Data is Stupid

AAEAAQAAAAAAAAQfAAAAJDY5MWY1YWU0LTE0NTEtNGUzMy1hODI1LTYyNjFmNTU1YzI2Ng

Data is stupid!

Data is not clever.

Nor is it helpful and insightful.

Data is very, seriously, stupid!

And yet data can be honed, with a little thoughtful craftsmanship, into something amazing, beautiful and genuinely powerful. Data can become intelligence.

Why is Data So Stupid?

Data is simply a collection of stuff and noise. Certainly, it has a vast amount of potential but data is like a pile of wood. A pile of wood is stupid. You can’t ask a pile of wood for shelter. One would frown upon such silliness.

What you can do, to keep this metaphor rolling along, like a log down a hill, is shape the wood, perhaps into planks. You could add in some careful planning, a bucket of nails, some experience in architecture and with hard work, build a home. Or a school. Or a barn. And now you have some shelter.

Data, is merely a material that one can use, if used properly, to shape a strategy. It could be used as the particulars with which one could grow a business. Data, without planning and effort and whittling and waxing, is inherently dumb.

The Leaky Bucket Of Web Analytics

As if stupid data wasn’t worrying enough, even when somewhat prepared, our reactions can cause data to be quite stupid.

Bryan and Jeffrey Eisenberg have a great analogy in their book Call To Action to describe a common, obtuse way that data can be used and misunderstood. Let’s say one looks at one’s website data and realizes that they’re only converting a handful of visitors into customers. Often, the reaction is to drive more traffic to the site. What does one do? Simply drive more traffic, get more sales?

The stupid data says “you’re converting visitors to customers.” Get more visitors and you’ll get more customers? To quote the Eisenbergs:

Trying to increase sales simply by driving more traffic to a website with a poor customer conversion rate is like trying to keep a leaky bucket full by adding more water instead of plugging the holes.

Smart analysis begins with a series of questions:

  • Is my bucket leaking?
  • Where are the holes in my bucket?
  • What is the easiest way to plug these holes?

And only then does the quick-witted analyst/bucket handler add more water.

The key to teasing out the answers is in the questions one asks. The answer to these questions is within the stupid data. And only when you’ve found the answers, will you have insights.

A Real Life Leaky Bucket

About a year ago, I had a conversation with the owner of a yoga studio, Amanda. She was using Adwords to drive traffic to her site but had a disappointing number of visitors actually signing up for classes. When we met for coffee, Amanda was thinking of increasing her advertising budget to get more students.

Amanda hadn’t crafted a real strategy surrounding the keywords so we began with a question that considered her ideal business goals. The question was: where are my best opportunities to grow my student base without increasing my ad budget?

A high bounce rate, alongside a low conversion rate was a big clue. The problem with such a big clue is that it points to many potential leaks. Stupid data, see? When we looked for the leaks, we found them.

Here are a few I remember offhand:

  1. Too broad of a geographic target audience
  2. Specific yoga classes were mentioned in the ads but the landing page was simply the homepage
  3. No clear call to action on the home page
  4. Site was slow to load

Together, we looked deeper at various segments. We noted that the highest bounce rates and the lowest conversion rates occurred in the following traffic segments:

  • Mobile visitors
  • Visitors outside the immediate geographic range
  • Visitors who clicked through on ads specifying new student discounts on classes
  • People who clicked on the ads for Ashtanga or Yin classes.

A high bounce rate often indicates a site speed issue – the page isn’t loading quickly, so new visitors shrug their shoulders and leave. I like starting with technical issues because they often have clear solutions and offer an easy victory. Some 3rd party plug-ins were indeed delaying the loading of the page. Amanda’s web guy made some tweaks to the site code which allowed the main content to load quicker. A small yet significant decrease in bounce rate was evident right away.

We noticed that mobile conversion rate was lower but because Amanda’s mobile traffic was so low, we decided to leave that for now. It simply didn’t warrant investment at this stage. This is another area where analysis beats data. Data demands you “create a mobile site” while analysis ask if it benefit me to build a mobile site today.

The geographic range issue was also easy to pinpoint and fix by tightening up the geographic targeting in Adwords. While this didn’t increase her number of students, it saved a worthwhile amount of money by avoiding clicks by people who were highly unlikely to sign up for a class in a neighboring town. Thus she was now able to put more of her budget toward likelier audiences.

We also decided to look at which landing pages were performing well with organic search traffic. We noticed that pages specific to yoga styles like Ashtanga or Yin or Vinyasa seemed to have high engagement. Shockingly, the paid search traffic about Ashtanga classes didn’t land directly on the Ashtanga page. Once this was rectified, the results were delightful.

Finally, we added a pop up to the home page with the new student deal clearly stated with a time sensitive call to action. This catered to those visitors who clicked on the ads offering new student discount.

I expected good results and I was not disappointed. The leaks were plugged, the student sign up rate increased to a profitable degree and Amanda was able to confidently increase the studio’s Adwords budget and grow even further.

To belabor the point somewhat, by setting our goals, by stating the questions clearly, by looking for specific answers and by being open to finding those answers, the stupid data became intelligence. Dimensions and metrics were transformed into analytics.

Why Should You Care About Stupid Data?

Well, stupid data can do a lot for you and your business when it is carefully analyzed and turned into intelligence. It can tell you when you have a leaky bucket. It can also tell you which hole is biggest or the easiest to fix. With analysis, you can even learn which tools you need to fix the hole.

So it’s time to draw my metaphors to a close. Remember this:

Data is stupid. You’re not!

Dean Levitt
About the author: 

Dean Levitt is the Founder at Teacup Analytics.  Previously he was the Director – Online Support at GoDaddy.com and the Chief of Culture at Mad Mimi, LLC.  Follow Dean on LinkedIn and Twitter.