Home > What’s UP > blog > Customer insights in action: 4 complex customer insights to boost digital banking sales

blog

11. 14. 2018

Customer insights in action: 4 complex customer insights to boost digital banking sales

Simple data-driven customer insights offer great benefits to banks but why stop there? To maximise results, you might also want to give more complex insights a go. That’s what we’re focusing on in the last part of our series exploring how customer insights can drive digital sales. Ready to dive in? Scroll down for examples and related use cases.

Over the past few weeks, we’ve covered how data-driven customer insights are getting ever more important as financial institutions are shifting from traditional demographic and geographic segmentation methods to hyper-personalisation when targeting customers. We’ve shown you some simpler customer insights and relatively easy-to-implement use cases, which can provide huge benefits. But the more complex a customer insight, the more use cases it can offer to financial institutions. Here are four such solutions and related use cases.

Predict financial difficulties

With the budget control insight, banks can not only predict whether the customer will face operative financial difficulties in the short run but can also tell how long they can sustain their current spending levels before their account balance drops to a dangerous level or slips into negative territory.

How? This solution uses machine learning algorithms based on a huge set of strong and weak variables. The insight is triggered by a high probability of a ‘zero-balance event’ in the near future and the bank can allocate meaningful actions (i.e. campaigns) to it. Possible follow-ups include sending an overdraft offer or some useful advice, depending on the financial context.

The use of this actionable customer insight can be complemented by specific customer segmentations, such as segments based on various types and regularity of income as well as bank account usage. The latter segments differentiate customers based on the relative amount of POS transactions, ATM withdrawals, bill payments and other transaction types.

Let’s see a use case. Alex, a retail banking customer, has recently bought a new flat and has started spending heavily on new furniture, consumer electronics and home appliances in the past three months from his savings in his regular bank account. With his income levels unchanged, increased spending has become a new trend in his finances and has affected his account balance significantly. Alex is warned by his bank that his balance is likely to drop to dangerous levels if he continues spending like this. The bank also offers him a credit card or as an alternative, a short-term credit to provide liquidity for the following two weeks until his next salary day.

Find new clients for existing segments

Our next example is actually a targeting tool rather than an actionable customer insight, although it can be considered a method for coming up with multiple new customer insights. Developing similar profiles, as we call this tool, allows banks to expand customer segments for marketing campaigns by adding new, similar customers to existing segments – based not only on demographic or other static data, but also using information on behavioural and spending patterns, and geo-location.

New members that can be included in current customer segments are identified by machine learning applications that analyse several hundreds of variables. A similar mechanism is used by Google’s AdWords service, which helps advertisers find potential new customers and simplifies targeting people who resemble existing website visitors. Facebook’s Lookalike Audience also relies on this method, identifying the common characteristics of potential new users who can then be added to the source audience previously selected by the advertiser.

This insight can help banks find new customers similar to those who, say, often buy premium Tommy Hilfiger clothes or eat out at Nobu restaurants when abroad. After these customers are identified, the bank can target them with a campaign offering credit cards, for example.

Tap emerging new customer segments

With the help of unsupervised machine learning techniques that can spot new lifestyle segments, data scientists can come up with new customer groups for banks to target, based on variables that haven’t been used or even considered before. Mapping new customer patterns can uncover new, previously unknown segments. They can be triggered by life events or other changes in a customer’s behaviour that moves them into a new group of potential customers.

One example is clients with flexible or freelancer income. People in this category have been using services like Airbnb, Etsy or Ebay, and spend money on Facebook advertising campaigns. They normally have foreign exchange transactions, which gives the bank a good opportunity to offer them savings products. The new segments profile can help banks predict future income trends, using external sources of data, and come up with new campaigns offering credit or savings to freelancers in a special programme tailored to their very needs.

Use a well-targeted merchant programme

Many banks run merchant programmes for retailer partners to better target banking customers. But more often than not these programmes are not accurate enough, causing frustration both for the partners and for the financial institutions. Well-targeted programmes can put an end to these frustrations and bring more satisfactory results. They provide information on when the customer’s likely to be at a merchant location and how much time they will spend near there. They also show how far the customer currently is from a merchant and what the right place, the right time and the right channel is for contacting them.

Based on the shopping habits and other behavioural patterns of customers, the programme allows for laser-sharp targeting and boosts conversion rates. In addition, targeted merchant programmes can also examine when, how and through what channel the customer responded to a campaign. Plus, it provides AI-based target audience recommendations, predicting the probability of purchases and also what the client’s likely to buy.

Here’s a potential use case to show how this works in practice. Based on location and mobile activity data, John’s bank discovers that he’s a cycling enthusiast, and often takes certain cycling routes at weekends and even bikes to work. After analysing device activity, speed and frequency, the bank can target John with third-party sports insurance or merchant offers from cycling shops he frequently passes by. These shops, in turn, can be selected from the bank’s existing portfolio of small and medium-sized corporate clients.

Importantly, none of the above four insights and use cases will bring the desired results unless banks go digital end to end in customer journeys. Just imagine how effective these methods are if the customer still needs to visit a branch to apply for a mortgage. Financial institutions must have simple, fast and seamless digital processes in place to benefit from insight-driven campaigns.

For more customer insights and use cases, download our white paper Segments of one: customer insights in digital banking.