Customer Segmentation Moves into the Fast Lane

It is a common marketing dilemma: Companies know that targeting their customers with greater personalization and timeliness will unlock value, yet customer segmentation strategies often fail to deliver the expected ROI.

The problem is that traditional approaches to customer segmentation are fraught with difficulties. Consider a typical example: A company is seeking to identify customer segments with significant growth potential for a new product it plans to launch. It has captured 50 customer attributes, ranging from age to account tenure, and wants to pinpoint the best six leading indicators to base its segmentation upon. It sounds simple enough, but there are 15 million possible combinations when choosing six from 50 – about the same chance as winning the lottery. If you Ddouble the size of the data set, and the possible combinations total more than 1 billion.
Companies need a more advanced approach  that can handle infinite numbers of customer attributes, while also taking into account the business objectives of the project, rather than relying on the analytical skills of statisticians. The good news is this exactly what the next generation of customer segmentation solutions, based on artificial intelligence technology, can offer.

Advanced customer segmentation projects begin by considering the business’s goals and drawing on attributes from different dimensions of consumer information, such as their attitudes, behaviors and value. They must also produce insights that can be acted upon. The key is to address three fundamental issues:

• Who to focus on? In order for the business to develop tailored investment strategies, each segment identified must be highly differentiated on core value dimensions such as revenue, cost and profitability.

• What to offer and what to say? Each segment must be comprehensively profiled so that the business can make the best possible product offer with the most appropriate messaging.

• Where to find them? Each segment must be developed so it can be applied to a customer database or used to inform a mass-marketing campaign. Otherwise the project simply isn’t actionable.
Artificial intelligence technologies address these issues, using advanced data mining techniques to link analytics more directly and explicitly to business goals. The technology also enables the business to adapt the project as it develops, fine-tuning the segmentation strategy as preliminary findings prompt an evolution of the original objectives.

To test all this, one large American retailer used a segmentation tool to re-examine a project previously carried out that had generated solid results using traditional techniques. The aim was to provide insights that would help the retailer set its merchandising strategies using a set of 240 customer attributes and focusing specifically on shopping preferences and customer demographics.

The results were impressive. Where the original exercise only identified one small segment displaying significant differences from other segments, the new approach identified a series of segments with significant differentiation, helping the retailer better understand who to focus on. The series included both large, superior segments and poor performers, enabling the retailer to adapt its profiling, targeting, messaging and product offerings accordingly.

For each segment, the tool was also able to offer unique profiles, providing the sort of rich insights that are crucial for the development of individualized product and messaging strategies. Quite simply, the greater the clarity of the segment identified, the greater the opportunity for marketers to produce something that is highly relevant to the target audience.
The Missing Link

One shortcoming of traditional approaches to customer segmentation has been the difficulty of answering the crucial question of who to focus on. The consideration has typically been about what to offer and the messages to use because the technology has not been capable of drawing on attributes taken from different dimensions of consumer information.

In particular, the absence from many projects of a value dimension (e.g., profitability, revenue or net promoter score) has made it difficult to develop an investment strategy for each consumer segment. That has also hindered the marketing function’s efforts to persuade other parts of the business to support and use its segmentation strategies. Marketing has lacked the ability to demonstrate the potential financial impacts of addressing the needs of a particular customer segment. Were it able to do so, the business would pay more attention. Advanced customer segmentation techniques can do this – and companies that have adopted such solutions are now finding clear benefits.

In one example, a financial services company was able to use next generation techniques to identify two key segments: one, approximately 5 percent of the customer base, accounted for 50 percent of profits, while the other, about 10 percent of the base, was eating into 10 percent of profits. This segmentation, which balanced value, demographics and customer behaviors, enabled the company to refine its marketing messages and customer acquisition strategies. The end result was a 600 percent ROI in the second year.
With results such as these, businesses that fail to adopt segmentation techniques that leverage artificial intelligence technologies, advanced analytics and big data will begin to suffer serious competitive disadvantage. Advance segmentation solutions will enable companies to identify segment profiles that are richer and increasingly comprehensive. In turn, those businesses’ marketing strategies will be far more effective as they personalize and target product offers and messaging to the most relevant customers.

(Author's note: To read more about customer segmentation, download Accenture’s recently released analysis “Taking a Fresh Look at Customer Segmentation: Winning the Lotto").

For reprint and licensing requests for this article, click here.