Step out of your data leader role for a moment and ask yourself a question: Would you tolerate inaccurate data in your personal life – in your investments, your healthcare, or even your car’s gas gauge? Probably not. Then why do we accept it, or hold it to a different standard, in our professional lives? 

In fintech, market data or financial information companies, the quality and accuracy of data are paramount. If you’re a credit rating agency whose role is to support efficient capital markets by helping market participants make informed, timely decisions , you require accurate data. That is what data governance is ultimately about: where do we need to have the highest possible level of confidence in our data? 

This is why I and my team are actively designing what I call a data certification process. These certifications tag each data set with confidence indicators that are set by the data owners who are primarily on the business side, so decision-making is based on a rock-solid foundation, and visible at the time that data is used. 

Would you prefer gold or bronze? 

Part of the inspiration for these conversations is based on a structure I implemented at a previous company. We actually certified data sets as gold, silver, or bronze. 

  • Gold meant that the data had met indicated thresholds for the highest levels of accuracy, so my colleagues could rely on it to make critical decisions
  • Silver didn’t meet as many requirements, but might be a go-to if no other data were available. It might also be suitable for less mission-critical decisions.
  • Bronze met fewer requirements than silver. You might use this data for indicative purposes but you would not make a significant decision on it until you did more research with other sources.

These certifications were not cosmetic. They were the visible result of our governance process. You could reference the metadata. Everything was well documented, which meant that we didn’t have conflicting sources. Gold data demonstrably came from the source of truth. In fact, based on our conviction of the importance of visible certification levels, our business performance and operational dashboards featured tiles where the gold, silver and bronze designations would appear in reference to the data upon which they were built.

Although we’d designed our data certification program more as an explicit measure for our executives, it quickly caught on. Six months in, executives viewing a dashboard without the certification badges would ask why they weren’t included. Even better, data owners started to put in the extra work to ensure that their data went through our governance process. Simply making a copy of a data source from a virtual storage bucket without inquiring into its governance status was not going to earn you Gold. 

Governance. Trust. Behavior change.

I mention all this not because I’m recommending an Olympics-style data race at your organization, but to point out that promoting data governance with a bit of “cultural innovation” can change and significantly improve day-to-day behaviors. Rather than saying, “You have to perform all these onerous tasks with your data or you’ll be publicly admonished,” give people something to aspire to. Make it part of fostering a broader data literacy. Data certification programs can build trust in your data, and therefore in your data culture, but they’re less likely to succeed if they’re viewed as a burden.

Remember, though, that behavior change across an organization can take time. Set realistic goals. And remember that data has to be centrally governed and easily located before you start assigning data quality certifications on top of it. 

The good news is, the more people trust in a process, the faster their habits and behaviors will change. What you’re trying to build is a culture where data is transparent and trusted. The right habits can form a strong bridge to get you there.