Try D&B Credit FREE for 30 Days! Get Started. Try D&B Credit FREE for 30 Days
Aim and Pivot: Avoiding 4 Dangerous Data Segmentation Assumptions

Aim and Pivot: Avoiding 4 Dangerous Data Segmentation Assumptions

Part 2 in the Series: The 8-Ball of Customer Portfolio Segmentation for Finance

You're standing in a smoky room and you've shot your break on the pool table, scattering the balls. It's time to understand the lay of the land. The end game - to shoot stripes or solids into the best pockets - is no different when you are eyeing balls of data, creating your winning customer portfolio segmentation strategy.

It's important to understand the rules of the game - to incorporate best practices, solicit advice from experts and ensure you have the right tools to do the job. You may be eager to dive in, but before you do, understanding some of the most common data billiard blunders will certainly help steer you on the path to success before you spend valuable time and energy constructing your strategy.

Segmentation efforts based on aging reports alone may keep your receivables from sinking your ship, but they won’t help you maximize opportunities or innovate on your approach to risk management.
Abigail Lutte, Dun & Bradstreet

To recap, why should CFOs and finance leaders be concerned with customer portfolio segmentation? If you're still enjoying your second cosmo with your Marketing colleagues, they will tell you that segmentation is integral to effectively wooing, winning and keeping the best customers - and that each segment of customers should command a unique treatment. The finance imperative is linked to this effectiveness. Without insight into financial data such as payment behavior, profitability or failure risk, the organization's sales and marketing segmentation efforts are shots in the dark, no matter how data-driven they appear to be. In Finance's corner of the cocktail party, segmentation efforts based on aging reports alone may keep your receivables from sinking your ship, but they won't help you maximize opportunities or innovate on your approach to risk management.

The difference between informational and helpful is key, according to Paul M. Perry of Warren Averett. "Knowing that a customer bought 12 widgets is informational. It is not helpful. But knowing a customer bought 12 items last month and six this month and has repeated that pattern for the last two years might provide insight you can use to better sell to that customer. Likewise, knowing that a customer's payment on credit has increased from an average of 12 days early in the year to more than 45 days at the end of the year can help you understand their struggles and alert you to move more resources to collecting that cash."

Now that you're ready to play, here are the most common areas for fatal blunders when performing your customer portfolio segmentation:

Assumption #1: What is visible is true. What is hidden is not.

Picture trying to pocket a ball in the dark. You can't see what you're doing, but you think that because you sense where the balls might be since you know they are, theoretically, on the table that you can feel with your hands, that your odds of hitting one are relatively good. And you assume that because you've played the game so many times, that you are in a good position to win without having the information that your eyesight can give you.

It's time to turn on the lights.

A Finance leader may understand the intricacies of data visibility and accuracy better than most. After all, SEC filings and Sarbanes-Oxley are legally binding reminders of how important visibility and accuracy are. The danger is in the dark room where your game is being played. You know what you know. You know how to play the game and you know where the table is. Yet, because you might assume that the data you're analyzing, collecting and acting upon is not only the most relevant and correct but also the only data that you need to make decisions, you might be especially prone to falling into the trap of eating a soggy bowl of data-frosted, gut-decision flakes instead of the meaty four-course conclusion meal you thought you were making.

The PwC 'Gut & gigabytes' global survey on data and analytics warns of the inherent bias in the data-analysis process. "Thus, big data needs human involvement to make sense of it. However, the process of analysing data also introduces a lot of biases that managers and executives bring to bear, particularly when looking at big data sets." The same survey reports that 71% of executives don't regard data and analysis as having the key role in their "big decisions." If that isn't frightening, nothing is.

Embracing data visibility means understanding and accounting for the quality, composition, velocity, origin and human manipulations of the data within your scope of decision-making. Furthermore, a responsible executive also asks the question, "What's not visible here?"

Sound like a big deal? It is.

The reality may be, when you turn the lights on, that your segmentation, in Finance, Sales, HR or anywhere else in your company has a high likelihood of being totally and completely wrong. Why? You may all be making assumptions based on the data sets that you know. Your data may be sitting in different systems that do not talk to each other. Your data may be generated by human hands. Layers of human analysis may have been performed prior to your review. Alarmingly, only 38% of companies share results of their analytic insights outside their departments, according to Dun & Bradstreet's 2016 Enterprise Analytics Study. Questioning what's visible early and often is a true pro tip. Once you and your colleagues collectively grasp the implications of the data you have, the data you need and collaborate on how to use it, the lights turn on.

Turning the lights on is overwhelming, but it's also invigorating. Once you see all the balls on the table, the possibilities open up. You aren't limited to looking at your customer segmentation strategy within the realm of financial ratios or payment behavior. A wealth of CRM data, industry profiles, social feeds, contact history and business linkages can and will enrich the way in which you look at your customers as a collective executive unit. And when you are all looking at the same table, with the lights on, the possibilities for growth are endless.

Assumption #2: All data is created equal.

Finance has a special relationship with Master Data Management, or linking key business information across the enterprise in cohesive records. Yet, even in the cleanest data sets on earth, more lurks beneath the surface than meets the eye. Each time a business transaction takes place, be it a customer service call, a supplier shipment or a stocked piece of inventory, a record is (hopefully) generated somewhere. It goes without saying that these records should be procedurally linked within an ERP or other system. The very act of creating a balance sheet each quarter necessitates some act of MDM, and yet, when data is not directly linked to a company's books, it can often take on a life of its own.

When a company goes to link any of its masterfully managed data with systems and data sets that aren't as closely scrutinized, the marriage between data sets can end as unscrupulously as it began. When customer segmentation is performed, it's tempting to stir your data in a cauldron and ogle at the colorful patterns emerging. But if there's one rule that every chef knows, it's that you can't mix cooked meat and raw meat together. The CRM, for example, can be a place fraught with data e-coli as a tool routinely managed and cleansed (or not cleansed) by human hands. Match your crisp and clean lettuce leaves of financial data with the Sales team's still-earthy customer profile data, and you might just have a recipe for a trip to the hospital. Ensure that duplicate records are accounted for, that data cleansing processes are understood and that linkages between departments and groups are identified prior to the great stirring of the segmentation soup.

Assumption #3: More analytics = better analytics.

Another temptation when performing enterprise-wide customer segmentation is to assume that more is better. As the great pool sharks know, it's not about having more stats - it's about making effective decisions within the game you're currently playing. Doing so requires powerful discipline. Organizations are awash in analytics today, but ensuring that the analytics are right for the strategy and the enterprise at large is an entirely different story.

Tell the right story with the words you have in your data sets. For example, you may decide to further segment and profile your customers utilizing their full credit limits by also looking at their likelihood to go out of business. Add information such as a score for propensity to buy and customer satisfaction, and your segmentation efforts can soon lose focus and gain complexity.

When segmenting your customer portfolio, do take some time to revel in the analytics available to you and establish what the best metrics are for measuring your desired path. But just as more shots in pool might mean fewer chances of winning, clarity and vision will ultimately help your enterprise succeed - not multi-page dashboards.

Assumption #4: The past is more important than the future.

Behavioral and descriptive analytical exercises typify the extent of many organizational efforts to segment customer data. It's the data devil you know, after all. It's far easier to segment your organizational data based on what's currently inside your systems. Just like any good game of pool, though, you have to develop an intuition about what your opponent's next move will be and how you will need to adapt in the future as you play the game.

Predictive and prescriptive analytics ask your data, "What next?" Don't be afraid to do some scrappy modelling on optimal pricing trends or on supply chain formations to scope out what the future might look for you. While what you have in your history is valuable, the audience watching you play pool cares only about your next move.

Ultimately, avoiding these common mistakes in preparing to segment your customer base will significantly increase your chance of doing it right the first time.

Now that we've gone over a few common areas of data danger, we'll explore a few best practices on customer segmentation in the next post.

Let's get in touch
See how Dun & Bradstreet can help you grow your most valuable relationships. Fill out this form and we’ll contact you soon.
How Good is your B2B Marketing Data? Get a Free Data HealthScan Get a Free Data HealthScan