The following Insights post is contributed by Gary Angel, Principal at Ernst & Young and expert JPK Group Presenter.
In the halls of enterprise BI, the idea of democratizing data has the same rousing emotional appeal that the Internationale does in left-wing Paris salons. Democracy good. Self-sevice good. Data good. It’s all good. But in practice, it tends to work about as well as the Russian Revolution did in creating a worker’s paradise.
The evolution of generalized tools like Tableau and purpose built tools like Google Analytics has indeed created a class of software that brings sophisticated and vibrant reporting and visualization to the masses. If you’re counting on technology to raise the analytics bar in your enterprise, though, you’re in for some serious disappointment. The challenges in democratizing data are formidable and are not primarily technical.
Here are the hard realities upon which most attempts to democratize digital data founder:
Too much data: Most users with direct access to digital data face immediate data overload. The common lament – “I have too much data to use.” This must be maddening to advocates of democratization but it’s the simple fact of real human experience. Most users don’t know where to start when it comes to using digital data.
KPIs don’t work: Democratizers have a simple answer to the “to much data refrain” – focus on a few actionable Key Performance Indicators (KPIs). Stripping away all the options does make reporting more accessible; it just doesn’t make it any better. KPIs are neither actionable nor even directional. Every KPI in digital, from uniques to conversion rate to NPS is ambiguous in interpretation. Change can be driven by a multitude of causes that make for radically different (from positive to negative) interpretations of the data. KPI driven reports are at best useless and at worst dangerously misleading.
Context and Connection are Missing: KPIs don’t work because they don’t show the connections between data. But it’s in the connections that all the real learnings and potential actions reside. Worse, even rich slicing and dicing won’t solve the problem. Connections in data can only be substantively revealed by statistical analysis – they don’t fall out of traditional cross-tabulations. You can endlessly explore most data sets in a visualization tool and never find the real connections that drive the system.
Reporting Fatigue: People get tired of looking at the same old numbers from the past. It’s too static. Top-line numbers rarely show much variation and give no clue to what might be interesting or different in the underlying data. The reports that your consumers were so excited about six months ago often sit completely unread in most inboxes.
There is a better way.
Instead of building reports that democratize data, build tools that democratize knowledge. Have your analysts construct models of key systems. Then embed those models into generalized data visualization tools. Embedding models in reports fundamentally changes the reporting paradigm. Instead of relying on the user to find the connections between the data, the model does that. Having a built-in model in a report solves every key problem in reporting. It helps the user understand what’s actually important in the data. It functions as a built-in alerting system by highlighting only significant variables and statistically interesting changes. A good model can illuminate what happened, forecast what will happen, and, best of all, allow the user to explore what might happen. When users have tools that allow them to explore the potential impact they can have on the system, they quickly grasp how the system works and what their options for change are. That’s powerful. Far more powerful than dumping any amount of data on their doorstep. And users appreciate it; this type of analytics tool doesn’t suffer from reporting fatigue and is far, far more likely to become a part of the everyday operations of the business.
When you integrate analytics and reporting, you’ve gone beyond data democratization. You’ve done something far better. You’ve democratized knowledge. That’s real liberation.
As the Internationale might have it:
Of the past let us make a clean slate.
Enslaved masses, stand up, stand up.
The world is about to change its foundation.
Widely considered one of the leading digital measurement experts in the world, Gary leads EY’s Digital Analytics Practice. EY acquired Gary’s previous company – Semphonic – in March of 2013. As Semphonic’s President and co-Founder, Gary led Semphonic’s growth over a 15 year period from a 2-person practice to the one of the leading digital analytics practices in the United States. Voted the most Influential Industry Contributor by the Digital Analytics Association in 2012, Gary writes an influential blog, has published more than twenty whitepapers on advanced digital analytics practice and is a frequent speaker at industry events.