A major hurdle corporations face, across industries, is how to break down the data silos that exist within their companies, so they can visualize it, analyze it, and draw insights from it — not to mention hire the talent needed to make that happen. 

“Just getting the data lined up properly, so that people can analyze it, is typically I think 80% of the problem,” Phil Snow, CEO of FactSet, said at a Fortune CEO Series panel this week on data and customer experience.

There can be massive benefits for those who get it right. For instance, retailer Old Navy aims to use predictive analytics to better manage its supply chains, said CEO Nancy Green. “We’re very fortunate to have a very large first-party database of tens of millions of customers, and we use that data in various ways,” she said. “We see the biggest opportunity going forward is really around the power of using predictive analytics to better forecast customer supply to better meet customer demand throughout our supply chains.”

It’s not just the data itself that’s important: Employees need to be able to understand it and put it to use. Hiring data analysts can be tough, though, as there’s a shortage of them. “The tech exists,” says Mohamad Ali, CEO of tech research and media company International Data Group. “Really the hard part is having the talent in place to utilize this stuff.”

Amid the difficulties in hiring data analysts, but there’s also something to be said for training pre-existing employees across a company’s lines of business to harness data and therefore become more effective in their roles, said Mark Nelson, CEO of data visualization company Tableau. “Your whole company culture has to want to be data driven,” he said.

Some industries are further behind others when it comes to cleaning up their data: CEOs on the panel specifically mentioned how hospitals and banks tend to lag in this arena. But at the end of the day, all companies should have this on their radar, they said, particularly after the supply chain challenges brought on by COVID. Besides, once companies can get past that initial hurdle of cleaning up the data, said CEO Peter Brereton of supply chain software company Tecsys, the analysis “becomes actually fairly straightforward.”