Everyone seems both excited and anxious about their organization’s data: Suddenly you realize that you potentially have gold in your hands, but to you, it still doesn’t look as shiny and flawless as gold. And most of all, you’re not quite sure what would be the best way to exploit these nuggets, or what price you should be asking for them. So what kind of gold is data?
Last year, Mélanie Clerc one of our solutions consultants was invited by Alliander, one of the 3 leading Dutch energy network companies to pitch in Amsterdam at their 2nd “Data-Driven Companies” leadership conference. From this experience she shares her own takeaway on the value of data, and how this impacts any business seeking to truly leverage this up and coming asset.
Data-driven strategy: who’s concerned?
Strongly committed to turning the long-standing company into a data-driven company thanks to their recent impressive investments, Alliander gathered representatives from various data-driven companies to share lessons learned, inspire each other, discuss common IT platform pitfalls and facilitate dialogue among experts. The participants represented a wide range of industries using data to improve their businesses, ranging from gas and heat networks through hospitals, and even included a data scientist from the Dutch football team.
As a consultant working for a startup specialized in online urban simulation platforms, I help cities and businesses simulate how the city could evolve over the next 25 years, and anticipate how this will affect their business. Our platforms are powered by simulation models that crunch a lot of urban data to predict what may happen based on the decisions you could take and assumptions you make. Through my assignments at ForCity I’ve had the opportunity to meet with many different urban players who need to anticipate the future in the city such as city planners, utilities operators, transportation network operators or even large bank corporations.
Everyone seems both excited and anxious about data: suddenly you realize that you potentially have gold in your hands, but to you it still doesn’t look as shiny and flawless as gold, and most of all, you’re not quite sure what would be the best way to exploit these nuggets, or what price you should be asking for them. So what kind of gold is data? Is it a shiny yellow gold that you should showcase in a beautiful cabinet like a piece of jewelry (and never wear)? Or rather like black gold, which you use to fuel your projects? And how much infrastructure do you need to set up in order to start refining that precious ore?
Start from problems you want to solve, not from the data you have
Many organizations wake up one day realizing they’re sitting on a huge data treasure and launch data science contests to find ways to leverage them for what they are worth. But a few years down the line, how many data hackathon ideas really became a success story?
Data shall support your business strategy, not the contrary. If you want to truly create value by leveraging your data, start from the problems you need to solve, not from the data you own. Take a step back and allow yourself to think about all those daily issues, including the overwhelming or insuperable ones. What is the use case, what is the pain you want to solve?
Once you have identified your strategic pains, you may start sourcing the data that is strategic to support your decisions. Now in the 30 or so different organizations I’ve met during my assignments, there is one constant observation I can share:
There is no perfect data, nowhere… and that’s fine.
Organizations own a lot of data, but it’s not always easy to access it. You will face reluctance, as people are afraid they will be judged on the quality of their datasets: unharmonized data in different formats or mismatching scales, missing values, misallocation… People out there might call it gold, but your colleagues only see unrefined ore.
You will never find a perfect dataset, but that should not stop you from leveraging what you have to support your decision-making. Your organization has always worked with imperfect information, and that’s precisely why your staff’s expertise will always be needed and will always distinguish performant from poor decision-making.
Some data may even be simply missing: don’t be afraid to use the power of guesstimates. For example, if you don’t have the energy consumption values for buildings, you can reconstruct a fair estimation based on the building’s 3D shape, its age and use. Fill the missing data with estimation models, you will always be able to refine your simulations by replacing estimations with real values when these become available.
Start with what you have, you will create a pull
So start with what you have because it will help create a strong traction.
Becoming a data-driven company is also a matter of change in culture, and all individuals contributing to your project need to see benefits quickly. Showing what you can do with data will allay initial concerns and will often get you access to more accurate, more up-to-date datasets which you’ll be able to use to improve your first guesstimates.
Furthermore, although it might appeal as a comfortable choice in face of the resistance inside your organization, don’t invest too much in setting up the perfect infrastructure with the perfect datasets, topic by topic and bottom-up, because it could become a dangerous time-trap. Instead (or at least in parallel), consider using concrete applications to get started.
Starting with an applied case derived from your operational needs will save you time and money. In fact, it will help you infer the specifications for your future datasets with guaranteed usefulness: what are the most urgent datasets, in what format and with which update frequency should they be available? Remember, there is no and will never be any perfect data, so you’d better prioritize pragmatically.
Support the change to truly reap benefits
The final advice I would give is about change in company culture. Becoming a data-driven company is not only a matter of investing in the right tools and processes. As mentioned before, using data often ignites fears, but it also opens new opportunities. Hence it comes down to fostering a creative culture in which people are open to these new opportunities. You will be able to overcome issues which, up until now, most people in your organisation had always considered as immutable constraints.
Becoming a digital company is not a brush-up over old ideas: it requires deep changes in your working culture and is best done when supported both technically and humanly. Find the right influential people in your organization or get support from experienced consultants.