Maybe you’ve heard this before – establishing a business intelligence (BI) strategy is hard work. It’s true. BI, big data analytics, data governance, and other buzzwords are really just pseudonyms for “hard work.” Specifically, for the hard work that it takes to collect data, manipulate it, interpret it, and leverage it to answer questions. Because, at the end of the day, when it comes to BI that’s what we’re after – answers to questions.
In this article, we’ll take you through the process of implementing BI and arriving at the answers to your questions. We’ve broken the process down into five steps: develop a plan, secure buy-in, choose a platform, ramp up gradually, and then dig into your data. But it’s not easy. Keep reading to learn about pitfalls to avoid and tips to achieve success when implementing BI.
What Are You Asking?
To get to answers, we need to start with questions. What questions are you trying to answer? In other words, what problems are you trying to solve? That’s going to depend on your business. The important thing is to know what your questions are. These questions are going to guide your data strategy and help you develop a measurement plan. Before you begin with any BI implementation, you should know what you want to get out of your efforts.
It’s crucial to keep your questions specific and measurable. A question like, “How can I drive visitors to my website?” isn’t measurable. A better question would be “Which of my marketing campaigns is most effectively driving visitors to my site?” because it’s specific and it’s measurable.
When you’ve identified your key questions, it’s time to figure out what you need to measure. For example, if you’re trying to figure out which of your campaigns is driving the most traffic to your site, then you would need to measure where traffic to your site is coming from.
When deciding what to measure, i.e. what data to collect, there are some basic factors to consider:
- Can the data be collected?
- Is the data relevant to your question?
- Can you ensure the data will be accurate?
- Will there be enough data to have significance?
If you want more information about setting up a measurement plan, see our Google Analytics whitepaper.
All Hands on Deck
At this point, if you don’t already have buy-in from leadership, it’s time to take your well-thought-out measurement plan and go get some buy-in. If you don’t have the necessary support, it’s going to be difficult to move forward. Luckily, if you’ve got a good plan, then you’ve probably got what you need to secure buy-in.
Even if you already have leadership’s backing, this is the point where many BI implementations end up stalling out. Here’s a common scenario: You’ve collected some data, maybe you’ve already got a dashboard to two and some slick-looking visualizations… So, why aren’t the answers to your questions hitting you over the head?
In addition to securing buy-in from leadership, it’s also important to consider if your company-wide culture is ready for BI implementation. Is your company undergoing a “digital transformation” and if so where is it in that process? Depending on where your company is, culture-wise, you might have more or fewer hurdles to implementing BI. For BI to work, everyone needs to be on board. It’s also essential to remember that BI is a means to an end, it’s not a silver bullet. BI can help you answer specific questions, but it can’t singlehandedly solve a business’s underlying problems.
Choosing the Right Platform
Okay, if you’ve got support from leadership, and your company culture is amenable to implementing BI, then it’s time to start thinking about BI platforms. We’ve covered this in greater detail here [link].
Two big questions to consider before you compare specific platforms are: What functionality do you need? and Who will be using the platform? Speaking very generally, platforms with greater functionality will often be more complex and harder to use, at least out-of-the-box. Ideally, you want to find the platform that provides all the functionality you need, along with the clearest, cleanest, most attractive user interface – so that everyone who needs to use the platform will actually use it.
Free demos can be a helpful way to “test drive” different platforms. But you should keep in mind that limited-time demos will only give you limited insight. You likely won’t get the whole picture until you adopt a platform full time. You might also consider leaning on an agency partner in comparing platforms. At Creative Anvil, for example, we have first-hand experience with many different BI tools, from Looker to Sisense to Tableau.
More Isn’t Always Better
One “rookie mistake” we’ve seen companies make when they’re starting out with BI or analytics is trying to jam-pack way too many metrics onto a single dashboard. We get it. It’s tempting to pad your dashboards with vanity metrics. They make your dashboards look sophisticated and complex. But more often than not, simpler is better. Those vanity metrics take up space and distract from the core metrics that are going to help you. For data to be worth reporting, it’s got to be relevant.
Something else we’ve seen more than one client get hung up on is fixating too much on “big data.” It’s a buzzword for a reason. All else being equal, more data is usually better. But, with that said, don’t forget about the little data. The little data matters too. (Big or little, what’s most important is the quality of your data. Is it accurate?) Usually, you should collect as much data as you can. But for certain metrics, you just won’t be able to collect very much. Using this “little data” might involve more creativity on the part of you or your analysts, but it can be useful. Don’t ignore the little data.
Like with other aspects of a digital transformation, when it comes to implementing BI or starting out with analytics, it’s best to take a “crawl-walk-run” approach. Don’t try to do everything all at once. When it comes to BI, this could mean limiting yourself to a few critically important questions regarding your business and answering those first before moving on to other questions. For example, it might be more important to ask yourself what’s the best way to get visitors to your website, before you drill down into analyzing how best to get visitors to complete a particular action on your website.
Make Your Data Talk
You’ve developed a data strategy. You’ve secured buy-in from leadership and team members. You’ve decided on a BI platform, or several, to make use of. You’re not trying to do too much, too quickly. You’ve identified some critical metrics and you’re watching them like a hawk.
Congratulations! At this point, you’ve probably already answered some of your key questions. The good news is that you’re on the right track, and you’ve got more wins in your future. The “bad news” is that the hard work isn’t over yet. The truth is that when it comes to BI, or business in general, the hard work is never really over.
But let’s end with some more good news. If you’ve collected enough relevant accurate data, some of your questions are going to be easy to answer, but others won’t. The good news? Those tough questions are also usually the most fun.
Here’s a quick example from our Decision Science team. Recently, one of our clients became concerned that their Amazon marketing efforts were cannibalizing sales from their own e-commerce website. Our Decision Science team took that hypothesis and tested it using sales data from the two systems (the client’s site and Amazon). We weren’t able to find any significant correlation, but we did notice some odd trends in the sales of individual products. This is where we had to really dig in. We broke the data down and did three separate tests on the top-selling products. By doing so, we identified one specific product that was selling more on Amazon at the expense of sales on the client’s website. We shared this data with our client and our marketing strategy team so that they could tweak their campaigns to address the concern.
Tough questions can’t always be answered after the first attempt. They require you to do more than look at the data, but to really analyze, understand, and interpret. They require you to dig in – make your data talk. That’s the hard part, but it’s usually the fun part, too.