Last month, we discussed the different forms of data analysis and their applicability across different organizations and different uses cases.
Understanding the landscape of available data analytics methods and determining which tools are appropriate for various projects and use cases is critical to succeeding with your analytics initiatives.
Define the Questions and Break Down the Silos
For most organizations, however, determining the right analytics method shouldn’t be the first step on the road to analytic success. In fact, it shouldn’t be step two either. Remember, the first step in any successful analytics initiative is always to define the specific business question you want to answer.
That’s because even the right analytic method will ultimately fail to produce a successful outcome unless you first break down the structural information and communication silos that still exist in most organizations.
Don’t Blame the Analytics, Blame the Silos
In my experience with customers, when an analytics initiative fails, it’s not because of improper selection or application of an analytic method. It’s not even because the organization didn’t have the right tech tools in place to implement the selected method, no matter what vendors and marketers would like you to believe.
It’s because the organization had the wrong structure and the wrong culture. An organization with a culture and structure marked by information and communication silos is one that simply cannot succeed with analytics. It’s the analytic world’s version of ‘set up to fail.’
The Growth of Stealth Silos
So, let’s take a step back and forget about specific analytics for a moment. Let’s look instead at the types of information and communication silos that exist in most organizations, understand how they can stand in the way of analytics success and then explore some ways to break them down.
One of the many reasons silos are so common and their existence so hard to prevent is that they develop organically in parallel with the natural growth and progression of a business and often go unnoticed until they are firmly established.
When Good Intentions Turn to Silos
Think of a business as being like a beautiful park. At first, the landscape is simple, pristine and undisturbed. Then, development starts innocently enough with a hiking trail or two. Pretty soon, a small visitor’s center crops up to disseminate information.
Before you even notice it, there are three restaurants, two hotels and a shopping center, all within walking distance. Every addition was well-intentioned and harmless by itself, but the cumulative effect was to change an undisturbed natural landscape into a bustling tourist destination.
Now, visitors might have to talk to three different guides to what they thought were simple questions about the park.
Silos, Silos Everywhere
Your business develops in much the same way, with new team members, new departments, new processes and new tools slowly added over time. Each step in the growth and development of your business creates information and communication silos, which become entrenched parts of your company’s culture if left unattended.
The more complex your organization and the more diverse and autonomous the collection of businesses within it, the more silos are likely to proliferate.
Are Your Silos Self-Sustaining?
With respect to analytics, the silos that tend to exist in most organizations can generally be categorized along the lines of ingestion, processing and understanding.
It’s not uncommon for an organization to have one team responsible for collecting data for analytics, another team responsible for processing the analytics and a third responsible for making business decisions based on the outcome of the analytics.
Each of these teams has its own leadership, its own KPIs and its own definition of what is and isn’t a successful outcome. Just as important, each of these teams is doing the right thing when viewed through the lens of its particular silo.
How Silos Create Roadblocks to Success
You can see how silos can quickly become roadblocks to success. Without cross-functional communication and sharing of information, you may do a great job ingesting and collecting data, but it might not be precisely the data needed for the type of analytic process your team wants to execute.
Or your data scientists might flawlessly execute a given analytic process, but they did so without really understanding the business decision making questions.
Or the business decision makers might jump to a conclusion without really understanding what data was used or how to interpret the analytics.
When Data Lakes Become Silos
It’s not just structural silos either. Technology silos are equally common and often equally damaging. Take the example of a data lake. Many businesses started down the path of creating data lakes because they thought they would serve as simple, cost-effective ways to store data and centralize data collection activities.
In the process of building the data lake however, the finance department decided it still needed to maintain its data warehouse, the HR team utilized a few data marts for some one-off projects and the marketing team invested in a few SaaS solutions. In the end, the data lake simply became yet another structure to manage, maintain and optimize. In other words, another silo.
Pursue Transparency, Not Perfection
So, what’s an organization to do? For starters, accept the fact that it’s unlikely you can completely eliminate silos. Remembers, analytics is ultimately about making the best possible use of information to make the best possible decision.
Nothing’s ever guaranteed to be perfect, and you’re just trying to make the best possible decision based on the information at your disposal. The same is true for any effort to break down analytic silos. It’s never going to be perfect. Your aim should be to create as much transparency and collaboration as possible based on the unique circumstances of your business.
After all, breaking down silos isn’t a zero-sum game.
Silos Never Happen Deliberately
Once you’ve taken on the right mindset, the next step is to understand the topography of your organization. You can’t break down silos unless you can pinpoint where they exist, and how and why they came to exist.
Understand your data governance process and your security processes, as well as the tools your organization has invested in. The reality is that most teams aren’t deliberately operating in silos but things just naturally progressed to that point.
Being able to point out where and why a silo exists is the first step to breaking it down.
Pair Analytics With Project Managers
A smart next step would be to make sure that every analytics project, has a project manager who understands the overarching purpose behind the initiative and is closely involved to provide oversight and ensure consistency at every stage.
You’re looking for someone who can say to the team collecting data that based on the business questions you’re trying to solve, particular data is or isn’t relevant. You want someone who can clearly articulate to the analytics experts what data was collected so they can make the right choice in terms of analytics processing. And you need someone who can help business decision makers understand what it is they’re seeing when they look at an analytic.
And it doesn’t have to be just one person either. The more people who are involved with an analytics project in holistic, cross-functional ways, the more knowledge you’re creating for that and future projects.
Crawl Before You Walk
With C-level executives in all industries eager to leverage analytics to drive their businesses forward, it’s only natural that organizations are eager to learn about and invest in analytics technologies.
Ultimately though, companies that make the leap before they’ve assessed the landscape of their organization, identified silos and taken steps to break them down, will see their analytics initiatives fall short of expectation.
However, those who exercise patience and follow the critical steps of understanding their topography and aligning their teams properly before leaping into a series of technology investments are the ones who will succeed. In analytics, as in life, you must crawl before you walk if you ever hope to run.