Don’t Let Your Analytics Deceive
If John F. Kennedy were an internet marketer, I think he would say, “Ask not what your marketing can do for your site. Ask what your site can do for your marketing.” Kennedy understood the idea of cause and effect in his original quote: that people make a nation effective, and not the other way […]
If John F. Kennedy were an internet marketer, I think he would say, “Ask not what your marketing can do for your site. Ask what your site can do for your marketing.” Kennedy understood the idea of cause and effect in his original quote: that people make a nation effective, and not the other way around. Online, the site makes marketing effective because the majority of what determines your ROI is not the media itself but what happens after the user clicks. And that’s what marketing analytics needs to focus on.
In the world of marketing web sites, we tend to treat the site as a given: fixed and unchanging. Search campaigns focus on driving “qualified” traffic, display campaigns focus on external branding and generating clicks to the site, email is pretty much the same story, as is social, mobile search and even print, radio and TV advertising that feature prominent URLs in the ads. We’re all trying to do the same thing online as we do in the real world. Get people in the door.
Online, we’ve figured out how to connect the dots more closely: unlike the real world where users come in the front door, a web site can welcome users to a specific product or category page, and we’ve seen how positively that has affected the performance of campaigns. Today, an effective marketing campaign drives relatively few clicks directly to the home page. On the contrary, a lot of marketing budget is spent on identifying users with specific needs and dropping them in front of hyper-related content.
But we’ve made a huge mistake along the way. We’ve told ourselves that the media has become more effective as a result of these “optimizations,” which is wholly untrue. The media hasn’t changed—we haven’t identified a new, better audience. The site hasn’t changed, either. We’ve just done a bit of the work the user would otherwise have to do on their own, and they’re showing some appreciation.
So on paper, it looks like the campaign is more successful. Conversion rates are up, bounce rates are down, maybe page views per session are down (which can be good—people found what they’re looking for quickly) and all of that wonderful stuff. But still, we didn’t change anything about either the media or the site.
What vs. why (yes, again…)
According to the Google AdWords keyword tool, about 12,000 searches are done monthly around the topic of web marketing analytics. This is great news because we know that people are interested in getting more out of their marketing. But it’s also potentially sad news: these searches have to be related to one of three things, and I think we all know which one it is:
- Learning how to optimize your marketing to maximize return
- Learning how to optimize your site and post-click experience to maximize return
- Learning how to put a bunch of numbers on a spreadsheet
The sad truth is that most people think of analytics as reporting. An analytics tool is a reporting tool: it’s the place you go to get your reports. So, it’s not much of a stretch to assume that most people searching for “web marketing analytics” are really just looking for ways to report what’s going on with web marketing. The problem is that a report tells us what happened, but not why. And worse yet, it often is very misleading about why.
Let’s walk through an example:
If you buy the keywords “basketballs” and “tennis balls,” and your conversion rate on basketballs is double that of tennis balls, we may assume that “basketballs” is an inherently better keyword, right? Well, that assumption could be completely wrong.
Next week, the opposite could be true when Wimbledon is on TV and people get excited about the sport. Or you might find that your price for tennis balls has its decimal point in the wrong place—people don’t want to buy a $374.00 can of tennis balls. Maybe your stock photo for tennis balls is bad, the user reviews on that page are all negative or the shipping cost of a single can of balls exceeds the price of the balls themselves. There are myriad explanations for why this may be happening, and the report captures exactly zero of these explanations.
When we optimize this campaign using report data alone, we would probably shift more budget to basketballs, potentially bid the keyword into a higher position, and probably scale back on tennis balls. Truthfully, the site itself is probably the last thing in the world we would look at to improve the campaign. But what if we figured out that the tennis ball product image was broken, fixed it, and now tennis balls was a better performer?
You can see where this is going: your marketing campaigns do not simply carry intrinsic value or implied return. Rather, you need to figure out why things are happening to know the right changes to make to your campaigns and your site. This begs an obvious question: do some keywords, banners, emails and tweets have more intrinsic value than others? Sure. But is what we’re seeing in performance reports necessarily indicative of this? No, not even close.
So, how do we do web marketing analytics?
What makes web marketing work is a series of things we can control influencing things we can’t control. Here’s how someone might interact with a banner, for example:
- Exposure (impression)
- Interaction (click)
- On-site experience (potentially complicated—these are the big why questions)
- Conversion (purchase)
These are things we can control:
And the things we can’t:
- Interaction (clicks)
In other words, we put content out there in our media and on our site, and if users like it, they will click or convert. We can’t control clicks or conversions: those are actions that users decide to take, and are measures of how effective the things we can control are. Looking at these metrics is how you get started with your analytics. If you see something that sticks out, don’t simply believe that’s just how it is. Start digging. You need to figure out why.
In my mind, the real meaning of web marketing analytics is also seeing your marketing budget as tuition, educating you how to make your site better. When we look at cause and effect, post-click experience is without a doubt the key driver of ROI in your marketing, but site experience is usually not a part of making web marketing campaigns better. While it may be a challenge getting marketing people more involved in content development (and the other way around), the payoff of getting these people working together is enormous, particularly if the site gets the majority of its traffic from search engines and marketing channels.
JFK was way ahead of his time, and would have made a great web marketer: “Change is the law of life. And those who look only to the past or present [reports!] are certain to miss the future.”
Opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.
New on Search Engine Land