2.2 What are key types of Marketing Analytics? - Video Tutorials & Practice Problems
Video duration:
14m
Play a video:
Video transcript
<v ->So let's take a deeper look</v> at what you can do with marketing dashboards. Now, obviously it's a way to visualize your data and that visualization usually takes the form of a table or a graph. And if you think about the things we do for reasoning with numbers, it's going to help us do that aggregation, but also the segmenting and filtering that let us compare things against each other. And a really specific type of comparison is a trend line where you're basically looking at how things change over time. These are all kind of basic functions that marketing dashboards do. Now, in order to use a marketing dashboard, you need both marketing and analytics skills. Now, remember don't be afraid because the analytics skills are those three things of aggregating, segmenting, and comparing. So everybody's able to do that. Now there's actually three kinds of analytics. We mentioned earlier descriptive analytics, but there are two other kinds too. So descriptive analytics basically tells you what happened. And so an example of things that help you with descriptive analytics would be a dashboard or maybe kind of the screen on Facebook insights 'cause that's a kind of dashboard too. That's a specific dashboard. Dashboard's usually good at to telling you what happened. Now, predictive analytics are different. They're trying to tell you what's going to happen. So something like a customer retention model is predictive analytics 'cause it's trying to tell you which customers you're in danger of losing. Google ads has a traffic prediction model. And what that will tell you is, it's going to give you a guess as to how much traffic you're gonna drive with the ad that you're contemplating running. There's a third kind of analytics called prescriptive analytics and prescriptive analytics use the predictive data in predictive analytics, but instead of just showing it to a person to take action, prescriptive analytics actually put the analytics to work by having the computer take an action based on the analytics. So account-based marketing is an example of that where the computer might actually be targeting, sending emails to prospective customers that have certain characteristics or realtime content recommendation. The way Amazon does when it shows a product and says other people who looked at this page bought this product. And so these are ways that the computer is actually automatically taking action. So descriptive tells you what happened, predictive tries to tell you what's going to happen, and prescriptive actually takes action on what's going to happen. And these are all three different ways of thinking about analytics. And so, as you are using them, you're basically going from problem to insight and with prescriptive analytics, you're even going all the way to action, where you've made a decision and you're taking action on it. And so this is really what the journey is when you're using analytics of how can you go from problem to insight even to automated action. So let's look at some examples of dashboards and most of these are going to be descriptive analytics. You might imagine, you're not gonna have a dashboard of predictive analytics 'cause it's just taking action. You can have dashboards for predictive analytics and we'll show you an example of that too, but let's break these down according to what they actually help with. So this one helps you attract people to your website. So the question this dashboard asks, is "which channels are performing?" So is email doing better than display or search or social? And so this can be helpful to compare results across programs because that can help you to prioritize spending in your marketing mix. And so what kind of insights can you draw from this type of dashboard? Well, you can look at trends to see if some channels are up and down. You can look to see if there are campaigns that you planned for omnichannel to actually optimize across many channels at the same time. Are they doing better than single channel campaigns? And maybe you could actually segment by business unit to see if there's success in different channels that vary by business unit. So these are all things you can look at in this type of dashboard. Let's look at another dashboard that's also focused on attraction. You might ask yourself, "which keywords are we missing from our search marketing programs?" So the reason this can be useful is because you might be optimizing your pages and SEO for certain keyword, but not buying it in paid search. Maybe you just overlooked it because there are two different teams. So what kind of insights can you draw from this? Well, you can obviously figure out if there are additional keywords that might drive traffic to existing content because you're going to add it to the campaign it was missing from. You also can do something that helps you to identify whether you can increase conversions for SEO because those things are usually focused on for paid search campaigns. Or if you can lower your costs for paid search because SEO really focuses on the quality of the page. And by raising your quality score, you can actually reduce your spending in paid search and still get just as much or even more traffic. Let's look at another dashboard. Again, this is an attraction dashboard. It asks, "which content is shared and why?" Well, what does this do? Well, what it's really looking at, is which pages on your site are shared the most on social media and which ones draw the most inbound links from other websites. So these are all ways of you sharing your content. It might give you an idea of which content is somewhat popular with influencers and other people who might draw attention to your site. And so, we know that content that's shared by others converts at a higher rate than when it comes direct from you. So knowing that this content is being shared could be really important. Now, what kind of insights can you draw from it? Well, maybe you could segment the content by what topic it's about and see whether there are some topics that are shared more than others or maybe you could segment it based on whether the content seems to be branded or not. So is it mentioning your brand names? Is it about the products? Or is it maybe earlier in the sales cycle and it's kind of unbranded, generic content and which one is being shared more? And maybe you can compare it against your competitors because you can look at their numbers for shares on social media and inbound links. You might be able to compare your content against your competitors content and see which one has a higher market share of being shared. And so this is some things you can get from this type of dashboard that focuses again, on attracting people to your website. And all of these are descriptive analytics dashboards. Here's another one that says, "which landing pages need work?" Now we've shifted from attract to retain because which landing pages need work might be having to do with which ones have a higher bounce rate. And so if you work to attract the traffic, it would be great to retain that traffic by improving the landing pages. And you can look at some insights that say. "how do bounce rates differ based on channel source?" Then you're segmenting by channel and seeing if where they came from influenced the bounce rate. You could also look to see if targeting is the problem or the content or both, but what do we mean by that? Well, suppose for example, you have a high bounce rate on a landing page when people come to it from your email campaign. Well, is the problem that the landing page is bad or did you just send the email to the wrong people? So that's the difference between targeting and content. Or maybe the email content isn't very good and the landing page doesn't really pay off on it. So by looking at these things, you can kind of take a guess as to what needs to be improved next. You could also segment the landing pages by topic, as we talked about before, and see if maybe certain topics tend to have higher or lower bounce rates. Let's look at another retain dashboard. This one asks the question, "which questions don't we answer?" So what does that mean? Well, what if you looked at each page on your website and check to see what the search keywords people were typing in to your site search keyword box on each page. You would probably find that there are patterns that on certain pages, there are certain keywords that are being typed in over and over again. You know what that might tell you? That those questions are not being answered on that page. Nobody looking at that page can find the answer to that question and that question's coming up for them a lot. And so what kind of insights can you draw here? Well, you might know which marketing messages aren't working. You might also segment and see, "Hey, are prospects and existing customers still having the same keywords that they're entering? Do they have the same objections, the same questions?" And maybe you could do it over time and see whether maybe some seasonal differences occur. Are there unanswered questions that change by the season? Now, all of these things I'm showing you are just examples. So you might have different questions you're trying to answer in your organization. You might segment by different things 'cause you're looking for different insights, but I'm showing you these examples to just get your brain moving and start to think about what kind of dashboards you want to solve the problem that you've selected. Here's another one, another retain dashboard, "which topics are emerging and ebbing?" So again, if we split things up by topic, we segment by topic, maybe we'd find out whether certain topics are trending up and down in terms of popularity. You can also do this on your competitor's site because if you can classify your pages by topic, well, you could crawl their site and classify theirs as well. And that might be useful because it might drive prioritization of which topics your content should be about in your content calendar. Now, what insights can you draw here? Well, you might find out that your competitors are focusing on topics that you are not and maybe that would be an interesting thing to experiment with. You could also segment by seeing whether the emerging topics are being consumed more by different segments than by others. So maybe you could segment based on what industry people were in or what geography they're coming from. And so segmenting might give you some more insights and you can also do, as we mentioned before, you can look at the market share of emerging topics for you versus your competitors to see if maybe you're getting ahead of them in terms of grabbing the attraction and the attention for people against the topics that really are the ones that are climbing the fastest. So here's another retain dashboard. This one shifts gears a little bit. All of the ones we've shown so far are descriptive analytics, telling you what happened. Here's one that's actually using predictive analytics. So this one asks the question, "which pages will have high exit rates?" So it's going to predict based on the qualities of the page, whether that page is going to have a high exit rate or not, maybe even when you first published the page because sometimes a low traffic page might take months before you get enough page views that you even find out that you have a problem. So obvious insights are which pages need improvement and whether high exit pages might be written at a higher grade level and maybe that's one of the reasons why they have high exit rates. You could also ask whether our pages have higher exit rates than our competitors because if you can run this predictive analytics against your pages, you can run them against your competitors pages as well, just by crawling them. And here's our last dashboard. This one's about convert, "which pages are the most valuable?" So you can look at a list of pages ranked by their contribution to conversion value. Now, remember you wanna use causation, not correlation. So you don't wanna just look at pages that are correlated to conversions, but you might wanna throw some of them out and maybe make some adjustments based on which products are the most popular, make some adjustments of where things are in the navigation, the closer they are to the conversion, the more likely they are to correlate with a conversion. So you can make some adjustments that help you show what their real contribution is rather than just their correlation. And so, by finding that out, maybe what you'll find out is that high value pages have different qualities than low value pages. And maybe you can emulate those qualities in the low value pages to improve them. You can also draw some insights that for example, say, "hey, is there more value for this page for email versus search? Or do high value pages have lower grade level content?" Or again, segment by topics and see which topics the most valuable pages are about. These are just a few examples of dashboards. There are many more, they're limited only by your imagination. So if you didn't see the dashboard that's good for the problem you selected, make one up, come up with your own dashboard. What do you need it to show you? How can it help you solve the problem? That's what your job is, to think through your problem so you know what dashboard's right for you.