Drowning in data? You need to stop focusing on everything.
In this era of being data-driven, some people are being driven nuts by the volume of incoherent data they have at their disposal.
P.S. I’ll reference Google Analytics quite a lot in this piece because that’s the most common/popular web analytics tool.
Jumping into the world of analytics, it is very easy to get lost in data. The first time I dived into the deep seas of Google analytics, I felt like a kid at Toys r us. I wanted to check everything out. Page flows, time spent on a page, session duration, screen resolution of the user, everything!
In retrospect, that’s a rookie move.
An analytics rookie is more focused on the data set and the dimensions at his disposal, that’s the most important thing to him. From these hundreds and sometimes thousands of data sets and dimensions, he plans to work his way up to the big picture. The problem is, more often than none, he will get lost in the maze. When you’re engulfed in the hundreds/thousands of little stuff at the bottom, it is very tough to get yourself to see the hypothetical big picture.
On the ground level, there is so much to see, and even worse, when you think you’ve seen a lot, there is always more. You open the analytics tool to evaluate the performance of one of your Ad creatives and two minutes later you’re checking TreeMaps and content drill down.
Yes, it is important to know what pages your users visit, it is important to know how they navigate your website from their entry point to the exit point but was that your goal? And is this directly connected to your goal which in this case is to evaluate the performance of one of your Ad creatives? You can now see that you’ve gotten lost in data, like a kid at the candy store.
As with a lot of other things in and outside marketing, efficiency is increased when you don’t just dive into the task, you increase your efficiency and success rate by taking out the time to analyze your goal, the parameters required to achieve that goal, measurement plan, and the form of information/result you expect at the end of the task.
I know it sounds like I’m just waxing lyrical, but I’ll give an example using the hypothetical task mentioned above.
Goal: To evaluate the performance of one of your Ad creatives
Data required: Since it is an Ad campaign, the key parameters required would be common ad metrics and metrics that are relevant to the ad campaign’s objective. This includes Impressions (you may not be able to get this on Google Analytics depending on your campaign type though), Users and Sessions, CTR, Target Conversions (this may be sales, downloads, or even unique viewers/visitors if it is an awareness campaign), Conversion rate, then ROAS if your campaign goal is to drive revenue or something around those lines.
I’m not even halfway through and we can already see clearly what we need to check to gather information on the performance. Let’s carry on.
Measurement Plan: Now after pulling the aforementioned data, it doesn’t do much to tell you how well that ad creative is performing. It really doesn’t. To measure something, you need a standard or a yardstick. This yardstick can be your best-performing ad creative or an industry-standard — Wordstream does a great job at compiling this, but the accuracy of those standards aren’t quite spot on when you are looking at markets like Africa or Asia.
The form of information you expect at the end of the task: My use of the word “form” is intentional. I am not implying that you propose the result, if you do this you are susceptible to fall for a confirmation bias – observing and holding on to data that supports your proposed result. What I’m asking you to do here is to determine the form you expect your result to take.
In the case of this example, I expect my result to tell me if this ad creative is performing better or worse, tell me the metrics that were used to determine this conclusion, and show me the percentage by which it is better or worse. E.g. Impression is at par with the yardstick, but CTR is X% less, and Conversion rate is Y% less, hence this ad creative is performing poorly compared to the yardstick.
Piecing together this structure I’ve just illustrated may take you up to 10 minutes, but after spending those 10 minutes, you’ll have a birds-eye view of the situation and it will take you the least time possible to fetch the data you’ve identified to be important to your cause and to interpret this data.
I should mention that you do not have to write it down as I have, it can be arranged and laid out in your head, but ensure you do this before diving into the sea of data.
More often than not, reaching one data conclusion will raise another question, all you have to do is repeat the process (mentally or written down) and then take another dive. For example, if on comparing the performance, I see that the CTR and the conversion rates are at par with the yardstick ad creative, but the number of Impressions and number of conversions is lower, this will take me back to observe the ad delivery. It shows that the ad creative is being served less than the other ad creative. The next question is why? Then I head over to the ad server to see how it was set up, and so on.
With this “birds-eye view approach”, you will achieve more and get a clear analysis of your marketing performance and product performance without spending time drifting from one data set to another hoping something clicks.
I hope this was helpful. ✌🏽
