How to Drive a Data Transformation

Let’s start with the bad news. Most organizations have a major data problem, and it’s an impediment to many improvements, such as automation and AI. Our own business is fairly straightforward, and even we’ve spent the last year clearing up our data in order to automate reporting. data transformation

The clean-up can be complicated. It consists of a large number of different problems, some related, some unrelated, some very large, some very small. Often these have surfaced over a period of years. And often the benefit diminishes, so you don’t necessarily finish—you may just target “good enough,” which is very difficult to define objectively. 

Now for the good news: it’s an important and complex job, but certainly not impossible. After all, we’ve done it for ourselves and dozens of clients. If you need to clean up your data, here is our brief guide to the process. 

1. Make data transformation a priority with a budget. Try to identify the scale of the potential benefit of getting it right and use this to justify a budget. Identify internal experts for whom this will be a priority—and work through how to keep it a priority alongside their existing commitments. 
2. Get clarify on reporting requirements. Often reporting requirements are a moving target, because people don’t know what they want until they see the wrong thing, and each report gives rise to a new question. So get clarity on who needs what and when they need it—but accept that the answers will be imperfect! Reporting requirements may change and develop, so without flexibility you risk a stuck job. 
3. Clarify your terms as well. What exactly do you mean by revenue, by a client, by profitability? Often the most basic terms have different meanings within a business, which causes inconsistencies in how data is collected and reported. Resolve these terms before you start.  
4. Prioritize the issues. Data can seem overwhelming. The only way forward is to start breaking down the problem. Build a list of the types of data issues; categorize them in terms of do-ability and benefit; and then order them in terms of priorities. Review the list and the progress frequently, because as we mentioned above, factors will change. Perhaps the list was incomplete at the start but improves as the fog lifts. 
5. Work out your governance process. There is no point cleaning up data unless you keep it that way! It doesn’t have to be overly complicated, and the process can depend upon the size and resources of the organization. But it does have to be clear and ongoing, or you risk going back to step 1. 

In short, the key is to be systematic about data—about getting it clean and keeping it that way—so you can use it to its best advantage.

Data is the basis of so many potentially transformative improvements in a mid-market business; give it the attention it deserves, and you’ll see a real difference in your own company.