Home / Banking Strategies / Merchant attrition to merchant addition: Analytics as a payments catalyst

Merchant attrition to merchant addition: Analytics as a payments catalyst

Sep 5, 2017 / Consumer Banking / Payments

The payments industry has grappled with rising merchant attrition for quite some time. No longer is traditional customer service enough to retain a merchant. As a result, the industry has started to explore analytics to get a better handle on this problem.

But here’s the issue: Technology as applied to merchant attrition has delivered underwhelming results—and no significant headway has been made in finding an effective solution. That is not to say technology can’t solve this problem: merely that the industry needs to reinvent its approach to analytics. This can be done by gathering and analyzing data that allows payments professionals to see what merchant clients think of their services—and identify potential problems before attrition begins.

The attrition condition

By some estimates, more than 20 percent of merchants are likely to move from their current provider-independent sales organization or acquirer. Within that number, 16-17 percent represents voluntary attrition and hence, is possibly preventable. Industry consensus also indicates that turnover is expensive: It takes roughly three new merchant clients to make up for each one that leaves the portfolio. Due to alarmingly high attrition rates, the industry spends more than $1 billion every year to replace merchants they have lost.

It pays to look beyond price

Most conversations on attrition focus on the single biggest reason merchants leave a provider: price. What’s more, processors often consider price reduction the most important tool for retention. But price is not the only factor contributing to attrition.

In addition to new pricing dynamics, providers face evolving technologies and a new wave of competitors, all making it more difficult to retain existing clients. It’s much easier to retain a current client by noting and proactively addressing their concerns rather than scouting a new one. If providers don’t follow the issues as they bubble up—and only become aware once they receive a request to terminate services—it will be too late. Thus, they must shift from a reactive to proactive stance. That is where analytics can help.

Patterns and payments pain points

Payments firms have started to utilize predictive analytics to identify pain points in merchant relationships and understand when merchants are likely to leave. However, these efforts have been very restricted due to input data used to conduct the analysis. As one can expect, most predictive analytics are based today on segmenting the merchant and then analyzing their transaction feeds. Any reduction in transaction patterns, whether by volume or amount, is thought to indicate the likelihood of their leaving.

Though this indicator forms an important component of attrition analysis, it is definitely not the only, best or earliest one. Payment processors and their partners have barely begun to scratch the surface of what’s possible through analytics.

Clear data = clear insights

To produce effective predictive analytics, all possible factors that relate to merchant attrition must first be identified. For example, price can have multiple nuances, such as whether pricing is high relative to the other merchants in the segment or as a proportion to the merchant’s revenues. Next, we also need to identify other indicators that can show whether there is danger of a merchant leaving. Collected across merchant transactions and settlements, servicing and other contexts, the indicators should be tangible and measurable (such as merchant transaction and call patterns).

We have identified more than 40 such data points that represent “significant variables” businesses can easily make available for this analysis. These factors and indicators form the basis of building hypotheses around this question: Why does attrition occur and how it can be detected before the merchant’s departure is fait accompli?

From pattern anomalies to data possibilities

Once the data is in place, it’s time to decide analytics-based approaches to determine the likelihood of attrition. Regression analysis is almost a given, as it will determine the importance that these indicators play in helping understand the likelihood.

Approaches such as pattern anomaly detection—of transaction, engagement and service request patterns—are easily accessible in a structured format and hence quickly usable. So if call information is already available with a call center, it might be also available to the retention/relationship team. But with typical functional (and data) siloes in many organizations, this is not a given. Thus the analytics route could also serve as a mechanism to improve data flow across teams.

Pitting it all together: An attrition solution

The next phase in analytics looks at approaches such as entity resolution, network analytics and voice/sentiment analytics to leverage unstructured data that is often “outside the enterprise” in an analytics equation.

Entity resolution along with network analytics can help the provider identify each specific merchant across multiple channels and also gauge how much they might influence attrition of other merchants.

Voice, text and sentiment analytics look at data from voice recordings, call transcripts and contact notes to understand the “tone” of the merchant’s communication with the organization and what the merchant is saying—sometimes, implicitly. Recent advances in deep analysis of speech and text using machine learning make these approaches feasible.

It’s also extremely critical that the insights are presented in an intuitive, easy-to-consume manner to the people needed to take action—for example, relationship managers. But this aspect often fails to get the attention it really deserves.

Prescriptive analytics mark the next move towards the Holy Grail of any analytics program today: one that suggests exactly what needs to be done by the relationship, service, product and other teams to retain them.

There is still much work to do. But right now, analytics heralds a future working towards more successful merchant relationships. And so beckons a tantalizing analytics endgame: that merchant attrition itself might succumb to attrition.

Want more Banking Strategies? Sign up for our free newsletter!

Rajesh Kamath is head of financial services solutions and incubation at Incedo, a technology services firm specializing in data management, product engineering and emerging technologies. Kamath and his team help clients innovate in emerging technologies including big data, data automation, text analytics, machine learning, chat bots and predictive modeling.