Five questions to ask your data before inviting it to your CDP party
The power and the beauty of a CDP comes from its ability to harness customer data to create meaningful personalized experiences. But there's a big difference between data and wisdom, and it's important to take a good look at your data to make sure you can get the most value out of bringing it into your CDP.
The questions to ask your data are: Are you available? Are you linkable? Are you complete? Are you correct? And are you fresh?
Below I'll breakdown each of these questions and why they're important to your CDP implementation.
Are you available?
The first thing to figure out, is if your CDP use cases require data that you are tracking today. It might be really cool to create segments based on topics your clients want to hear about, but if your preference data only captures opt out, you will need to get a little creative, or make some changes, to create a preference-based segment in your CDP. You can update your preference center, so your customers can tell you what they want to hear about. You can also look at topics your clients are engaging with on your website or in your email campaigns and create segments that way. If you want to create segments based on your client's engagement with your loyalty program, but you don't currently have a loyalty program—that would be a much heavier lift.
Are you linkable?
The Holy Grail of a CDP comes with a unified profile where you can link data from disparate systems to see all the ways that your customers are interacting with your brand. To do that, we will need a key that unlocks our ability to link data together. CDPs give you many ways to create links: email addresses, names, phone numbers, locations subscriber keys, device data—in various combination—to create a unified view. Many CDPs will give you the ability to create segments based on a single data set, but with this approach you run the risk of creating disjointed and potentially duplicative interactions with your client.
Are you complete?
This question looks at the data you're capturing and discusses how much of it you actually have. For example, you might be asking your customer service desk to capture certain bits of information during customer interactions, but if the customer declines to give the information or your service agent doesn't put it in the system, while you are tracking that data point, you may not have enough of it to create a meaningful segment. In this instance you may want to implement training or process changes before bringing this data into your CDP.
Are you correct?
While you might be tracking customer information, the data in your system may not be correct. This is especially true for information that is entered by a human being. It's a good idea to gut check your data set and look for values that may be commonly misspelled, entered incorrectly, or just don't make sense. If you find this the case, take a minute to be curious because you may be able adjust the user experience when entering the information to make it more likely that the correct information is captured.
Are you fresh?
To create personalized real time interactions, you want to base those touchpoints on where your customer is today, not where they were weeks and months ago. If your data sets are old, you may want to evaluate if it makes sense to bring them into your CDP, or what the level of effort it would be to start tracking that information on a cadence that would be more meaningful.
But what if I find out my data is terrible?
Everyone has skeletons in their data closets, but most people can find enough data to get started and create a road map, based on priority and level of effort, to get the most important bits of data into shape.
You can create parallel paths in your implementation. So, while you are implementing one set of use cases, you can pick another set where you work towards the process and policy changes you need to make within your organization to get the data ready for your CDP.
The great thing about CDP data as opposed to some other data initiatives is that fresh data is the best data. For many use cases, this eliminates the need to clean up large swaths of historical data. You can fix the problem now or start tracking a key piece of information and the data will be ready to go by the time you are ready to put it in your CDP.
The best way to eat an elephant is one bite at a time
Taking on a large data project can feel intimidating, but it doesn't have to be. Here are three tips for success:
- Prioritize the data that will support your highest impact use cases.
- Start small and work in bite sized chunks.
- Start with the easy stuff. You will learn a lot along the way that will help when it comes to tackling the more complex challenges.