If you have searched online for "what are data silos?" you'll notice a fair amount of articles about how bad data silos are. Some titles say something to the effect, "What is a data silo, and why is it bad for your organization?" with the assumption that data silos are categorically bad. Perhaps, in learning what data silos really are and how they are formed, we will find out that data silos themselves aren't necessarily bad but instead could be an indicator of more complex organizational challenges.
What is a data silo?
A data silo, also known as information silo, is an insular management system in which one or more information systems, or subsystems, are incapable of operating with other systems or subsystems that are conceptually or organizationally related. For example, a company may have a client management system and a web analytics system with two different vendors. Both are complementary data about clients, that is, client relationship information can be linked to client web behavior producing strategic insights. If the two systems cannot communicate operationally, then the two data sets cannot be linked, and they each would be considered their own data silo, effectively they are invisible to each other.
Silo, as a term, is not new to the business world. It was applied to organizations before data and describes an organizational silo as a separation of people, like departments or working groups, which erect barriers or are incapable of inter-departmental communications, ultimately preventing information sharing.
How are data silos erected?
Silos can pop up like mushrooms, suddenly overnight. Companies that do not remain vigilant and actively integrate information sharing amongst their working groups will inevitably see silos in their ranks. Perhaps more damagingly, people will form cliques, inter-office politics will ensue, and information hoarding and protection becomes the company's cultural norm.
But sometimes, silos are also formed on purpose. Big organizations have many large departments, and despite the interrelatedness of each department's information to the whole company, opening up all department information to each other equates to information overload. Departmental separations do create discrete information domains, but management must maintain the inter-communication protocols to facilitate the right information sharing, otherwise risk isolating teams or causing information traffic jams.
In a way, this is like seeing departments as separate rooms inside the company. Each room with its own open door allows the passage of information. The doors are small compared to the rooms, symbolizing that there is more information inside than passing through. When all the doors are open, conceptually, the company has no silos, just a fluid exchange of information.
In this scenario, because the protocols have been formed, the company is alerted when doors close. Maybe the sales numbers are not being analyzed along with the marketing figures for the monthly report, and the CFO can see this. A red flag appears, the issue is addressed, and the door opens back up.
In other disorganized organizations without standardized communications, leadership may need to rely on intuition to uncover silos. They may have to sniff out signs of poor information sharing or data hoarding through a lot of face time with their team, as well as asking many questions to get a full view of the company and where isolation is forming. Maybe this organic and casual style works for particular teams. However, standardizing information sharing, even a small degree, will go miles for these companies in preventing unintended silos from occuring.
In his book, Silos, Politics and Turf Wars: A Leadership Fable About Destroying the Barriers That Turn Colleagues Into Competitors, New York Times best-selling author and management expert Patrick Lencioni has this to say about silos:
Silos are nothing more than the barriers that exist between departments within an organization, causing people who are supposed to be on the same team to work against one another. And whether we call this phenomenon departmental politics, divisional rivalry, or turf warfare, it is one of the most frustrating aspects of life in any sizable organization.
Now, sometimes silos do indeed come about because leaders at the top of an organization have interpersonal problems with one another. But my experience suggests that this is often not the case. In most situations, silos rise up not because of what executives are doing purposefully but rather because of what they are failing to do: provide themselves and their employees with a compelling context for working together.
This notion of context is critical. Without it, employees at all levels—especially executives—easily get lost, moving in different directions, often at cross-purposes.
It is a valid question to ask whether an identified silo is good or bad. However, asking questions about how the silo aligns with the company's goal is a more useful direction to enquire in.
- Does the silo serve a purpose? How?
- Is the silo a necessity for security, data protection, etc.? How?
- Is the silo causing inefficiencies or worse ineffectiveness in its purpose? How?
- Is the silo detrimental to other goals? How?
- Is the silo negatively affecting the company culture? How?
- Is it possible to integrate the silo? How?
The point is to inquire about the usefulness of silos, rather than assuming that because they do have seemingly negative effects that they are bad.
It is important to understand that data silos are extensions of organizational silos. Though data silos are tangibly separate and within the computer domain, the information within those silos are typically owned by someone or some team or department. If those departments are making their own IT decisions (even if there is an IT department, shadow IT plays a significant role in propagating data silos), they could be creating numerous silos and isolating data. Take a case of a marketing department that has deployed several vendor digital marketing platforms; it has already introduced that many silos through operating with many disparate vendors.
When addressing organizational silos, or the data silos that form out of them, each case is unique. But, by even just a cursory review of how data silos are erected, what are the challenges they present, how to break them down, and how to use them, teams and companies are better equipped to manage data silo issues.
Data silo challenges
Prominent as grain silos towering above croplands are the real risks of silo challenges that exist in the spaces unseen. Now, reverse this towering silo image and replace it with an equally silo-shaped hole in the ground. Then, you might have a more accurate representation of the hidden nature of these challenges. Gazing over the horizon, the hazards are invisible, making it impossible to determine their depth of risk.
How deep of a risk? Gartner puts this into perspective, reporting that "more than 87 percent of organizations are classified as having low business intelligence (BI) and analytics maturity." In other words, businesses that neglect tapping into their data likely have their data siloed across a number of platforms, systems, and possibly databases and spreadsheet files. These low BI maturity organizations have certain characteristics which inhibit modernizing BI and create challenges when it comes to scaling:
- primitive or aging IT infrastructure
- limited collaboration between IT and business users
- data rarely linked to a clearly improved business outcome
- BI functionality mainly based on reporting rather than analysis
- bottlenecks caused by the central IT team handling content authoring and data model preparation
Companies with these characteristics have tended to neglect the importance of data and how they keep their data. Consider the following challenges that occur when data silos go unmanaged:
- Limited Company Data View —Data silos renders a complete view of a company's data challenging, silos tend to present only a fragmented picture of business activity. The risk is in missing valuable business insights hidden within the data. And, just as limiting, keeping simplistic data sets, like storing data in spreadsheets, compared to using analytics-enabled databases, like Google Analytics, risks stopping short of uncovering deeper insights within the raw data. (Related article: Why It's Time To Ditch Excel Spreadsheets For Business Apps)
- Weakened Data Integrity—Data silos inherently create data fragmentation. And, fragmentation leads to potential data integrity decay. Fragmentation must be managed; otherwise, data may be duplicated, worse even, duplicated data may never be updated and become stale. These pockets of data weakness can cause analysis problems later, even costly damages. For example, the Canadian power company, TransAlta, while using spreadsheets to store, analyze, and move their data, made a simple cut and paste mistake that cost them $24 million.
- Duplicated Resources—Data has a financial cost, too. Storing data incurs infrastructure costs. Moving data incurs transportation costs. Collecting and using data costs. Data silos, unwanted or planned, will consume resources, and factors like data redundancy, maintenance, and duplication will demand more resources.
- Hindered Collaborative Work—Data silos emerge out of the organizational silos formed from organizational separations. As these layers of separation build on top of each other, creating both cultural boundaries and technical incompatibilities, collaborative work becomes more difficult. In today's world, data silos as a technical issue for most companies is a solvable problem, there are many solutions on the market to connect company data, and many consultants willing to propose custom solutions. Signs such as data hoarding, data turf wars, and data negligence are indicative of something beyond technical issues. Look at how the organization dynamics are working, and address the collaboration problems there.
Data Silos have challenges, but jumping to the conclusion that data silos are inherently bad doesn't seem fair. For instance, scenarios that require special security configurations may turn a "bad" data silo into an "intended" data silo. However, concerns are also warranted about how data silos impact business. They seemingly have a greater potential for negative impacts than positive ones.
Working with data silos
Let's cut to the chase; ultimately, fixing data silos starts with fixing organizational silos. Organizational silos supply the energy behind the formation of data silos. Without understanding the underlying cause that creates data silos (namely the motivations within team members), chasing a technical solution may prove short-lived. And this actually has a name, silo mentality which is the "mindset present when certain departments or sectors do not wish to share information with others in the same company" says Brent Gleeson, in his Forbes article The Silo Mentality: How To Break Down The Barriers where he gives us five holistic suggestions to breaking down data silos.
- Create a Unified Vision—At the highest levels of leadership, every member must understand and believe in a unified vision for the organization that believes in an open and collaborative environment, and one that this data-centered. Core buy-in and understanding trickles down through the organization, encouraging trust, and reinforcing the organization and departmental mentalities.
- Work Towards Achieving a Common Goal—Leadership teams must then execute based on goals and objectives that practically support the higher company unified vision. They need to search out conflicting goals that are causing silos. They must identify every sore point in the culture, perceptions, inter-personal relationships, being honest with themselves to address shunned and neglected problems (elephants in the room that are costing the company simply by existing). They must also aim their culture in the direction of data literacy.
- Motivate and Incentivize—Each team member needs to understand how their contribution relates to the whole, and how that will relate to them. How does the unified vision enhance their job performance? How will this incentivize their work efforts? However, because each team member has unique motivating requirements, team managers capable of forming and motivating cohesive teams are essential. They are crucial in translating higher-level vision into directions for team members.
- Execute and Measure—If it can't be measured, it can't be improved, or so the adage goes. Routine and constantly reinforced data gathering on objectives and goals will create a protocol on which teams can rely.
- Collaborate and Create—The steps above are a significant step towards managing organizational and data silos. However, this is not a onetime fix. Teams must continue to collaborate, create, and seek new knowledge and understanding about their current situations. Data silos are like mushrooms, springing up when organizational silos form.
Only after teams are aligned in their shared vision can a company pursue a technical solution to their data silos. Most companies may benefit from single-platform data solutions or tools that integrate data between different platforms. However, there are sophisticated tools that provide a holistic all-in-one solution that can cater to the needs of companies without extensive IT departments such as Kintone's data, app development, and workflow automation solutions.
Data silos are often written about with a data-centric focus. But these discussions tend to miss the point that there is that a human factor driving the creation of data silos. Sometimes smaller data silos will just sprout up throughout the normal working day. For example, a team member uses multiple spreadsheets to record their client data on their desktop, or the department subscribes to several vendor platforms where their data is stored but lacks integration. But in large-scale cases, systemic data silos are often created by leaders in an organization. Silos become a tool of power for politic or inter-office dominance, with the business suffering as a result. In such cases, the solution is to look at the organizational makeup and dynamics to understand what specifically is the cause, and then to address it. Only then will a true technical data silo solution stand a chance of helping the company.