Data mistakes, part 2: Analysis problems

data mistakes (Click here to expand)

Unfortunately, data mistakes don’t stop at the collection stage. It’s all too common to see marketers misunderstand perfectly clean data by employing sloppy analysis practices. If you’re worried whether your data analysis is lacking, here are a few common culprits that may be leading you down the wrong path—and how to correct course.

Analyzing with assumptions already in mind

When you look at your data thinking you already know the answer to what worked and what didn’t, you’re going to miss key insights. You’re also more likely to mistake correlation for causation, and misattribute changes in your marketing’s performance. Instead of looking for ways that your data can support a theory you have already formulated, look for patterns in the data and draw conclusions from them. This practice may force you to reconsider what’s working and what’s not working in your marketing, and that’s a good thing.

Confusing correlation with causation

It’s tempting to find meaning in data, even when there isn’t any there. For example, take a look at the results of this survey of Reddit users, in which data on users’ employment status and cheese preferences was collected. If you confused correlation with causation when looking at this, you may be tempted to think that being retired makes you more likely to enjoy cheddar cheese—or perhaps those who like cheddar cheese are more likely to be retired? Obviously, one of these doesn’t cause the other, but the correlation makes it easy to think so. In your marketing, it’s essential not to fall into this trap.

If your new ad campaign is associated with a drop in traffic, you need to look a little closer before assuming it’s all because of your creative approach. Have all ads you’re running seen a similar drop in the same time period? If so, it’s equally possible that there’s an outside cause. Instead of immediately drawing conclusions about a specific campaign, do some digging. Has the use of a new ad blocker recently spiked? Are there holidays or other events that might be causing your audience to be less present where you’ve placed your ads for the time being? Conduct similar detective work when you see positive changes in performance, so you can rule out things unrelated to your marketing choices before making campaign changes based on considerations beyond your control.  

Measuring data against the wrong benchmarks

When assessing whether your marketing is performing in a meaningful way, you should be comparing the numbers against benchmarks set during the strategic planning process. However, this can cause you problems if you have not set appropriate benchmarks.

What would be an inappropriate benchmark? If you’ve simply chosen an arbitrary number as a goal, it’s not going to tell you anything about your current actions. For example, if you’ve made your Q3 benchmark goal a thousand leads a quarter when you are working with a specialized B2B organization that has been reaching one hundred leads a quarter in the past, you’re not going to learn anything when you fall far short of the goal. Instead, look back at your performance and take note of what your average growth rates are. Then, factor in new marketing tactics, predict how they’ll benefit those rates and base your benchmarks off of that number.

Instead of setting this sort of goal beyond what you can hope to meet, create a series of challenging but achievable goals, which build over time to reach the numbers you need. Each of the smaller progressions you make will give you, through robust data analysis, more of the insights needed to build on marketing’s success.
Are you looking for more insight on how to keep up with the rapidly changing world of digital marketing today? Read our white paper, Return on acceleration for advice.

marketing budget

Download white paper

Comments are closed.