We live in the golden age of business analytics. Too bad that so many people turn to analytics to stop discussion instead of enliven it, and believe in analytics instead of common sense.
Don’t get me wrong: I love the analytics we have access to these days. And with the “we” in that sentence I mean you, me and every business that does anything on the Web—and what business doesn’t? Twenty years or so ago, businesses lived with marketing budgets that were mostly guesswork. In the early days of Palo Alto Software I’d spend on advertising without more than a dim hope of knowing whether or not the money was actually working. Back then it was print, not Web. Not that we didn’t ask customers how they heard about us, but only a third would even answer that question; and of those who did, barely half mentioned media we’d actually put ads in. So we guessed, and we hoped.
That’s quite different from today. I’ve read marketing experts seriously recommending that businesses only do marketing yields immediate analytics. I think that’s an exaggeration, but there’s a kernel of truth in that idea...somewhere. Clicks, click-through rates, conversions, e-mails opened, page views, hot spots…it’s like magic, compared to what it once was.
And I’d like that burst of available information, if it weren’t for the way it so frequently functions as a discussion finisher, the end of the thinking; when it should be the beginning of the thinking. Analytics should be the catalyst, not the end product.
Can analysis be creative? Of course it can. Here are a few examples.
1. Market segmentation. How do you divide your market into meaningful groups? Forget the traditional demographic or geographic segmentations and get into the right mindset. For example, a business selling to small- and medium-sized businesses divides its market by the decision process—owner decides, committee decides, tech people decide, etc.—rather than by size of company and industry. I’ve seen companies completely revise their marketing effectiveness by changing their segmentation scheme.
2. Creative skepticism. Look at the analytics and explore how results may have been disfigured unintentionally. This is where you notice, for example, that the survey of small business owners on technology is done by physical mail instead of on the Web. Look for the analytic design factors that can affect the outcome. Wording of questions and the selection of users and systems are often flawed. Just think of how outcome changes when you send something by e-mail vs. via Twitter or Facebook…and how small changes in the wording of a question can lead to major changes in the result reports.
3. Creative explanation. Take an extra step to look beyond the obvious to explain the changes in clicks or page views. Was it the change in the Web headline that you assume, or was it the snowstorm in the East, or the change in the holiday schedule, or the competitive announcement that drew your traffic away? Don’t just accept results on the surface—explore the hidden possibilities.
4. Watch the change over time. Watch for changes in analytics over time more than just the gross numbers. A conversion rate of one percent is neither good nor bad; if it’s up over time from half a percent to one percent, that’s one thing; and if it’s down from two percent, that’s an entirely different thing. Remember to keep definitions and parameters constant, so you’re comparing apples to apples, not oranges.
5. Guess the future. Take some time thinking about what might happen to change those key indicators over time. Look at opportunities and threats. Tell yourself the story of what might happen, and how it would impact results.
Conclusion: Love the analytics, use them, collect them, but use them to start the conversation, not finish it.