By Bret Greenstein
April 7, 2021
Every business leader pledges their commitment to be more data driven these days. It’s also no secret that artificial intelligence will be key to competitiveness and customer-centricity for just about everyone, and that success depends on applying it to the right data. But it turns out that the data that is often the most valuable to companies is not where they think it is. It may not even be inside their companies.
During the pandemic, data’s centrality became decisively more apparent. Having your finger on the pulse of data turned out to be critical for adapting quickly in a rapidly and radically changing environment. But not many companies were good at that. Too many struggled because they didn’t understand what data matters.
Almost every business leader I meet is still making “gut feel” business decisions off of spreadsheets. Most companies generate thousands of reports that nobody looks at. And in any case those reports are backward-looking, historical. Too many companies just think about optimizing the data they have, instead of working to become a forward-looking company and considering the data they need.
How do you unlock the value of data? I’ve been working on artificial intelligence and data for years. But the pandemic helped me refine my own understanding about how it can be most useful. The key, I’ve realized, is knowing which data matters to your business, making it accessible to the people who need it, and making sure it is timely and well-governed.
And what too many companies miss is that what will make them competitive will not just be the data they control. A big revelation for me in the last 12 months has been that companies that were looking at their own operational data were completely missing all kinds of other relevant data–including hyperlocal information and signals from social media, as well as data about policy and other developments that affected their business. Why, for example, did we have to keep running out of toilet paper? Then flour ran out in grocery stores almost everywhere. But a company watching social media could probably have seen that everyone was baking bread at home, because they couldn’t go out and eat.If you trained your AI to look for market signals like this, maybe your organization could have seen what was coming, and been better prepared.
For example, we worked with a large home goods store. They wanted to know where people go after they leave the store. We were able to get anonymized location data based on mobile phone traffic. That’s publicly available and you can pay to get it. We overlaid that data with geospatial and map data. We used all of that to look for people who were in store A and see where they went for their next stop. It turned out it was most often a fast food chain. That gave the retailer lots of ideas for promotions and other ways to stay better in touch with their customers. We are increasingly helping our clients tie together diverse data types like that.
No one company can control all the data that will turn out to be useful for its business. You will often need to aggregate data from all sorts of other places and combine that with what you control, in order to make the best decisions. It isn’t always easy to get the data you need, but when you make it work the trouble is worth it.
There is a famous scene in the TV show Silicon Valley where the rich investor connected the cyclical patterns of cicadas emerging from their years-long nymph state with the price of sesame seeds, which apparently they ate. Then he made some quick money buying sesame seed futures. But it doesn’t take an eccentric genius to do that kind of thing anymore.
What we need is a new mindset that encourages sharing data and what it can enable. Many companies and organizations control data that could be valuable for other companies. And conversely, pretty much every company can benefit from gaining access to data controlled by others.
Information sharing between companies has evolved tremendously over the years. I’ve been watching that for my entire career. We used to build shared File Transfer Protocol (FTP) boxes, so if I put something into it you could “FTP” it out. It was crude stuff. But then B2B exchanges took sharing to the next level. Since then companies like Dropbox, Box, Onedrive and other have been used to enable collaboration between companies. Now the innovative cloud company Snowflake has made a step function improvement in what’s possible, by securely providing direct access to data.
People think of Snowflake as a warehouse in the cloud. But in fact it’s a warehouse with a door that connects to your warehouse and to my warehouse and to lots of other companies’ warehouses. It ultimately can become a superhighway of warehouses. Think of Snowflake as the world’s largest shared Dropbox. The important thing is it has control capability. Customers can share their data without replicating it. You can use my data but it doesn’t have to replicate, you just get access to it. They’re breaking down the barriers to information sharing. It’s not the main part of their business right now, but it could be. And Snowflake will not be the only such platform.
What companies are going to increasingly want is “API-callable data”. Say I want to see your delivery schedules, and you’ve decided to share them with me, but you can cut me off tomorrow and I’ll never get it again. That becomes a controllable spigot of information for the people you decide to share with. Companies will want to get access to one another’s data but not replicate it all over the planet.
The economic potential of this future of unlocked and shareable data will be huge. It could help companies make decisions in real estate and lending, for example. You could combine data about people who are moving, from a company like U Haul, with data from a furniture company like Raymour and Flanigan, plus data about property values and home sales from companies like Realtor.com or Zillow. If that were available to a bank, it could help them figure out how best to underwrite mortgages and what risks they should take. Or you might run AI against public webcam data to understand traffic patterns of customers going in and out of stores to help you estimate the revenue in these stores. How is traffic changing each day? Does it correlate to weather? Is it correlating to COVID infection rates? Is it correlating to some other macro trend? If you further overlay geospatial data from a company like S3, you could get insights into whether property values will go up or down.
Every company is sitting on a pile of meaningful data. We each uniquely know something about the world. When that knowledge is combined with insights from other companies, it can be highly valuable. Locked up data doesn’t create value. AI and analytics will become more and more important as companies gain insights from data they get from other partners as well as public sources. And corporate partners themselves may generate new income streams by making their data accessible. In the future, companies will do better if they’re better at sharing data.
Bret Greenstein leads Cognizant’s Global Data Practice, focused on helping Chief Data Officers transform their businesses through Data Modernization. Bret will be speaking at our upcoming Health+Wealth of America event, April 20-22 – learn more and register.