kentoh - Fotolia
Experts and surveys agree: Data silos are a big issue for enterprises. Unfortunately, breaking down data silos is easier said than done.
In an Adobe survey of 1,000 IT decision-makers published in March 2019, 37% of respondents cited data silos as the top challenge facing IT teams in creating a single view of the customer -- right behind integrating legacy systems.
"It's one of the largest unsolved problems in IT, and it's getting worse," said Kendall Clark, CTO and a founder of Stardog Union, a Virginia-based enterprise data unification startup. "[Data silos] are all old and enduring ... you have to just focus on unifying their contents and not shutting them down enough to activate all the political dimension of the organization."
A major issue that arises when data is stored in different silos is that the data structures are often in incompatible formats. To tackle that integration problem, companies are increasingly turning to cloud platforms -- and realizing it's a people problem as much as it is a technology one. Here's how organizations like NASA, Dell, AllianceBernstein and Goop are breaking down data silos and minimizing their negative effects.
NASA's solution to data silos
Government agencies are no stranger to dealing with data silos. NASA, Stardog's biggest client, spent years struggling to analyze the relationship among its many tests, faults, experiments and designs. Embracing a unified view of all its underlying data, with help from Stardog, changed that.
"The answers come back instantly, as if you were querying an ordinary database," Clark said. He added that Stardog's solution to NASA's problem was not in moving siloed data -- a move that could spark budgetary turf battles -- but in virtualizing all the data.
In doing this, Stardog faced "disaster areas" for multiple systems, including incompatible formats. Some data was owned by various governments, and some of it was owned by contractors, Clark said. By keeping all the data silos in place, NASA had avoided the knife fight, Clark added.
Even then, it's not easy convincing others that change is necessary.
"What can you do to remind people that there's an issue … if they've gotten used to seeing something the same way?" said Andrew Schain, a former cross-program data integration manager for manned space flights at NASA who worked closely with Clark.
One issue is that forcing a new data silos strategy onto those within the organization inevitably disrupts business practices, Schain said. If programmers don't like the new system, they won't embrace it. "Where it's emotional or technical, you introduce loss into the transfer of that data," he said.
Solving the people problem at Dell
Sometimes, a company doesn't have to go outside to address the issue of breaking down data silos.
Dell had a wholly internal challenge to overcome. Immediately after the big merger in 2016, Dell needed to bring together three disparate data systems: their own and those of EMC and VMware. How would they manage the data and integrate the networks and legacy systems so the people who worked for Dell, EMC and VMware were all able to interact?
Marc Stein, senior vice president of Dell, said that, right after the merger, the most immediate problem was coordinating calendaring and telephone systems so employees of the new firm could reach one another.
This looked like a data issue, but it actually turned out to be a people issue that was solved by improved collaboration, Stein said.
There's a natural human tendency in times of uncertainty to hole up into a shell. "In a world in which, as they say, 'Knowledge is power,' people have an incentive to control that knowledge and control that power," Stein said.
Employees will "not necessarily [try to control access to data] to the detriment of others but will use that data as their personal differentiation," Stein said. From the leadership level, therefore, came a request: Work collaboratively.
"We needed to get that coordinated not by phone calls and Excel spreadsheets, but by real-time tools and dashboards," Stein said.
With help from those tools and dashboards, Stein said Dell worked to create a solution that it used internally called the "early warning system." It started to democratize all of that data from calls and solution opportunities to things going on in the supply chain back to things in billing and invoicing back through things in service and support so that global account customers can get early visibility to issues popping up, Stein said.
AllianceBernstein's cloud-based approach
Once an enterprise addresses the people problem, experts said it can usually look to the cloud for its technical solution to breaking down data silos.
LeePatrick McintireSenior manager of technology infrastructure, Goop
Often, the problem is simply organizing and analyzing all the data necessary for an enterprise user to apply what is publicly available to achieve some degree of BI. That was the case with New Jersey-based investment firm AllianceBernstein.
"The amount of information and data was so large that its value was often lost," said Kevin Rosenfeld, senior vice president and head of BI at AllianceBernstein.
To address the data silo problem, Rosenfeld said it needed a SaaS platform that could serve the company in real time. That's when AllianceBernstein turned to Virginia-based BI and analytics vendor MicroStrategy. It took nine months to set up, but since AllianceBernstein started using MicroStrategy in August 2007, the company has saved $75,000 a year, Rosenfeld said.
"We have been able to embed predictive analytics within Salesforce.com and deliver personalized web and mobile applications that empower our sales, marketing and support teams with proactive, real-time updates," Rosenfeld said.
Data integration at Goop
For Goop, the seller of beauty and lifestyle products based in Santa Monica, Calif., owned by Gwyneth Paltrow, data "was all over the place" and difficult for employees to find, said LeePatrick Mcintire, Goop's senior manager of technology infrastructure. Photos that could maximize online sales were especially hard to categorize, he said. Mcintire said Goop's goal was to have its images in a location where machine learning could be attached to them.
Many enterprises, especially those with legacy software, have the same issue. In the case of Goop, Mcintire said his team was faced with the question of how to match the data of yesterday with the programmable source code needed to maximize sales and profit, as well as sharpen algorithms going into the next fiscal year. To address that, Mcintire said he looked to a platform-agnostic, cloud-based data management vendor.
If there's any advice that customers can offer when breaking down data silos, it's to plan and then plan some more.
"Do as much planning as you can to make sure that your data is structured in a way that's as accessible as your people or your product need it to be," Mcintire said.