Under the Hood: Measuring Purchase Intent of PPC Traffic In Google Analytics

Is paid search predominately a research mechanism? Or do visitors have strong purchase-intent when clicking on paid advertisements? Does PPC predominately expose new visitors to your company? Or does it seal the deal? Out of the box Google Analytics doesn't return this data. Google Analytics tracks the last referring source and overwrites the previous. So, if a visitor is exposed to your company via PPC, but then makes a purchase during a later visit via organic search, the organic visit gets the credit thus blinding you to your PPC's conversion rate.

Here's what we did to try and get at this data (still not sure if it will work).

The Goal

We wanted better insight into exactly how PPC traffic was behaving for a catering client.  Some keywords generated high click-through-rates (CTR) and decent conversion rates; but other keywords had even higher CTR's but lower conversion rates. Before abandoning the "poorly converting" keywords, we wanted to be sure we had the whole story.

Google Analytics + The Purchase Cycle

As I mentioned previously, Google Analytics tracks the latest referring source. Direct traffic doesn't overwrite the previous source, but all other referring sources overwrite one-another (organic search, PPC, referral, banners etc.) Let's identify a possible (simple) use case:

  1. Billy-Jean is searching for "catering in Washington, DC".
  2. Billy-Jean clicks a PPC ad, lands on the site, checks out prices and the menu then exits.
  3. The next day, Billy-Jean searches Yahoo! for "Catering Company Brand Name" and drops $100k on a whole lot of soda.

Why it Matters

If we were viewing Billy-Jean's process in Google Analytics, we would see a Yahoo! organic search visitor searched for "Catering Company Brand Name" and purchased $100K of soda. We would also see that the Google AdWords keyword "catering in Washington, DC" generated one visit but no conversions.

You can see why this is a problem. We're left thinking the PPC keyword "catering in Washington, DC" isn't producing revenue and is a big waste of money. But in reality it was the rainmaker.

Our Solution

Using the Google Analytics User Defined Variable (setVar); Tony wrote some JavaScript that automatically tagged every paid search visitor with the user defined variable of "CPC". We then created a GA profile that filtered in only visitors with the variable "CPC". The best part? The user defined variable (UDV) doesn't expire, and it doesn't get overwritten unless you tell it to. In other words, that profile contains multiple referring sources (direct, referring sites, MSN organic); however, we can identify all visitors who arrived via paid search because of the UDV.

Want to know how to do it? Check out Tony's post on "GA User Intention Tracking" on our developer blog.

How We're Planning to Use the Data

Well, like any marketing shop, we're going to let the data talk to us, not the other way around. But, here are a few ideas of how we might use this data:

  • If we discover PPC is simply an exposure piece, we might consider building a “purchase” PPC campaign that contains keywords such as "order" or "affordable" to capture visitors with a stronger intent-to-purchase.
  • If we know people are using PPC for research but usually purchase after 10 days, maybe we'll have a section of the homepage say, “Welcome back! Last time you visited you check out Boxed Lunches. Still interested?"

The important thing is we have the data, which GA couldn't give us out of the box. Now it's about using this data to make smart marketing decisions on behalf of our client.

Any questions?

Josh Chambers

,
Posted in Article Category: #Strategy
on