Cătălin Pop is the lead Project Manager at Tecknoworks. He has extensive experience spearheading client initiatives in Data Analysis, Business Intelligence, Office Automation, Machine Learning, and AI.
Implementing a data analytics program is crucial for competitive edge in today’s business environment – but nobody said it was easy. Despite tremendous growth in the application of BI tools over the last 20 years, it’s brand-new territory for most organizations.
According to Gartner, 80% of BI projects do not deliver the planned business outcomes. But you absolutely don’t have to end up in that unfortunate statistic.
If you’re at the beginning of your BI initiative, or are considering taking the plunge, here are a few of the most common errors to avoid.
1. Regarding Analytics as a Technology Project
Of course, data analysis is achieved through technology, but the algorithms all by themselves are not going to get you to your objectives. Think of your data project as a business initiative, not simply a task for your IT team.
For example, let’s say you’re interested in predicting churn. Identifying the right customer demographics or geographic locations to start with would involve stakeholders and team members across departments.
Your IT team is naturally going to do the technological heavy lifting. But the project has a much greater chance of achieving your goals if you bring more people to the table. View the data project as just one element of a much larger business plan, identify a use case that propels your overall strategy, and put together a strong interdisciplinary team.
2. Making it Hard to Adopt BI Across the Organization
There’s nothing wrong with starting small when it comes to your data project; in fact, it’s almost always the best approach. But staying small because you can’t get users on board is a problem.
To maximize the success of your data initiative, use a data tool with functionality appropriate for every user’s need and skill level. It should be just as easy for the sales team to access and use the tool as it is for the CTO.
At all costs, avoid implementing different tools for different users and requirements. Doing so leads to the problem described below.
3. Creating Data Silos
Data analytics can and should be self-service. When your entire organization is able to leverage your BI tool and make data-driven decisions, everybody wins.
But all too often, siloed teams prepare their own data and create their own reports, shared only with each other. When several teams are doing this in their own silos, it’s called “multiple versions of the truth.” And it’s not hard to see why that’s a problem.
The issue is magnified if several data tools are in use. Everyone’s sourcing from their chosen data pools, preparing data their way, inadvertently letting errors slip in, and using different calculations. The unfortunate result is that people make decisions based on poor or incomplete data, totally missing the point of the entire initiative.
For example, let's say the sales department is using Excel, the finance department is using Domo, and R&D is using Power BI. There's no way to ensure the data is being cleansed consistently, and no way to make truly valuable connections among insights.
Ensure that your data strategy is founded on one properly governed and well-organized platform, with data that reflects the single, real version of the truth.
4. Using Unreliable Data
Even though I’ve touched on data governance above, it’s such a crucial point that it deserves a larger explanation.
No one sets out to purposely use poor data, obviously. But when you first start a BI project, the right data to use (and what to do with it) isn’t always immediately clear.
What data are you currently collecting? How is it managed? What data is missing, and what is inaccurate? This can come down to something as simple as duplicate entries, missing form fields, and even typos – and can totally invalidate your results.
Take the big-picture view of your organization’s current status and future goals, identify the data you’re going to use (or start collecting it), and enact a strong governance method. This will get your project off to the right start, and ensure reliable results as you continue to explore analytics.
5. Overlooking Data Storytelling
Narrative is a powerful communication method, engaging both the logical and emotional parts of the brain. If you want real buy-in from your team, clients, or stakeholders, you need to tell the full story of the data.
Don’t just tell them that a product isn’t selling as expected; show them the interconnected, data-driven reasons why.
This is where your dashboard and visualizations come in especially handy. Multiple views of data sources, conclusions, and metrics tell the how and the why, not just the what.
Yes, data science is challenging, and the endless pace of technology can make things even more complicated. But BI has been around long enough to establish the best practices and strategic approaches that provide real returns.
Having this big-picture overview of common pitfalls is a good place to start. From here, you can begin considering your goals, getting your team together, and deciding which internal and external data is needed.
Ultimately, think everything through carefully, knowing that BI tools achieve results only when implemented correctly. Steering clear of these major pitfalls means you’re already ahead of the competition.