These days you can’t read a marketing trade publication or go to a conference without hearing about Customer Data Platforms (CDP). I have been in and around CDP for 6 years at three different companies: client, vendor, and agency. In each environment, I’ve seen the same question come up again and again in different ways: how do I measure the ROI on my CDP investment? In my experience:
Here are some of the lessons I’ve learned along the way.
Clients are generally unprepared to measure the investment
The vast majority of clients and practitioners are ill-prepared to understand the impact of a CDP on their business. I have seen two divergent schools on this and neither has generally been successful:
I once asked a room of prospective clients at an online retailer how they planned to know success. It was January of the new year, and the client had revenues of over $5B dollars. The senior-most person in the room spoke up above the silence and said something to the effect of, “we budgeted an increase of $25 mm dollars for this year.” While $25 mm dollars may sound like only 0.5% lift, that was a 100:1 projected return on their investment for year 1. Note, the implementation of the CDP had not yet begun. Plus, there was no plan for what would change in order to get the 100:1 return in the 8-9 months that would remain after implementation. Furthermore, there was no measurement plan to decide which revenues came incrementally from CDP – meaning it would be lost in the noise of a large business. Needless to say, this client was disappointed with their CDP investment. This type of measurement approach is, sadly, not uncommon.
More commonly than the unrealistic is uniformed. When I run workshops pre-implementation or even post-implementation with existing CDP clients, I always ask the question “how will you know if this is successful.” The most frequent answer I get is 5-8 seconds of silence, followed by “I don’t know, how do other clients do it?” A key difference of the uninformed from the unrealistic is that generally the unrealistic are very senior-level management. The uninformed are generally the execution level directors and below who are tasked with the implementation and adoption of the CDP.
The majority of customer data in a CDP is usually data that exists in data warehouses, ESPs, CRMs, etc. CDPs make it possible to move the most predictive data into activation systems – serving as a well-integrated layer to route those data and decisions. Thus, the measurement of the CDP depends on exactly what you’re trying to do with the CDP. Generally, these break into three schools of thought:
Most commonly, a marketing team is responsible for the adoption and usage of the CDP. The marketing team is powering customer experiences like ads, site personalization, email, push notifications, and the like. Therefore, most clients end up looking for efficiencies in the ROAS, LTV, Conversion Rates, and the like. The challenge becomes that a CDP shares audiences and user-specific attributes. The marketing team is using those audiences and data in campaigns. Unfortunately, most marketing technologies and even ad technologies make it difficult to understand what audience was reached in your campaigns. Unless you’re structuring your campaigns around audiences, this is not a trivial task. As such, we have a series of recommendations to consider which campaigns are CDP-powered and which are not:
One CDP client I've worked with automated the creation of hundreds of campaigns per month leveraging external data science, high-volume & high-velocity ticketing data, and some limited site interaction data. Prior to a CDP it took the client days for each email and ad campaign to launch – again, at a scale of hundreds per month. This meant many teams in many geographies were creating redundant audiences and were coordinating tons of audience data. This client measured their internal efficiencies based on the person-hours saved from the process. They attributed 37% & 80% improvements respectively to their ads & email time to launch.
This approach generally works best for clients who integrate CDP deeply into their product. The savings might be cost, time, personnel, etc. All of them should ultimately be a dollar value that is measured against the CDP.
In 2014, I was running a marketing team for a $50mm online retailer. When we looked ahead at the goals for the year, a key Q2 OKR was to be able to deliver the same personalized message to a user whether they were onsite, on Facebook, or in email. My team was experienced with a high volume of campaigns Facebook, our ESP(s), and our site. However, generally we were blasting the same message out with some more isolated personalized messages. To do the audience integration to all three channels, a CDP solved an obviously missing piece of the puzzle. For me, simply being able to “Add the Capability.”
In some less common scenarios, a binary success metric such as my example above will be the success criteria. That decision-makers who have this clear of an objective, I’d still recommend considering a longer-term ROI metric on that capability. For example, how much more effective is cross-channel marketing than the old business as usual? How many incremental dollars have been unlocked by the capability? That can unlock additional investment in your CDP since an individual capability may not be defensible in the long run.
Get in touch with a Lytics expert