This article originally appeared on Forbes.com.

Talk to CEOs about data analytics and they will tell you it’s a business challenge, not a technology issue.

They can tell you (or you can read for yourself in the management press) that gleaning insights and creating value from data isn’t something that can be handed over to IT. It requires new hybrid teams and roles that understand the business, as well as how and where to find the data that creates competitive advantage.

We agree; indeed, we tell this often to clients.

However, it’s also true that companies can’t reap the full benefits of their analytics efforts unless their technology organization is up to the challenge of managing the data that makes it possible.

Unfortunately, most CIOs and other senior IT executives believe their companies are not prepared to make full use of the waves of data coming into their organization. A Bain & Company survey last year of 325 global companies found that 59% believe they lack the capabilities to generate meaningful business insights from their data. In another Bain survey of 250 companies, 85% said they will require substantial investments to update their existing data platform, which includes consolidating and cleaning data, simplifying access and rights management, and improving access to external data sources.

Given that these technology executives already understand how far they have to go, what keeps them from getting started? In our experience working with hundreds of companies around the world, we see three major obstacles.

The divide between ownership and stewardship of data: IT may not always know where the value resides in data, while executives on the business side may not understand the intricacies of data storage and management. This disconnect can have expensive consequences if the business (which owns the data) and IT (the data stewards) make decisions without a solid understanding of each other’s perspective.

Getting data into the hands of the right people: Historically, IT departments focused on how and where to store data and how to keep it secure. But companies don’t create value by storing data or managing access rights. Value comes from putting data in the hands of business people, and this is where a lot of companies stumble.

Old processes that no longer fit business needs: Data warehouses used to be the centerpiece of the IT organization, walled gardens that protected a company’s most valuable data and restricted access to only a few. That model fails in a world where companies need to allow more people to access the data and make discoveries. For these uses, cloud-based infrastructure can work better than on-premise systems because they can be scaled up, provisioned on-demand and paid for on a consumption basis.

These are common barriers, but most companies can learn to become more like analytics leaders by creating tighter working relationships between IT and the business, and updating policies that may be limiting the value they get out of data.

One imperative is to bring the business closer to the development process. Agile development methods, which build applications more incrementally with frequent input and review from the business, are an essential step in that direction. So too is the concept of DataOps: Just as DevOps improves efficiency by increasing collaboration between application development and operations, a DataOps mentality focuses the organization on improving the way that business leaders, data scientists and IT managers work together to discover insights.

This close collaboration should help the technology team continue to prioritize its investments where they can deliver the most value. Keeping the focus on long-term business priorities will help executives look beyond current fads and buzzwords and toward building sustainable analytics capabilities.

But the technologists have an important role to play, too.

The risk in prioritizing by business needs is that it can make it hard to invest in underlying issues that may have no immediate upside but that can hurt productivity in the long run if left unaddressed. One common example is companies that store customer data across several different systems, often as a result of mergers and acquisitions. Business executives sometimes find it hard to see the value in rationalizing these systems, but the technology group needs to explain how doing so will save money over time and enable new digital capabilities such as the ability to serve customers with a common knowledge, whatever channel they’re coming in through.

Plenty of CIOs will fail to push their IT departments beyond traditional data management into a realm where their technology makes it easy for executives to access data and derive valuable business insights. Successful CIOs, however, will be those who move from a technologist’s view to one that allows a broader swath of business users to access data, experiment with analytics and derive new insights that deliver meaningful value to the organization.

Steve Berez is a partner with Bain & Company in Boston.

Rasmus Wegener is a Bain partner in Atlanta and the founder of Bain’s Advanced Analytics practice.

Paul Callahan is a principal in Atlanta, where he works with Bain’s Global IT practice.