Writer, classic rock lover, dog rescuer, company founder, software exec, and now independent management consultant--I speak, blog, and pester my friends about these topics. My current focus is getting IT and business organizations to collaborate more effectively and not kill each other. I also talk and write about big data, why analytics is fundamentally strategic, how to pitch business execs on IT projects, and why not to buy a dog from a pet store.

I’ve lived in London, Paris, and Sydney, but call L.A. home. #weatherwimp. I cultivate an organic vegetable garden and friends with issues. I’ve written three books, co-authored a fourth, and contributed to a bunch more. (I have another one in my head waiting to come out, but it’s crowded in there right now.) I prefer Def Leppard to Bon Jovi, mashed potatoes to brown rice, fly fishing to golf, Pinot Noir to Zinfandel, and nice people to assholes. I have a tattoo. I’m not telling you where. I feel guilty that I go hot and cold on social media, that I don’t spend enough face time with my friends, that my French is rusty, and that I ate that whole bag of Kirkland peanut butter cups in less than a week. I have to live with those things.

The (Changing) Rules for Data Governance Success

The (Changing) Rules for Data Governance Success

In which Jill shows how her clients are modernizing their data governance playbooks.

I recently watched a YouTube video interview of Mark Sisson discussing his work in the nutrition movement. Sisson is the author of The Keto Reset Diet and an expert on what he calls "ancestral health."

During the interview, Sisson mentioned the microbiome, bemoaning that he'd been talking about it years earlier than other experts only to be queried about it recently. To him, the microbiome -- the human digestive environment full of bacteria and other organisms that can protect us against disease -- was old news, something he'd already moved past. In the health and fitness movement, however, the microbiome is still very much a thing.

I tell this story because it struck a chord. I've been talking about data -- its hygiene, integration, enrichment, provisioning, and value -- for more than two decades. (My book e-Data was published in 2000.) Nowadays data's not just a hot topic; every CEO wants his or her company to be "data-driven." Many even claim to have transformed their companies into data companies. It seems so 2008.

That's why when I received Bill's question below, I initially felt a little pang of "been there, done that."


I'm very interested in what you consider to be the 5 to 10 keys to successful data governance programs. I am at an organization that is beginning a data governance initiative in 2019 and I would like to know what is most important to implement.

-- Thanks, Bill

My first reaction was to dig out a copy of the well-worn "10 Mistakes to Avoid When Launching Your Data Governance Program," which I wrote with Kimberly Nevala for TDWI (yup, back in 2008, and downloadable here). Companies were guilty of treating data governance like a project (Mistake #4) and prematurely pitching data governance (Mistake #7), and were starting to recognize their errors. What was left to be said?

Plenty, as it turns out. We couldn't have predicted some of the changes that would occur a decade later. Artificial intelligence, the digital revolution, the Internet of Things, even the blockchain -- they all lean heavily on their ability to crunch relevant data.

Here are a few updated pointers for Bill -- and you -- to ponder as you recast your data governance efforts or launch them anew.

Rule #1: Consider your culture

Sure, we covered that back in 2008. Mistake #6 was "Overlooking cultural considerations." The difference now is that companies have more rigor in their development processes and more orthodoxies. Agile development has either taken hold or it hasn't. Shadow IT is out of the closet and lines of business own their own applications. Everyone agrees that customer data transcends a single line of business. In the early days, data governance programs sought to upend existing cultural norms. Many went too far and failed.

Rule #2: Use commercial tools

Back in 2008, data stewardship workbenches were clunky, homegrown, or both. A decade later, business people are wrangling their own data and data curation is the price of entry for any earnest analytics effort. Companies such as Collibra, Alation, Ataccama, Tamr, and others offer data management and enrichment toolsets that make it easier for laypeople to bring the right data together and get on with the business of analyzing it.

Rule #3: Don't ignore your data lake

Data lakes weren't around in 2008, but companies and their vendors are serious about them now. Hot off the presses, Amazon just announced AWS Lake Formation, a managed platform to support formal data lakes.

Data lakes have varying levels of data rigor. As with data governance a decade ago, different companies define their data lakes differently. Some have become a veritable bouillabaisse of raw, heterogeneous data with no unifying purpose. Others are scrupulously governed, managed, and secure. Neither model obviates the need for business-driven data policymaking and oversight, which is what data governance is.

Rule #4: Get small

Back in the day, data governance was automatically assumed to be a large, enterprise-scale activity where everyone sat in a room making decisions about how data would be owned, managed, and maintained. It never worked that way, of course, with some data constituents being marginalized (hello, manufacturing!) while others got an abundant share of resources and time. When I introduced the concept of data governance "regimes" in 2009, many of my consulting clients reacted as if I'd just enrolled in clown school. The joke was on them; small teams abiding by core guiding principles get so much more done.

Rule #5: Treat data as its own skill set

These days, business intelligence is a commodity, advanced analytics has broken free of COEs, and artificial intelligence and machine learning are redefining the need for data-specific skills. Being data-savvy in and of itself means nothing. Understanding the data life cycle, key data sources, enrichment functions, and usage needs is increasingly carrying a high price tag.

In short, data is the hard part. I think Bill gets this already, hence his question -- which, I'll admit, is more relevant in 2019 than it ever was.

Thanks, Bill!

Original article published on TDWI.org.

5 Questions CEOs Are Asking About AI

5 Questions CEOs Are Asking About AI

The Internal Disruption of AI

The Internal Disruption of AI