Do Something With Your Marketing: Explore Your Data

In this week's do something with your marketing, I want you to do more with your data than just accept what the vendor gives you. Part of marketing data science is applying that sort of scientific thinking, that scientific curiosity, that inquisitiveness to your data, rather than just blindly going to an application and saying, okay, there's my data, and you look at it and chart either went up to the right or down to the right and called a day. Yeah, I get it. We're all busy, and for some folks, there you may not have a formal background in in statistics. That's fine.

You don't need to have any of that. You just have to be curious and want to get more out of your data. So let's look at an example of how might you do this. So let's go ahead and bring this up here. Here we have stock Google Analytics, right?

Now, what we see here, just by looking at it, most people look at this and go, okay, well, December is kind of a crappy month because it started uh high and then kind of went down, and okay, so it was it wasn't a great month, right? The holidays, all that stuff, and you move on with your day. That's not great, right? That doesn't tell us anything. It doesn't give us anything to make decisions with.

All we can see is there it is. Remember, data without decisions are decorations, right? Or they're distractions. Uh so let's see what could we what else could we get out of this? Well, if we were to export this from Google Analytics using the export button here and choose Excel.

I already have already opened this in Excel. Uh let's bring that over here. Here's our data set. Now, what are some things that you might want to know? Maybe you might want to know, for example, the uh the arithmetic mean, which is the average, right?

So we use average. What's the average number of visitors that month? 392.774. 393 visitors, right? Let's leave that out there, make sure that's not messing things up.

Now, what about the median? The median, remember, is the central number. If you reordered this data set, it would be the number square in the middle. Let's go ahead here. Median 395.

So isn't that interesting? The average is lower than the median. Generally speaking, when with marketing data in particular, when the mean is above the median, it means there's more anomalies on the upper side. There's more stuff happening on the upper side, so that it's you're pulling uh you're you're pulling your data up because there's the the classic statistics example is this you know, you and four of your friends are sitting around uh a table having a meal, and Bill Gates walks in the room, suddenly the average, the mean income on the room goes to like 100 million dollars, right? You know, you didn't get any richer.

Um, but that skew from that one outlier uh pulls everything up, whereas that doesn't happen with the median. That's one of the reasons why we say it a lot in marketing statistics and marketing data analytics, use the median rather than the mean a lot of the time. But the two together can give us little tiny insights. In this case, my mean is below my median. So I had a worse than average month, right?

I had a worse than normal month because things are on the downside. If we if we were to color this in, let's let's do this. I'm gonna do just a little bit of conditional formatting with the fun little data bars feature, and extend that out so we can see it. Now, what happens if I resort this column here by largest to smallest? Like so.

Right? 395, there's our median. It essentially was having it, we had a worse month on the downside than we did on the upside, and that's why the mean is lower than the median. Now, if I had uh a fantastic month, let's I'm going to manipulate this data here. I'm going to put the 15,000 visitors, right?

And now, of course, the mean is substantially above the median. The data didn't really change uh except for that one cell. But you can see how that changed. So this little indicator, just comparing these two things, tells us what kind of month did we have, right? Or what kind of week or quarter or year.

This is a useful insight that tells you uh and specifically it tells us I didn't do a great job with campaigns, right? You typically get anomalies in marketing data, like that 15,000 example I just showed you. When you've done a big campaign, a big push, some kind of big launch, right, and you get that big one or two days that month where the traffic is off the charts. And so the average, the mean is above the median. So campaigns went great, right?

Maybe the rest of the month maybe not as good, but campaigns went great. In this case, I didn't run any campaigns in December, right? Uh it was clearly a mistake because my mean is below my median. So for January, I better have some kind of campaign, something to drive traffic to the website, drive conversions, drive all those things that we care about so that my mean is above my median. It means that I'm doing a good job with campaigns.

Right? Now, if there was a substantial difference between these two, as we just saw. Uh, clearly that would show that maybe I'm a little too reliant on campaigns. Um, but in this case, I'm just not doing a good enough job. So that's what I want you to do with your marketing this week as your homework.

Start looking at just some basic statistics. This is high school level statistics at best, right? Actually, it's not because I think one of my kids just went through this in like seventh grade, eighth grade. So um start trying these techniques. Just play with some of your data and see what it tells you, what it hints at, and then have two numbers, just two measures right now.

Compare them and try to make a decision. As you saw, my mean was below my medium, decision made. I need to campaign at least one in January so that I don't get a repeat of this. So I need some kind of campaign to push that mean back above that median. So that's what I want you to do with your marketing.

Try this on your own data. As you can see, we're not using anything sophisticated. There's no advanced analytics, there's no machine learning. This is just straight up good old statistics in a spreadsheet. Try it out, see what decisions you can make.

As always, please leave your comments below, subscribe to the YouTube channel and the newsletter, and I'll talk to you soon. Take care. Want help solving your company's data analytics and digital marketing problems? Visit TrustInsights.ai today and let us know how we can help you.