What is the data hairball?
As I mentioned earlier, I use “hairball” as a metaphor to describe the complicated jumble of interactions, applications, data and processes that accumulate haphazardly when companies are unprepared to handle information from a wide range of sources. It’s different than the “data deluge” or the “sea of data” we’ve all read about; sticking with that kind of imagery, the data hairball would be the shoreline after a tsunami, but prior to reconstruction.
To me, the data hairball embodies both the
So, how can you unravel all the complications and start harnessing the power of new data-driven marketing strategies? Once again, it’s critical to take it step-by-step. My upcoming book explains the process in much more detail, but here are the basic eight points I suggest you follow:
1. Define the vision. What customer experience do you want to deliver? Research the customer journey as it is now. Then, paint a picture of the future. You goal is to make that vision a reality.
2. Outline the questions you need to answer. Which business questions do you want the data (and the team) to answer? Too many projects end in disappointment because they fail to keep the outcome in mind. If you aren’t sure what questions to ask, conducting a discovery session with all the key stakeholders can help them bubble up organically.
3. Assign the right team with the right sponsorship. Make sure you bring together people who “get it.” You’ll need senior-executive alignment and support, and you’ll need the team to reach deep into the organization, across multiple departments and geographies. Plus, everyone involved needs to be willing to challenge the status quo as needed.
4. Identify the data requirements. Be certain you understand what types of data you’ll need to drive the desired customer experience. Look at the data you can currently access, and then map your future needs as they relate to your present abilities. You’re bound to find gaps, and that’s okay, because next you’ll need to . . .
5. Find the source of the data you need. As you take inventory of what data exists and where, make sure you look across the entire enterprise. Who knows? Other departments, such as R&D, customer support, inventory management and business operations, may actually be collecting and storing the data you need. Add this information to the map you established in step four, then examine the remaining gaps and determine what additional data you need to collect.
6. Identify and ready the single source of truth. Most organizations typically use a combination of technologies to achieve a single source of verified data—what I like to call the “truth.” These enabling systems usually include a data model or organizational structure, an enterprise data warehouse to provide a single repository for organizational data, a big data analytics discovery platform that collects structured and unstructured data for analysis and insights, and a master data-management solution to build a single source of customer information to be used as the so-called golden record.
7. Consolidate, integrate and iterate your data. Once you have a single source of the truth, you need to populate it by bringing all the data together. Begin by consolidating and integrating the data to inform the strategy, campaigns and initiatives that will elevate the customer experience. Complete the process of unraveling this part of the data hairball by devising new collection processes to use going forward, and add governance policies based on what you learn.
8. Test, expand and evolve. Remember to measure and assess your progress. Start by answering the business questions developed in step two, and verify they are indeed the right questions. Look to deliver a few quick wins and identify landmines that need to be addressed before going any further. Chart out crucial points on your journey where more data will be available to improve a customer interaction or campaign, and then test those, as well. This iterative approach will improve results and build confidence in the data.
And remember: Start small.
Small-scale, pilot projects are valuable because they let you test your data strategy’s feasibility with fewer resources and less risk than larger projects. They’ll also keep you from feeling overwhelmed by the sheer size of your data hairball. As you focus on one strand, you may realize that your pilot project gives you the experience and knowledge to unravel the next strand. Experiment, test and learn. Keep untangling the knots, project by project. It won’t be long before you realize that your data hairball is actually becoming more manageable!
What’s coming up next week? I’ll discuss Step Four: “Make Metrics Your Mantra.”
