Let’s say you asked all the right questions before getting started. You followed the best practices. Your technical team has deployed a solution that works. Reports and analytics are pouring out of the system. Now is the moment of truth. It’s time for action.

What do you do? How do you start?

One of the simplest techniques to get started is the “3 layers of so what” test. Avinash Kaushik talks about it in his book “Web Analytics an hour a day”

Ask every metric, report, or nugget of analytic insight that comes your way – “so what?” three times. Each question provides an answer that in turn typically raises another question (a “so what?” again). If at the end of the third question you don’t get a recommendation for action, you have the wrong metric/report.

Here is an example of the so what test:

• Sales are up month over month compared to last year. So what?
• This shows that our XYZ campaign may be having a positive outcome. So what? 
• Perhaps we should try a similar campaign in ABC  region where sales are down or reallocate additional funds to sustain this campaign longer. So what?

At this point, it’s time to test the XYZ campaign hypotheses and the select between one of the alternative courses of action. In other words, we have arrived at actionable next steps!

Most metrics that organizations report on fall into 3 categories. The nice to know, sounds important, or looks cool category. Such vanity metrics have limited value. Ultimately, we want to answer business questions and allow us to take action – do more of something, less of of something or perhaps do it in an improved way. Without this bias towards action, you will be leaving money on the table.

If I had an hour to solve a problem and my life depended on it, I would use the first 55 minutes determining the proper question to ask, for once I know the proper question, I could solve the problem in less than five minutes. -Albert Einstein

The right question can be a disruptive agent, cutting through years of complacency to redirect a team or a company’s focus. Sometimes I think the ideal superpower for a data analyst or data scientist is the ability to ask all the right questions.

So, what are the questions you need to ask before you embark on a data project?

In the book Profit Impact of Business Intelligence, Steve and Nancy Williams suggest asking the following 7 questions about your organization’s BI needs.

1. What business information do you need?
2. For what business analyses?
3. In support of what key business decisions?
4. That impact which core processes?
5. To deliver how much business value?
6. Via what changes to people, process and technology?
7. And what are our competitors doing with BI?

While these questions are deceptively simple – most data initiatives can’t or don’t answer them with any level of clarity. It is not vitally important to know what your competitors are doing but the remaining questions (1-6) are pertinent and should be addressed. Documenting your responses before starting a project will help you get a higher return on your investment.