Approaching the Future of Analytics with Governance

"Feed me, Seymour!" The carnivorous Audrey II begs in 'Little Shop of Horrors'. Desperate to grow and thrive, the plant in question requires a very specific diet to achieve its goals. Much like our green friend, your analytics need a special diet to grow and thrive.

With the advent of such technologies as Natural Language Processing, Artificial Intelligence, and Machine Learning it becomes more and more important for organizations to embrace the concepts of information governance before the opportunities offered by such advancement passes them by. You may be thinking, "aren't these things supposed to make my life easier, and 'crunch the numbers' for me?". To some extent, that is true but take for a moment the following example.

Imagine you are an analyst at a company that has numerous sources of revenue: services, product, investment, etc. This financial data may be sourced from multiple systems, and rolling them together is a challenge. In the "traditional" world, an Enterprise Data Warehouse may be leveraged to flatten, dimensionalize, and combine the multiple systems into a single source of the truth. But even then in that ideal world: Are the field names, relationships, calculations all consistently named, stored and retrievable? Two fields, one called "Service Revenue" and one "Investment Earned Dollars" look quite different to a Natural Language Processing system. It's possible they will not be correlated, even though they are both incoming funds.

The old mantra "Garbage in, garbage out" can apply to the most advanced platform or the simplest spreadsheet.

The time is now to look at your underlying data, and determine how to apply consistency and order to it. Using governance tools is just half of the battle, and should come as a follow on to establishing top-down support, aligning data stewardship, and agreeing on a methodology for your custom process. And that is the nut... It has to be a custom process.

Every organization has its own culture and ways of dealing with conflict. That's right, conflict. Every good governance conversation will yield a series of challenges, and hopefully civil discussions about areas of the business where there are differences of opinion. Ask 13 people to define revenue, and you're likely to get 13 different answers. That's where a good governance process can help make sense of the chaos. Establish candidates for corporate definitions and create consensus around them. Approve the candidates, and facilitate the implementation of them.

The beautiful thing about these cutting edge analytic tools is that they provide a perfect demo, on top of ideal sample data. To avoid disappointment when you point them at your own data sets, spend the time now to create consistency. The tools are great, but they can only work with what they are given. You can work with your business partners, establish professional stewardship, and develop world class processes. It is not unattainable, but it starts with a single step.

Then when your own Audrey II seeks its next meal - you will be ready.