In our new viewpoint, Getting unstuck: Accelerating results from digital transformation, we discuss some of the challenges organizations face when trying to see results from digital initiatives. One common factor that hampers innovation across industries is the less than ideal use of data. Moreover, as some experts have said, successful digital transformation requires data transformation.
I often see this first hand when working with clients, they ask me to meet with the data science team and determine what they need. They can’t understand the lack of progress in any initiative and why the team, despite their experience and skills, are not able to provide any insight into their organization.
Usually, within the first few minutes of the initial meeting with data science team, the issue is clear.
Imagine hiring a scientist specializing in molecular chemistry. They were the top of their class from the most prestigious university. You know that they are fully capable of making the latest discovery and can’t wait to see what amazing formula they create. They are excited too, as they put on their white lab coat for their first day of work. “Let me show you to the lab!”, you say as you gleefully walk them to their laboratory. “Well, here it is! I’ll check on you in a month.”
The chemist opens the door, flips on the light, and their face immediately drops. The room is a disaster. All the equipment is behind locked cabinets for which they don’t have a key. Even through there are jars of chemicals, none of the jars are labelled and it’s not clear how long they have been sitting out on the counter. The chemicals they really need are nowhere to be found, and after calling around they find out that those chemicals are on backorder and can’t be delivered for months.
A month later you walk back into the lab expectantly, ready for the latest new chemical formula that will send your sales soaring. “I have nothing” says the chemist. “Nothing?” you say. “You’re the top chemist in the country!”
Would you expect the chemist to produce anything of substance after a month? They don’t have the foundation of what they need to do their job. Even if they did manage to mix something together that worked, they couldn’t explain what chemicals they used or the formula to repeat those results.
Most organizations hire a data scientist and put them in the same situation. The data scientists don’t have access to the data they need. The data that they can access has no associated documentation and the age, relationships and relevance of the data is unclear. The data is also spread around the organization causing them to go on a scavenger hunt through different departments. If they are able to build an algorithm or machine learning model, it’s difficult to explain or repeat the outcome.
An organization’s data scientist team needs a data engineering team that can help gather and catalog the available data. They need the right tools to source, identify, and promote to production. They often also need a trusted partner – like CGI – to provide strategy and implementation for their data science initiatives such as data engineering, analysis, and data governance practice.
Done right, data strategy and implementation support can be the catalyst (chemistry pun intended) for real transformation.