A Tale of Digital Marketing Attribution Woe

In Attribution Models, General, Google Adwords, Google Analytics by Kirk Williams

Janet and the Cows

ppc cowsJanet likes cows.

Ever since Janet first visited her grandparents’s dairy farm in small-town Wisconsin, Janet has had pleasant thoughts of her bovine friends. As she got older, Janet could not visit the farm as much as she would have liked, so she compensated by purchasing cow memorabilia off of the internet.

If you could put a cow on it, Janet would buy it.

One day, Janet’s friend told her about a new site, Cow Crap, LLC (https://cow-crap-it-all.com). Apparently, Cow Crap would take anything you mailed to them, stamp a cow on it, and send it back.

She was delighted, and immediately Bing-ed the term (Janet was not a fan of Google).  She found an ad by Cow Crap and clicked on it. 7 hours later, she left the site and headed back to work. She couldn’t get her mind off of Cow Crap, however, and she pondered which of her items to send in to be cow-stamped.

The next day, Janet woke up and checked Facebook immediately… upon which she saw an ad for Cow Crap! She couldn’t believe her luck, and wandered back into the site.

Over the course of the next 5 days, she would visit Cow Crap 12 more times.

4x through Facebook Ads, 3x by typing the URL directly into her browser, 2x through the emails she was beginning to receive, and 3x through ads in Bing.

cow crap llcAfter much deliberation, she ultimately decided to send in her bed-frame to be cow-stamped; so she typed in Cow Crap into Google, clicked on the add at the top, clicked the CTA on the LP, and finally gave Cow Crap her money.

Now, “attribution” answers the question, which channel gets the credit for this sale?

Here is where things get complicated in many Digital Marketing reports, because Google Analytics reports sales with Last Click (Non-Direct*) attribution. When the Cow Crap Data Analysts look at their Google Analytics reports, they see that this sale came by way of Google AdWords (at least, the last click was AdWords). They have noticed a lot of AdWords sales coming lately, which makes them feel warm and fuzzy since they are spending so much money on AdWords.

Their CMO on the other hand, is fresh out of her previous people management job (Cow Crap felt her “strengths” for this job were her people skills as opposed to her marketing skills), and she has been looking at Channel Reports.

“COW CRAPPERS!” She calls, demanding a meeting. “I have noticed that we spend far far far too much money on Facebook Advertising.” When I look at where our sales are coming from, it is clear they are not coming from Facebook Advertising.”

“But…” began a low-level analyst.

“No buts” She interrupted. “We need to put the money where it gives the best return. I want you to pull our Facebook Ads and send that budget over to doing all we can to increase our referrals and AdWords traffic, that is clearly where we make the most money. I want creative ideas on radio advertising, city bus ads, billboards, whatever you need to do to get the word out in a way that will increase referrals. Oh, also, boost our AdWords budget because that sends sales.”

Cow Crap did this, and unfortunately their sales began to dry up. You see, while AdWords branded keywords sent many sales like this, and thus AdWords got 100% of the credit for those sales, it was by no means the entry-point for most users (think about it, they typed in the brand so clearly they had to be familiar with it already). When they pulled their Facebook Ads budget, they ended up strangling (quite gruesomely, I might add) the Top of their Sales Funnel and drying up their sales.

The moral of this story? Don’t make foolish budget decisions based on Last Click only data. Rather, partner with or hire analysts that will help you get beyond Last Click in finding true value in all of your sources.  People often do not make a buying decision with one visit, make sure you understand the journey of your buyers before you make a decision you will regret.

*EDITS 4-25-17: Thanks to Peter O’Neil for correcting my original article to state that Google Analytics used a Last Click model, rather than a Non-Direct Last Click model. The post has been updated to better make sense with AdWords as an illustration rather than Direct. The basic premise of the article remains unchanged.