blog

It’s October again, and that means it’s time for retailers country-wide to fight it out to acquire the millions of potential customers looking for the scariest, sexiest, and creepiest Halloween costumes around. But while the prospect of grown men wearing Sesame Street costumes may be terrifying, we’re here today to present some equally (if not more) frightening facts about mistakes you might be inadvertently making in managing your online marketing campaign. Read on, if you dare.

Odds are that you, or a marketer at a company you work with, uses Google Analytics to help manage online marketing campaigns. That’s because it’s a free tool that appears to track marketing, and so has become the default site metrics measurement tool for many online marketers. The truth is, relying solely on Google Analytics “visits” as a metric of marketing performance can be misleading at best, and a dangerous business practice at worst. If you take a close look at how Google counts visitors across Google Analytics and AdWords, you’ll realize that oddities in some of their counting mechanisms, while subtle, have the potential to drastically alter where you, as a consumer of this data, think it makes sense to spend your advertising dollars.

At Convertro, our goal is to provide marketers exact visibility into their customers’ marketing interactions prior to purchase. When they start to dig into our data, prospects invariably ask why our visits count doesn’t match what they see in Google Analytics (of course, they expect GA to be the gold standard). The development side of our company had been relatively skeptical of Google Analytics’ numbers, but the client services side wasn’t convinced, so we did some research to find out who was right and who was dead wrong.

The most glaring piece of data we discovered involves the fact that Google Analytics considers any number of visits to your page by the same user within a thirty minute window to be a single visit, rather than multiple instances of a user landing on your page.

Visits represent the number of individual sessions initiated by all the visitors to your site. If a user is inactive on your site for 30 minutes or more, any future activity will be attributed to a new session. Users that leave your site and return within 30 minutes will be counted as part of the original session.(http://www.google.com/support/analytics/bin/answer.py?hl=en&answer=57164)

Why is this an important fact to note? Because if you under count the number of times users land on your page, you’re also under counting the number of ways they get to your page, and thus you’re missing information on what drove those users to your site in the first place.
We speculate that GA is combining multiple visits to reduce the number of inserts to their database, not because they are trying to skew the numbers to their benefit, but the result is the same.

Consider a user looking to buy a sexy Halloween costume for this year’s festivities (pretend, for the moment, that you are a purveyor of such goods). She searches on Google for “halloween costumes” and clicks on your enticing PPC ad, bringing her to your home page, CrazyCostumeCreations.com. After looking around a bit and not seeing anything all that sexy, she decides to try another web search, this time using Bing. A PPC ad on Bing with a creative actually mentioning your large selection of sexy costumes delivers her back to your page, where she finds the perfect Big Bird tube dress and checks out about half an hour after starting her search. One more customer satisfied!

In the Google Analytics world, her purchase was driven by one referral — because they both happened within the same half hour, her two visits to your site are considered just one. But in the real world it’s clear that both Google and Bing PPC (and, importantly, the two different creatives) are responsible for her purchase. The PPC click from Bing, however, won’t show up in your Google Analytics numbers, making your spend on Bing seem like a poor investment. In contrast, the Google PPC click ends up seeming more effective than it truly was. Scary, isn’t it?

Now let’s pretend our user is a Google loyalist and instead her whole search process takes place on Google. After her initial “halloween costumes” search, ad click, and dissatisfaction, she decides to give Google another shot and searches for “sexy halloween costumes.” This time she sees your more targeted creative on Google, clicks, and the rest is history. How are your GA numbers leading you astray this time? First of all, GA thinks this is one visit when it’s really two, so right off the bat you’ll be over-estimating your conversion rate. You’ll also think all the revenue from the Big Bird dress is the result of one visit instead of two, so you’ll over-estimate your revenue per visit. And, to add to the terror, this purchase actually cost you twice as much as the visits count in GA would lead you to believe.

So why not integrate your Google Analytics and AdWords accounts, and then use the “clicks” metric in GA that’s available if you do this? This metric is supposed to reflect the clicks that come to your site regardless of when in a user’s session they occurred. But now we must deal with the issue of over-counting clicks, because invalid and spammy clicks are reported as clicks in GA but aren’t counted in AdWords. Now, why should this strike fear into your heart?

Say that our friendly user lands on your site after her first search, hits back to do her second Google search but then suddenly remembers that she did see a link near the bottom of your page that she could have sworn started with the letters S-E-X. She hits forward to check it out again and makes her purchase. In this case GA will count one visit (remember the half hour rule) and AdWords will count only one paid click thanks to some magic deduping logic (read about GCLID’s [here]). But GA will count two clicks because Google doesn’t apply this filtering logic in Analytics. Unfortunately for you, this over counting of what seem like paid clicks will artificially deflate your conversion rate and your revenue per visit (you think you have more clicks than you do). Not to mention, when you see those extra clicks in GA coming from Google PPC, you’ll runscreaming to your accountant and demand to know how your cost per visit got so high.

The simple fact is that Google Analytics isn’t designed as a platform for managing marketing. It’s a tool meant to enable analysis of aggregate data about your website, and to give insight into common referrals and the paths users traverse on your site. Google AdWords, on the other hand, is an excellent marketing platform for managing marketing — as long as that marketing only uses Google products and properties. If data-driven marketers want to help their clients make a killing this Halloween season, they need to do their job with the best resources available. That means understanding exactly what Google Analytics is capable of and using it appropriately, while leveraging other tools that are designed specifically for accurately tracking marketing. So this October, whether the prospect of sexy cartoon character costumes sends you running for the hills or not, don’t let your GA numbers scare you into misallocating marketing spend!