Many years ago at a leadership conference, I attended a breakout session about setting and achieving goals. The lecturer's advice boiled down to a two-step "backwards" process popular among successful leaders.
Step One: Imagine yourself immediately after achieving your goal.
Step Two: Imagine what you did immediately before that. Repeat this step until you are in the present.
Planning -- let alone executing -- success can seem daunting when viewed prospectively. By working backwards, however, you force yourself to retrospectively script the minutiae that will logically lead you to success. Plus, this "backwards" method is emotionally easier because it requires that you start with the fun part -- imagining your success.
That advice has stuck with me throughout my professional life -- and it applies to analytics just as well as any other area.
Customer response surveys are a great example. As AllAnalytics.com community editor Shawn Hessinger and blogger Sandra Gittlen have both pointed out in recent posts -- Survey Fatigue Could Boost Social Media Mining and How to Battle Survey Fatigue -- the American public is getting fed up with these surveys. It's easy to see why: Companies don't know why they're surveying.
Oh, sure, executives have inklings that it is generally good to know when and why customers are dissatisfied. There, however, the understanding ends. Companies don't know what they're going to do with the data once they obtain it.
Accordingly, when implementing customer survey efforts, companies cast far too wide a net, going on a fishing expedition for data -- any data. They don't know what they're looking for. The result, all too often, is a 20-minute series of awkward questions that will probably get ignored (such as I described in my post, CRM Analytics Takes More Than Analytics).
Worse, by asking too many questions, it's all the more likely that the company will ask the wrong question (as I discussed in another post, E-Chat Today: Tedious Questions & Other Potential Survey Killers).
Another example is online sentiment analysis. Companies know that mining the Web for expressions of positive and negative sentiment yields valuable information. Indeed, online sentiment analysis has been used to successfully predict important future events -- including sales receipts, stock prices, and even election results. The problem for many is that they don't know how to fully realize that value.
Many misunderstand the use of online sentiment analysis, thinking of it as little more than generic polling. As AllAnalytics.com blogger Cordell Wise has pointed out, "Just because a bunch of people on [Facebook] launch a campaign to bring back butter pecan [at a local ice cream shop] doesn't mean it's a widely held view." If 85 percent of sentiment about your brand online is positive, that doesn't mean 85 percent of the people love you. Perhaps almost all of the positive sentiment is about your product, whereas almost all of the negative sentiment is about your customer service. Context is important.
In whatever form they may take, your analytics efforts need focus. Otherwise, you risk finding yourself wondering "Now what?" with each new dataset. Retrospective step-by-step planning will help you focus on what you are going to do with the data before you determine what data you want to collect and analyze. This not only will help you avoid red herrings but will also allow you to plan the proper way to present your findings to colleagues -- folks who may have varied perspectives and diverse goals. This is especially important if political obstacles at your organization stand in the way of implementing "shake-up" solutions.
Finally, taking the retrospective approach will keep your analytics efforts straightforward and efficient, avoiding counterproductive guesswork. This translates to less time and money spent on what could otherwise turn your big data campaign into a big disaster.