eBay commissioned a 26 page study through their research lab with one of their researchers and some reputable sources (a Univ. of Chicago Assistant Professor & UC Berkeley Associate Professor), citing some reputable resources (like Google Chief Economist Hal Varian), doing some pretty in depth research and even formulating their own algebraic equation to express ROI to explain and back up their conclusions.
I have to say I don’t totally disagree with it.
Here’s the abstract of the study:
Internet advertising has been the fastest growing advertising channel in recent years with paid advertisements on search platforms (e.g., Google and Bing) comprising the bulk of this revenue. We present results from a series of large-scale field experiments done at eBay that are designed to detect the causal effectiveness of paid search advertisements. Results show that brand-keyword ads have no short-term benefits, and that returns from all other keywords are a fraction of conventional estimates. We find that new and infrequent users are positively influenced by ads but that existing loyal users whose purchasing behavior is not influenced by paid search account for most of the advertising expenses, resulting in average returns that are negative. We discuss substitution to other channels and implications for advertising decisions in large firms.
Now before the PPC marketers out there get all wound up, here’s a quick recap with details per the study:
- eBay is calculated to spend $51 million on AdWords annually
- eBay wondered what would happen if they stopped buying on their brand keywords like “eBay”
- Bing/MSN was included as the first experiment in the turning off of branded search. Google remained the control. Only 5.9% of traffic is estimated to have been unrecovered by organic on Bing.
- The same brand keyword test is rolled out with Google, resulting in a 3% loss unrecovered by organic.
- eBay wondered what would happen if the stopped buying as much on some of their 170 million (historically speaking) unbranded keywords in certain DMAs. They turned off spend in 30% of their 210 DMAs.
- A random sampling of those DMAs left are split into control and test groups.
- The non-brand left to run is calculated to attribute only a .44% increase in sales.
It goes on to explore customer and product response heterogeneity (types of customers/visitors as they relate to the bigger picture) as well as channel substitution and how they derived ROI with this fabulous equation:
It’s nice that they were able to afford what I’m sure was a very costly study to prove one thing any search marketer could have told them- “you’re a major brand/household name. If someone types in “eBay XXX” you’re going to show number one for organic and they’ll just click that instead of the paid ad since your SEO isn’t a train wreck.”
Google is of course going to show the most relevant “eBay XXX” it can find in its index. It’s not a shock that turning off your brand keywords led to an increase in organic sales and saved you some money. Any competitor that’s going to go after eBay brand terms is going to get nailed in CPC and Quality Score as well as be banished to the right rail for ad placement more often than not. That’s how the system is supposed to work.
So it’s pretty cool actually to show some numbers and proof that super mega brands could save themselves a few bucks by not bidding on their brand…despite what their Google reps might say and even myself. (I’ve been known to encourage and buy on branded terms, but then again, the client wasn’t eBay, Amazon or Macy’s.)
The Non Brand Keywords
The non-brand side is where I have some disagreement and wonder if the researchers in question were search marketers themselves and how much they understood in regards to the finer points of PPC to draw their conclusions from.
170 million keywords is an awful lot. $51 million in spend means that Google AdWords reps had to be pretty heavy handed and involved in the processing of these monies. I know from personal experience that when you come to them and say “I have a million dollars to spend” they get really helpful with all kinds of places that you can spend that million with mixed results and advice.
Now don’t get me wrong, the people behind these suggestions are often times brilliant, super nice and mean well. But let’s face it, there’s a vested interest. It makes you wonder how the account got that big and how many calls with Google ended with “we’ll build out all those new campaigns for you right away!”
It’s been a long time running joke with PPC marketers on the quality of the eBay, Amazon and Target ads. Those guys will buy and sell anything- literally:
The Portent PPC team used to spend Friday afternoons googling around for the craziest or most nonsensical ads we could find, thanks to a liberal use of dynamic keyword insertion.
While I don’t think eBay needs to be handing $51million over to Google AdWords every year, they might want to take a second look at the quality of the buy they were making.
In conclusion: can eBay live without AdWords?
Duh, yes. They’re eBay.
Should they cut everything off entirely?
Well…the researchers accounted for seasonality, brand awareness, DMAs, volume, types of customers (frequency and recency) but not the part where the quality of the work being done was scrutinized. I’m not into drinking the Google Kool Aid, but I am in defense of good PPC work.
All else, they’ve got enough money for for twelve 30 second Super Bowl ads now.