Friday, July 27, 2018

Get Ready for CDP Horror Stories as Customer Data Platforms Enter the Trough of Dillusionment

It’s nearly a year since Gartner placed Customer Data Platforms at the top of its “hype cycle” for digital marketing technologies. The hype cycle shouldn’t be taken too literally but it does capture the growing interest in CDPs and reminds us to expect this attention to attract critics.

Sure enough, we’ve recently started to see headlines like “Customer Data Platforms: A Contrarian’s View”, “Why Your Customer Data Platform Is a Failure”  and “CDPs: Yet Another Acronym That Lets Marketers Down”.  It's tempting to dismiss such headlines as competitive attacks or mere attempts to piggyback on wide interest in CDPs.   But we should still take a look at the underlying arguments.   After all, we might learn something.

Let’s start with the “Contrarian’s View”, written by Lisa Loftis, a customer data industry veteran who current works for SAS. She offers to debunk two common CDP “myths”: that “CDPs solve a problem unique to marketing” and that “'marketing-managed' means you don’t need IT’s help”.

Regarding the first myth, Loftis says that systems to match customer identities have been available for decades and that departments outside of marketing also need unified data. Regarding the second, she states its best for marketing and IT departments to work together given the complex technical challenges of marketing systems in general and customer data matching in particular.

She’s right.

That is, she’s right that these technologies are not new, that unified data is useful outside of marketing, and that deploying CDPs requires some technical skills. So far a I know, though, she's wrong to suggest that CDP vendors and advocates (obviously including me) claim otherwise. False belief in these myths are not the reasons marketers buy CDPs.

To put it bluntly, the problem that CDP solves isn’t the lack of technology to build unified customer databases: it’s that corporate IT departments haven’t used that technology to meet marketers’ needs. That failure has created a business opportunity that CDPs have filled. It’s the same reason that people hire private security guards when the government's police fail to maintain order.

And, just as good security guards cooperate with the police, CDP systems must integrate with corporate systems and CDP vendors must work with corporate IT.  CDP vendors have designed their systems to be easier to use than traditional customer matching and management technologies, but that only reduces the technical effort without eliminating it. The remaining technical work may be done by the CDP vendor itself, by a service provider, or even by the corporate IT group. The term “marketer-driven” in the CDP Institute’s formal CDP definition is intended to express this: marketers in control of the CDP, which isn’t the same as doing the technical work.

“Why Your CDP is a Failure” offers an even more provocative headline. But hopes for juicy disaster tales are quickly dashed: author Alan J. Porter of Simple [A] only means that CDPs “fail” because customer data should be shared by all departments. Again, no CDP vendor, buyer, or analyst would ever argue otherwise. There’s no technical reason a CDP can’t be used outside of marketing and some CDP vendors explicitly position their product as an enterprise system. The reason that CDPs are not used outside of marketing is that companies fail to fund enterprise-wide customer databases, not that CDPs can’t deliver such databases. Your CDP is a failure for this reason only if building such a database was its goal. That’s rarely the case.

“CDPs: Yet Another Acronym That Lets Marketers Down” starts with the airy assertion that “When you strip all the nonsensical nuances away from these companies -- the CRMs, the TMSs [tag management systems], the DMPs, the CDPs -- they’re all one simple thing at their cores: identity companies.”  This will be news to people who use such systems every day to run call centers, manage sales forces, capture Web site, run advertising campaigns, and assemble detailed customer histories.

The article continues qirh assertions that “identity isn’t everything”, “brands don’t have a complete understanding of their customers”, and “behaviors without motivations teach us nothing."  Few would argue with the first two while the third is surely overstated.  But the relevance of CDP to all three is questionable.  It seems that author Andy Hunn’s main message is that marketers need the combination of anonymized third party data and survey panel results offered by his own company, Resonate. This may be, but Resonate clearly serves a different purpose from CDPs.  So there's little reason to measure one in terms of the other.

Let me be clear: CDPs are not perfect. Like many new technologies, they are often expected to deliver more than is possible.   We are surely entering the “disillusion” stage of the hype cycle when tales of failed implementations and studies showing mixed satisfaction levels are common (and prove nothing about the technology's ultimate value).  Critical articles can be helpful in clarifying what CDPs do and don’t offer.  It's easy to lose sight of those boundaries in the early stages of a product category, when the main task is building a clear picture of the problems it solves, not on establishing its limits.

This is why the most productive discussion around CDPs right now revolves around use cases. Marketers (and other departments) need concrete examples of how CDPs are being used.  In particular, they need to be told what applications typically become possible when a CDP is added to a company’s marketing technology stack. These generally do one or more thing: combine data from multiple sources, share that data across channels, and rely on real-time access to the assembled data. It's these applications that justify investment in a CDP.

Complaining that CDPs don’t do other things isn’t very helpful – especially if CDP vendors don’t claim they do.  Nor is it a flaw in CDPs if other solutions can achieve the same thing.  Buyers can and should consider all alternatives to solving a problem: sometimes the CDP will be best and sometimes it won’t. It takes a clear understanding of each possibility to make the right choice.   Blanket claims about the value or failures of CDP may be inevitable but they don't really advance that discussion.





No comments:

Post a Comment