Specifically, 47% of those surveyed indicated that data silos are hurting their digital customer experience initiatives.

It’s the latest example of how data silos are perceived as a huge problem — a perception that has been talked about for years.

So why are there still so many data silos? Here are five of the top reasons.

Painting a “big picture” can be really difficult in practice

Organizations may buy into the theoretical benefits of storing all of their data in one place, but actually determining what that looks like presents significant challenges.

One of the biggest: when establishing requirements for a unified data platform, various departments within an organization might not be able to clearly articulate their needs. Even more problematic is the fact that they often can’t anticipate future needs, raising the concern that a unified data platform might one day prove lacking as the organization and the markets it’s in evolve.

Budgeting can be tricky

For many organizations, establishing budget for technology used across the organization can often be difficult. Even in organizations that have IT departments tasked with supporting other departments, their budgets are often focused on technologies such as email, networking and cybersecurity. They may not have budget available for technologies such as unified data platforms, which might appear more appropriately purchased using budget from sales and marketing departments.

Lacking a department with a strong mandate to address the data silo issue, and the budget to do so, departments are effectively forced to invest in ways that create data silos.

Departments naturally focus on their own needs

Even if departments within organizations understand the problems presented by data silos and recognize the benefits of unified data platforms, the reality is that the people within those departments have jobs to do. If a data-related technology is needed to move a particular project forward, the data silo problem is realistically not going to be the primary consideration, especially if it presents a roadblock.

Put simply, the best tool for the job, as evaluated by individual departments, is often one that perpetuates the data silo problem.

Building a unified data platform can be costly

While there are numerous vendors offering unified data solutions, the costs of buying one can cause sticker shock. Even if the unified data solution itself seems reasonably-priced, the costs of integrating such a solution it into the organization, including into legacy systems, can be a wide-ranging and hard-to-cost-estimate undertaking.

This is another reflection of how the practical can come into conflict with the theoretical. In theory, unified data platforms should ultimately reduce costs, but stomaching high up-front costs is generally not easy and can be doubly hard given the other challenges above.

Customer Data Platforms Best Practice Guide

It’s seemingly easier to be nimble

Because of the vast number of SaaS solutions, many offering service at low cost and with minimal commitment, teams within organizations often feel that they can move more quickly by using disparate solutions.

Even when a unified data platform exists, if integration always requires IT support or is complex or time-consuming, teams have little incentive to favor it over SaaS solutions that can often be set up and put into use very quickly.

So what, if anything, can be done?

The challenges above are not insurmountable. Transformation is possible. But the challenges are daunting. And in some cases, it’s debatable as to whether or not organizations are wise to try to address them in the name of tearing down data silos.

While many of the arguments that data silos hold organizations back seem sensible, there is also an argument to be made that the things that contribute to data silos, such as low-upfront cost, low-commitment SaaS solutions, also offer significant benefits to organizations.

When coupled with the risk of failure when implementing a large unified data platform and integrating it into an organization, perhaps a contrarian view that the data silo problem is not as harmful as many believe is worth considering.