Why Your Customer Data Platform Needs a Brain
Deploying “smart” automation is difficult. It requires two challenging tasks: maintaining rich, accurate, and current customer data and making the right decisions when using that data. Here's the case to why CDPs should do both
A few weeks ago, a colleague used the term “smart” marketing automation. I was skeptical at first since the term implies there could also be “dumb” marketing automation. But, on reflection, she had a valid point: it is indeed possible for marketing automation to be “dumb” in the sense of applying simple rules regardless of the situation. This can easily do more harm than no automation at all. “Smart” marketing automation, by contrast, would use the power of automated processing to create added value by executing meaningful interactions across channels and at scale.
Deploying “smart” automation, in this sense, is difficult. It requires two challenging tasks: maintaining rich, accurate, and current customer data and making the right decisions about using that data.
We know that Customer Data Platforms create the unified, persistent customer profiles needed to solve the first part of the challenge. But can a CDP also answer the second?
Before answering that question, let’s take a step back and consider what a solution would look like.
Start with the premise that any treatment you provide to your customer involves a decision about how to act – even if the “decision” is simply to execute “dumb” automation that treats everyone the same regardless of the situation. Those decisions can be made in three places:
- Separately for each channel, by the delivery systems (email, web site, mobile app, call center) that manage those channels. This is the least desirable option since it requires great effort to coordinate treatments across channels, and any “smart” automation must be re-created separately for each channel. In this model, the CDP simply feeds raw data to the channel systems.
- In a separate, central decision system that pushes its choices out to the channel systems to execute. This avoids the inefficiencies of channel-based decision systems, although it does require connecting the decision system to the CDP. While CDPs are designed to allow such connections, there’s still considerable work involved to make them happen, and there are likely to be limits to how quickly and thoroughly the CDP data becomes available.
- In the CDP itself. This avoids the need for separate integration. This is much more efficient. But it only makes sense if the CDP provides a powerful decision capability. Otherwise, the system will only be an efficient way to deliver dumb automation.
Most companies today are stuck with the first option – of channels making their own decisions – even though it’s clearly the least efficient and least effective. The problem will only get worse as channels proliferate, and customers expect more personalized treatments that are coordinated across channels. Growing privacy regulations will also make it more critical for companies to centralize how customer data is stored and used.
All things being equal, most companies would rather buy a single system that combines CDP and decisions with two separate systems. But all things are never equal, so the best choice between options two and three isn’t always clear. It’s a matter of comparing the integration costs of separate systems against the power of an integrated decision engine.
Some systems are easier to integrate than others, and some embedded decision engines are better than others. So buyers need to look at the specific options available to them to decide which is better in their situation. Factors to consider include:
- the decision capabilities you need
- the features of your existing decision system, if you have one
- the features of alternative decision systems, both stand-alone and embedded in a CDP
- the costs of integrating the systems under alternative options
- the quality of integration available under different options
But let’s go back to the core issue: creating “smart” automation, which means providing customers with meaningful interactions across channels and at scale. Assembling a beautiful customer database is only part of the solution. Companies only gain value from that database when they use it to make and execute the right decisions. Running a centralized decision engine is more efficient than running separate decision engines in each delivery system. What’s not clear is precisely which decision engine is right for you or whether that engine should be part of your CDP.
If you’d like some help with those choices, I recently published a detailed paper sponsored by Optimove titled Why Your CDP Needs a Brain: The Case for Embedding a Decision Engine in Your Customer Data Platform. I encourage you to use it as a guide when thinking about these questions.