hold a vast amount of data on their customers, but not all of it has value in achieving their commercial objectives.
It’s really Wayne’s job to artfully identify and extract the most relevant statistics. For William Hill these include what sport people bet on, how often and to what value, and if they have placed a bet recently.
Statistics serves the purpose of identifying signal from noise, which becomes all the more critical as the breadth and depth of source data increases. While William Hill might also know a mosaic profile, which device the bet was placed on and the precise location when they placed it, in their models these variables have less significance in influencing customer value
A Bit About Personalisation
Second point: personalisation doesn’t need to be so complex. While creating a personalised experience is often touted as needing to be super detailed or precise, for William Hill:
“It’s about giving customers the softest possible landing when they hit your site and using the most relevant data points you have to make their life easier and find what they’re looking for.”
It’s important to serve customers relevant content, if they are a tennis fan make the website reflect their taste and when not much is going on in tennis world something else.
As William Hill have a business that is affected by seasonality, this is where more complex modelling comes into play: if you can’t serve (excuse the pun) that customer tennis because nothing hot is happening in the tennis world, what is the next best choice and then the third best choice?
As Nick said, “personalisation is not a technology: it is a process” – a comment that clearly resonated strongly with the Web Analytics Wednesday Crowd
Optimisation and personalisation is NOT a plug-in-and-play technology, it is a process. My favourite comment of the night #wawlondon
— Elise Maile (@E_Maile) March 23, 2016
Customer Modelling and Segmentation is the Real Key
Wayne and Nick work together to better target their customers, cross-referencing transactional IDs with clickstream data.
William Hill are better able to develop cross sell campaigns safe in the knowledge that the person has some interest in that product. A transactional error can be indication of churn prediction and they might use these as a way of tempting players at risk of lapsing with a win back offer.
By using this approach across all of the product verticals, William Hill has created a targeted “hot list” for each customer to drive CRM activity, with a resulting uplift in engagement from 30% to 225%
Analytical Common Sense and Top Tips
Wayne and Nick both say start small and scale, use control groups to A/B test personalisation, and within the data look out for outliers.
Finding oddities in data as our own analysts know comes with experience and clear interpretation in the context of a business. With William Hill you may get a high roller that messes up your averages. Technology can only get you so far: it’s having experience, enquiring mind and common sense to spot when something just doesn’t seem right.
The next Web Analytics Wednesday is 25th May
Find out more and sign up here
About the author
Lynchpin
Lynchpin integrates data science, engineering and strategy capabilities to solve our clients’ analytics challenges. By bringing together complementary expertise we help improve long term analytics maturity while delivering practical results in areas such as multichannel measurement, customer segmentation, forecasting, pricing optimisation, attribution and personalisation.
Our services span the full data lifecycle from technology architecture and integration through to advanced analytics and machine learning to drive effective decisions.
We customise our approach to address each client’s unique situation and requirements, extending and complementing their internal capabilities. Our practical experience enables us to effectively bridge the gaps between commercial, analytical, legal and technical teams. The result is a flexible partnership anchored to clear and valuable outcomes for our clients.