Customer segmentation enables companies to get more bang for their buck by directing marketing efforts to the most likely targets.

Segmentation is a widely used technique in marketing, a way of dividing customers into groups based on their demographics, attitudes, buying behaviors, and other attributes. Using predictive analytics makes it an even more powerful tool – by identifying precise and nuanced target groups for campaigns.

The result is a win-win scenario – customers are offered more relevant products and services in ways that resonate with them, leading to more profitable relationships for the company.

Machine learning lets the data do the talking

Segmentation separates customers into distinct groups with distinct needs.

Most often, marketers use available data on customers’ demographics and behavior, as well as value (revenue they generate, and costs of the relationship) to manually divide them into segments for campaigns.

In contrast, predictive analytics automates this process, letting the data generate the segments.  This is done by building a predictive model with sophisticated software like IBM SPSS Modeler, which finds patterns in the data that are too complex or too subtle to come up with manually.

As an example, the following case study shows how customer segments generated in this way can be leveraged to more effectively cross-sell to customers.

DekaBank – Marketing mutual funds more effectively with predictive analytics

On their website, IBM shares a compelling case study

DekaBank is one of Germany’s leading financial services providers, operating in the wholesale banking and mutual fund segments. The bank launched a new guarantee fund (similar to a certificate of deposit) and wanted to determine which customers would be most likely to purchase their new product.

To select customers who would most likely be interested in guarantee funds, the CRM and Database Marketing team relied on statistical procedures to develop a scoring model with ten appraisal classes. The model used current and past customer data, along with socio-demographic, product-related and geographic variables already stored in DekaBank’s system. All “candidates” were appraised on the basis of the model and ranked accordingly. The customers best suited for a targeted approach were in the first class, and those least suited landed in the tenth class.

DekaBank put together a service package containing a list of the customers most likely to be interested in the new guarantee funds, as well as sales and promotional materials customized to appeal to these customers. This enabled the bank’s sales partners – the local savings banks – to make the right offers to the right customers.

The campaign was a great success. The number of transactions of the 160 participating savings banks rose on average by a factor of 3.3 (the highest by a factor of 8.8) compared with the savings banks that didn’t participate.

Similarly, Beyond the Arc’s experience working with Fortune 500 companies in the U.S., including banks, has shown how powerful predictive analytics can be in increasing marketing effectiveness. Clients also want to understand if there is an underlying logic beyond “the algorithms indicate…” This may mean that an analyst has to spend more time trying to identify and understand the key implications that the model raises. We’ve found that data mining models are only tools – the bigger wins come from using the findings strategically, for example to shape the messaging, timing, and flights of ad campaigns.

Applying predictive analytics in social media

DekaBank’s guarantee fund campaign offers a persuasive use case for predictive analytics. Can this technique be applied to the new frontier of customer engagement – social media?

It certainly can for social media-specific campaigns, as we’ve written about. However, it’s difficult to connect existing customer data to online comments, so using text data from social networks to help generate customer segments is challenging.  One way companies can leverage these insights is to extend marketing campaigns to the social media channel through their communications strategy.

For example, DekaBank determined that the typical purchaser of their new financial product would be an older and often long-term, particularly intensive fund user – and adapted their communications to appeal to this customer, from mailing and flyer text and layouts to phone script guidelines.  Had they decided to promote the product in social media, they would have needed to adapt their messaging, design, and medium choices accordingly.

Benefiting your customers and your business

Ultimately, companies want to maximize the value of their customer relationships without alienating customers.  One way to achieve this balance is applying predictive analytics techniques to improve customer segmentation – ensuring that the right customers get the right offer, and minimizing the perception of ‘spam.’

The more a company demonstrates it knows it customers and understands their needs, the better their experience – which in turn can reap significant rewards for the business.