It is rare that databases appear on the balance sheets of companies. Yet customers are an asset to a company and form part of a deal when a company is bought out.
The connundrum lies in the difficulty of putting a value on a customer. They are unpredictable at the best of times: what they have spent with you so far? What they are likely to spend with you in the future? What percentage of prospects will convert and how much will they spend? What is the value of the insight that you have on your database?
Clients must realise has to place a high value on consumer data and they really must get to grips with putting a price on this treasure chest.
Many companies protect their information internally, but marketers need to open peoples’ eyes that over and above actual customer value and how insight (behavioural (like shopping data) and survey data) can make a hell of a difference to a company in its market place.
It's something that Tesco is very good at, but how many other companies have really got anywhere nears their level of sophistication?
Whilst I was at Craik Jones I worked on The Boots account and we developed a new segmentation model for them with the analytics company, 5one.
Boots is the operator of the UK's biggest loyalty programme, with their highest value card holders receiving all the mailings. We proposed that we should identify medium value customers who had the potential to spend more and be more selective and scientific about the offers we made to them, therefore extending mailings to another 5m card holders.
We needed to identify customer value and potential so we could change their behaviour. To do this we needed to identify:
– the key triggers that influenced spend, collection and redemption of points
– distance from store
– type of store
– attitudes to health
– fitness.
Here's how we did it:
1) We used a 5% sample from two years of transactions across 14 million customers.
2) We created 2 models: an Affinity model - which looked at the likelihood of a customer buying in a particular category, and Share Of Wallet model – which looked at the highest spend a customer has spent on a product in the past.
3) If there was no spend information for the customer we then took the average for that peer group (a group of similar customers in terms of demographics and shopping behaviour).
The model identifies the person’s affinity to buying a particular product. These are: primary - they have bought the product before, and secondary - they have not bought the product but have bought similar products based on basket analysis and tertiary – people who have similar profiles as the peer group who have bought the product.
4) When planning for the statement mailings, all the offers were matched against the database, the primary affinity customers are then selected then secondary and tertiary, this is then cascaded for all the coupons until they are all allocated. Default coupons are used to fill any gaps.
5) From this information each mailing is tailored to select the best 8 products for the individual from a pool of around 50 offers.
On average we were getting an 800% ROI from the mailings.
In addition, the team were tasked with selling insight to suppliers regarding their brand and the categories they traded in.
This took a very long time to get underway, as Boots were fiercely protective of their data. Little realising, like lots of brands, they were sitting on a gold mine of information that if used correctly, would actually strengthen supplier relationships and not give away all their secrets.
John Wallinger is founder of The Marketing Planning Practice