Dr Sam Short, Chief Data Scientist at Upside Saving, explains how customer spending data can support marketing campaigns.
Marketing campaigns have a wealth of data to draw on. A company’s own sales data, website analytics, and different kinds of social media data can give a view of the customer to help make better marketing decisions, but it’s not a complete view. By adding insights from consumer spending data we can fill the information gap and give a holistic view of the customer.
But what is spend data and how is it obtained? Thanks to a series of reforms, known as Open Banking, consumers can opt-in to securely share their individual shopping transactions, regular payments they make, and companies they buy from, collectively known as spend data. Examples of companies that consumers share their spend data with include personal financial management apps (to display all of your bank accounts in one place), credit scoring companies (to “boost” your credit score), and cashback apps (to earn money as you spend with their retail partners).
It’s worth noting that consumers can only share their spend data with companies who are authorised and regulated by the Financial Conduct Authority (FCA); getting your hands on insights derived from spend data either requires a rigorous regulation procedure or a partnership with an already regulated company.
Spend data insights enables marketers to not only understand the type of consumer that shops with them – their age, financial situation, where they live, their hobbies and interests, whether or not they have a propensity to buy from them – they also help to understand how they shop both with their competitors and more generally.
To illustrate the value of spend data insights in a marketing campaign, consider a fictional coffee chain with a love of jazz music: Jazzoffee. Spend data shows that Ahmed buys coffee and he spends £100 on coffee each month. However only £20 of his £100 is with Jazzoffee. We also know that Betty buys coffee, she spends £20 on coffee each month and it’s all with Jazzoffee. There’s also Chris – they don’t drink coffee at all – and Dana who spends £50 each month on coffee, but none of it with Jazzoffee.
Detailed insights enable Jazzoffee to run a one-time hyper-targeted, optimised campaign that efficiently uses the marketing budget to maximise the return on investment. As a result, Jazzoffee won’t target consumers who will never buy from them, like Chris, or consumers who are already completely loyal, like Betty, but will instead focus on existing customers whose share of wallet can be increased, like Ahmed, and new customers, like Dana.
However, spend data-driven marketing doesn’t have to be a one-off thing. As consumers are constantly spending, the potential for incremental revenues is constantly changing. A direct connection to spend data makes it possible to build smart marketing campaigns that are driven in real-time by both spend data insights and data science.
To illustrate what a real-time spend data-driven marketing campaign would look like, let’s continue the Jazzoffee example. The data shows that Ahmed buys coffee every weekday in the early morning. He usually visits Jazzoffee 2-3 times per week, but never on a specific day. The other 2-3 times per week he visits one or two of Jazzoffee’s competitors. What might a smart campaign look like?
- Step 1, initial offer: 10% cashback for a week
- Result: Ahmed spends 5 times that week
- Learning: Ahmed’s share of wallet can be shifted toward Jazzoffee
- Step 2, test loyalty: no offer
- Result: Ahmed spends only 4 times that week with Jazzoffee
- Learning: Need to work harder to make Ahmed loyal
- Step 3: second offer: 5% cashback for a week
- Result: Ahmed spends 5 times that week
- Learning: Ahmed will shift is loyalty to Jazzoffee with a smaller reward
- Step 4, build loyalty: 5% cashback for another two weeks
- Result: Ahmed spends 5 times each week
- Learning: building loyalty
- Step 5, monitor the loyalty: no offer but be ready to react
- Result: Ahmed spends 5 times each week for 4 weeks, but then starts spending less frequently with Jazzoffee and goes back to a competitor
- Learning: Ahmed’s loyalty needs a refresher
- Step 6, new offer: 3% cashback for two weeks
It’s worth noting that the campaign wouldn’t give any offers to Betty and Chris, but might give a bigger initial cashback offer to Dana, to get her to engage with Jazzoffee, but subsequently smaller and perhaps less frequent offers than Ahmed, as she is worth less in terms of customer lifetime value (CLTV) than Ahmed.
Thanks to modern data science techniques – machine learning and multivariate optimisation – we don’t need to outline the various steps like we just did (“if this happens, then do this, otherwise do that”).
Marketers just need to specify the bounds within which the algorithms can operate (i.e. don’t spend more than £10k on marketing in a given quarter, no cashback offer can exceed 10% for two weeks at a time, and don’t target people who live outside of Newcastle) and what they’d like to maximise (i.e. return on investment, change in wallet share) and the data science will optimise the offers based on a consumer’s propensity to engage with the personalised offer, taking into account their lifetime value, and react according to any future changes in spending habits.
Whilst the coffee chain example is easy to understand, spend data-driven, hyper-targeted marketing isn’t only applicable to the retail sector; it can be used to market financial or utility products to consumers, too. Whatever you’re marketing, spend data enables you to easily quantify the success of a campaign; this is because it enables you to know how much of a person’s wallet was actually up for grabs. Success then becomes, not how many “new” customers you engaged, but how much of a person’s wallet were you able to acquire.