There’s no I in Governance 🙃
As everyone knows, we’re within the golden period of the fashionable information stack. It’s by no means been so quick and simple to plug and play with extremely versatile and scalable information instruments. We’re agile! We’re data-driven! We’re residing the self-serve, democratized-data dream!
However with each upside, there’s a draw back: the better it’s to provide and devour information throughout an ever-growing set of programs, the quicker our beloved trendy information stack turns into a tangled internet of complicated interdependencies. We develop into much less agile and, as an alternative, spend increasingly time tending to an ever-growing backlog of damaged pipelines, transformation logic, dashboards, and extra. Self-serve assets start producing conflicting outcomes, irritating our core stakeholders and planting seeds of doubt in our capacity to make data-driven selections.
Inevitably, this turns into conversations round information possession, documentation, tagging, classifying, and extra; information governance turns into the silver bullet that may repair all of it! However the place can we even begin? Who’s accountable, and for what?
Sounds daunting, proper? It doesn’t need to be! Let’s work via an iterative framework to introduce information governance inside your group.
Initially, take the time to obviously outline what downside(s) you’re in search of to unravel that could be addressed by information governance practices. For instance:
- Adherence to evolving compliance/regulatory necessities requires time-intensive and redundant guide auditing of knowledge sources. We don’t have an ordinary definition of knowledge sensitivity, so the outcomes fluctuate based mostly on who performs the audit.
- Our month-to-month information storage and processing prices are out of hand. We have to lower prices as quickly as doable, however we don’t know who owns what or the way it’s leveraged throughout the group.
- Our staff spends 20% of dev time fixing damaged pipelines as a result of unreliable information high quality of upstream information sources. Knowledge high quality points ought to be detected on the supply and resolved by the staff that owns the info supply.
Let’s hone in on the third downside as we stroll via the framework:
Now that you simply’ve recognized the focused downside to unravel and have an thought of find out how to handle it, it’s time to set some concrete objectives to work towards.
Take the time to assume via the issue at hand: what motion can others take to resolve the problem? What objectives are you able to set for others to work towards? If all goes as deliberate, what measurable affect will this effort have?
Let’s proceed with our unreliable information high quality instance, the place we wish information high quality points to be detected at supply and resolved by the suitable staff. This implies we’re focusing on two objectives: 1) assign information possession and a pair of) construct information high quality take a look at protection for all belongings. Because of this, we anticipate to see a lower in time spent resolving damaged information pipelines.
Straightforward peasy, proper? We all know what downside to unravel, find out how to remedy it, and find out how to measure the affect. Let’s preserve movin’ alongside.
You realize that answer we simply got here up with in Step 2? Resist the urge to roll it out to all information belongings, suddenly. That is particularly vital once you’re beginning with 1000’s (or a whole bunch of 1000’s!) of assets and delegating out to 10’s or 100’s of people.
🤦🏼♀️PSA: It’s a dangerous thought to dump 180k column names right into a Google sheet and anticipate a whole Engineering Group to agree so as to add documentation simply since you requested properly. They’ll chortle you out of the room and you’ll cringe each time you concentrate on it for years to come back. 🙃🙃
As a substitute, begin with a subset of high-impact, low-complexity information assets and highly-invested stakeholders to staff up with you — that is the place we begin making information governance a staff sport!
Let’s return to our poor information high quality instance: we set a really lofty purpose of assigning possession and information high quality assessments to 100% of knowledge entities. We’ll chorus from attempting to hit that massive purpose suddenly, and as an alternative we’ll begin with a slender subset of assets which can be:
- Owned by highly-invested stakeholders
- Queried with a excessive frequency
- Leveraged in business-critical pipelines, fashions, studies, and many others.
By maintaining a slender scope of knowledge entities, you’ll be capable of focus extra on designing repeatable, cross-team information governance workflows.
You’ve achieved the work to outline what downside to unravel, the focused outcomes, and which assets to sort out; now it’s time to allow others to take motion. That is the chance to work intently together with your preliminary stakeholder group to test-drive the initiative to tell a larger-scale roll-out sooner or later.
Listed below are some questions to bear in mind as you’re enabling others to take motion:
- Is it apparent? “The Why” ought to be apparent to them; it ought to be clear why this work is worth it and impactful
- Is it simple? “The What” ought to be clearly outlined to allow them to deal with execution
- Is it collaborative? Don’t overprescribe “The How”; deal with the mutual finish purpose and encourage them to construct a course of/answer that matches into their present workflows
Final however not least, create speedy suggestions loops and proactively solicit suggestions. The place are your stakeholders lacking context? Which actions are tough for them to take? What kind of ongoing help will they want?
Leaping again to our unreliable information high quality instance, we’ve requested our highly-invested stakeholders to formally take possession of, and add information high quality assessments to, their high 25% most-queried, buisness crucial information belongings. Throughout this step, we ought to be centered on ensuring they:
- Perceive that the shortage of knowledge high quality assessments has a big affect on downstream assets
- Perceive our expectations of knowledge house owners
- Can simply create and monitor information high quality assessments on in-scope assets
Now that you’ve got a cross-team information governance workflow up and working, it’s time to measure progress and decide what’s working and what’s not:
- Are you making progress in the direction of your said objectives?
- Are the objectives nonetheless the fitting ones?
- Are your stakeholders supportive of the initiative?
- What do you might want to change/alter earlier than rolling out to an extra set of stakeholders?
You’ve made it to the ultimate step! Now it’s time to begin over from the start, apply what you have got realized, and refine these workflows as you go.
Don’t let perfection be the enemy of fine — simply because you possibly can’t completely govern each single information entity inside your stack doesn’t imply it isn’t value governing a subset. Give attention to affect, not perfection.
Get snug with reevaluating your objectives/goal outcomes and serving to your stakeholders pivot as wanted. Hearken to their suggestions. Keep in mind you’re asking them to tackle further work, so do your greatest to make it value their time (and never as a result of we set a purpose that one time, so we might as effectively do it.)
I’d love to listen to how you’re tackling information governance in your group — let’s iterate on this framework collectively! You possibly can contact me any time through DataHub Slack — I can’t wait to listen to from you!
