Customer Data Quality: What Is the Value-at-Risk?

January 7, 2015

VaRIf you work in financial services or, more specifically, Capital Markets, then you are likely to be familiar with the concept of VaR or “value at risk.” VaR is a statistical calculation used in finance to incorporate a quantifiable measure of the financial risk that an asset (or a portfolio of assets) will decline in value.

VaRIf you work in financial services or, more specifically, Capital Markets, then you are likely to be familiar with the concept of VaR or “value at risk.” VaR is a statistical calculation used in finance to incorporate a quantifiable measure of the financial risk that an asset (or a portfolio of assets) will decline in value. The details of VaR can be complicated, but at a high level it calculates the maximum loss possible on an investment over a given time period, given a certain probability that events will cause that decline.

VaR is not a perfect tool – after all, what form of predictive analytics is? But it’s noteworthy in its recognition that you cannot assume the face value of something – or presume that the face value will last over time. It’s particularly true in finance, since investment values are determined by fluid market factors. But the concept is clearly not constrained to finance. It is something that firms who depend upon the value of their customers should also consider. In an interesting blog posting from earlier this year, The Financial Value of Data Quality, Boris Huard of B2B Marketing cited Gartner Group research that by 2016, 30% of companies will be directly or indirectly trying to monetize customer assets through direct selling or bartering. Companies that exist through advertising models are the purest example. To quote Apple’s Tim Cook [arguably writing about advertising–driven Facebook] “…when an online service is free, you’re not the customer. You’re the product.”

That’s not necessarily comforting of course. Imagine a business where the consumer’s name, contact information, and demographic profile are effectively a product SKU, access to which can be sold commercially. Maybe we don’t want to consider ourselves – or others – as product SKU’s – I certainly don’t. But every company should look to their customers as assets with a value that can vary over time – and where that valuation is considered material to the success of the business. As the Huard blog indicates, customers have a financial value that lives beyond a particular transaction. Smart companies try to understand that value and maximize it. It is the basis for calculations such as CLTV (customer lifetime value).

In finance and mergers & acquisition, the term “good will” captures the value of business “intangibles”, things like patent portfolios, copyrights, business reputation, brand value calculations, etc… Customers should be considered part of that good will calculation. Classic customer loyalty/retention research suggests that it costs 5-10 times more to acquire a new customer than it takes to retain an existing one. Selling something to that retained customer will generate more gross margin, since you have eliminated much of the “cost of sale.” But you need to know who your customers are (and where/how to reach them) in order to retain them.

Knowing your customers – and reaching out to them – is more of a challenge than many business people recognize. Statistics indicate that annual changes of address involve up to 15% of the national population. And the transience of millennials as a demographic pretty much ensures that their COA statistics are quite higher. Millennials will represent 40% of the population by the beginning of the next decade, finding them could be important to your marketing efforts. Now conventional wisdom might suggest that physical addresses don’t matter, given digital marketing and sales, but that can ignore the fact that address provides a key to getting demographic insights.

And, what about digital addresses as a way to reach individuals? The Radicati Group estimates that the average user has 1.5 email accounts ( If you are targeting that consumer, which one do you use? Do you know whether it is the same or different people? What about phone number(s)? Consider that most working adults have business, mobile, and home landline numbers, meaning that they may have three potential contact points, with the home landline likely to be shared by multiple people in a household. But if you have multiple millennials sharing a landline, are they really a household? And, of course, if you are targeting millennials, the mobile number if more than likely the primary one – do you know which is which? Let’s not get too philosophical, but if a landline rings in an empty household, does it make a sound? Not if you’re trying to make a sale.

But we all know that social networks are the way to better understand people nowadays – right? What about those social accounts? Radicati claims that the average user has at least three different accounts. If you consider the aliasing and anonymity that is rampant in media like Twitter, etc., how confident should you be in appending social data to your consumer records for greater insight?

Customers are valuable assets to a business, but that valuation is directly tied to the accuracy of your customer information. Too many businesses are focusing on amassing more and more data about their customers and assuming the quality is inherently there. It is not. More likely, the quality is declining since the rigor of data capture is just not part of the data gathering process. Or they are presuming that quality is a minor concern and that trending and cohort segmentations are enough. But neither trends nor cohorts make purchase decisions. Individuals do.

The reality is that customer information is valuable, but that value is at risk when a company tries to turn that intangible into tangible results. We are increasingly reliant on automation to target and communicate with [potential] customers. The automation of marketing and customer care offers scale, but little flexibility or adaptability to the vagaries created by poor data quality. If you have executed marketing campaigns, then you know the constraints in terms of bounces and poor CTRs. Without proper attention to the quality of your customer data, targeting decisions are likely to be made incorrectly and executed improperly. The more finely calibrated your messaging and offers, the more likely the negative impact is will be more than a rounding error – it could be a serious hit to your P&L.

Investment firms know how to calculate VaR. Do you?