Many business leaders and entrepreneurs are fully aware that they should be data-driven. A NewVantage executive survey shows that 99% of firms invest in data initiatives, and 92% say the amount of investment increases year on year. However, despite the significant investment, many firms are not data-driven, with the primary challenge being people, processes, and culture.
While many organizations collect, process, and analyze data, they don’t know how to question it or what to do with the results. There are two critical issues with that scenario. First, it can lead to people around the business not questioning the data at all, failing to make use of the insight. Second, if they do ask questions, they don’t know if they are asking the right ones to get the requisite answers.
Imagine running a marketing team and spending a chunk of your budget on events at sports centers. The data shows that one sports center is generating an exceptionally high volume of leads, so you continue to invest heavily in events at that location. However, the analytics might only be presenting part of the picture.
Although the number of leads was impressive, they were poor quality, low converting prospects, and ultimately low ROI. A failure to ask the right questions of your data directly impacts KPIs and the business bottom line.
In this article, we will provide a framework you can use to ask better data questions.
When thinking about your data strategy, there are four critical steps to understanding the right questions to ask.
It’s a good idea to start by assessing your company’s health. Agree on which KPIs are most important for your business and how they are currently performing. Examine several KPI examples and compare them with your own. Consider how you want them to develop in the future. Is it possible for you to alter the course of events? Determine where improvements can be made.
Make sure you don’t waste time on unanswerable questions. If you notice an anomaly in a single month, it can be tempting to spend hours going down the rabbit hole to satisfy your curiosity. Nine times out of ten, these questions tend to give answers that don’t matter much anyway.
In the first step, you will identify a business problem that requires a solution. Your hypothesis should represent a strategic objective that meets that problem with relevant measures. For example, perhaps you discover your website has a high bounce rate, and you ask, “why is the bounce rate so high on this page?” The hypotheses can reflect potential causes such as page load times, a lack of call to action, fonts, content quality, or poor navigation. Tools like Google Analytics can report bounce rate trends over time, validating or nullifying the hypothesis.
Taking the bounce rate example, you need to define what success looks like. How do you know if the bounce rate is high? Against what are we comparing? What would be an acceptable bounce rate for that type of page?
Often time, you will find out that you were not comparing against the right thing or that the metric that you were looking at is not an outlier in the end. Defining what success looks like is a great way to avoid spending time on questions that had the wrong assumption from the start.
After working with many entrepreneurs, this one step is the one that is most often ignored. It’s easy to get greedy and run many analysis thinking that you will uncover a miracle way to make more money. Unfortunately, the insight you get from your data is just as good as the quality of the question you’re asking.
One example that comes to mind that we had to deal with recently. A client wanted to know how many affiliate links a user clicks on average per session. Answering that question would cost him around $500 in development.
So even though his question is well defined (step 1 – what do you want to find out), and let’s say we also know what the industry average is (step 3 – what does success look like). Even with that, we asked him what he would do with that information and he had no clear answer, it simply wasn’t worth answering. So we dropped it and moved on to more actionable questions.
Analysts cannot produce a magic bullet and it’s imperative you take adequate time to plan your questions. Random questions will get random answers.
Once you know what questions you want to ask, your hypotheses, measurements, and actions, you can think about what you will need to come up with the answers. Below is an overview of the six pillars to successful questions. Of course, there are more technical aspects to each of these that should form part of your market research and data strategy.
The question could be focusing on sales and revenue, website visitors, mobile app downloads, or anything relevant to your business.
You may have local or cloud-based data warehousing, Google Analytics, CRM, CMS, and countless other sources available. According to one survey, the average business has 400 data sources.
The frequency likely depends on the metrics and their relevance over different periods.
Linked to the frequency of data refreshes. Some metrics won’t change each day drastically, and weekly or monthly refreshes and insights could suffice.
For your hypothesis to be valid, you should compare the most relevant periods. For example, if your business experiences seasonal trends, month-over-month comparisons might show understated or overstated results.
This includes a clear objective, access to quality data, and the right technology, and support.
Asking the right questions can save you and your analytics team time while providing valuable insight. Having worked on countless data analysis projects, Systematik has the experience to help you gain data insights in the most efficient way possible. Contact the team today for your free 30-minute strategy session call with absolutely no obligation.