Search arbitrage is the practice of sending paid clicks to landing pages that offer AdSense or search partner results and then profiting off the click on your site. The idea is that you pay less to attract the traffic than you make from driving the traffic to the other advertiser, thus earning money.
The practice was rampant many years ago. AdWords policy updates and the increased challenges of earning a margin via this method have largely stamped it out. But, not entirely.
Ask.com in particular has been using this method for years. In fact, 6 years after Danny Sullivan pointed it out and Ask.com blamed it on a mistake by a vendor SEM agency, Ask.com arbitrage ads still litter the SERPs.
In early 2007, Google updated the Quality Score algorithm to factor landing page quality into the QS calculation; formerly, QS used signals from the ad text only. In September of that year, almost exactly seven years ago, Google announced that “arbitrage sites” in particular would be penalized with lower quality scores with the following snippet describing what is considered a “low-quality landing page”:
Arbitrage sites that are designed for the sole purpose of showing ads
This month, the AdWords policy center received another overhaul, first announced in June. One of the more notable changes was the introduction of a provision that specifically disallows “low-quality content,” which includes “sites that offer little unique value to users and are focused primarily on traffic generation.”
Previously, there were several disjointed content prohibitions such as cloaking and making false claims. Those are still there, and they now fall under a section of provisions barring “low-quality content.” The statement that banning “low-quality content” is an AdWords priority was seen by many as Google formally mirroring the principles of the Panda organic crackdown on poor-quality websites.
That association makes a lot of sense. Recently, many organic changes have been followed by similar updates on the paid side. In April, AdWords had query data stripped from paid ads, creating parity with the organic change to secure search and not provided keywords released in October 2011.
So where does arbitrage fall in this new “low-quality content” policy? It’s still called out as a prohibited practice, though the language has changed slightly:
Content that is designed for the primary purpose of showing ads
Examples: Driving traffic (whether through “arbitrage” or otherwise) to destinations with more ads than original content, little or no original content, or excessive advertising
The main difference compared to the 2007 language is that “sites” is now “content” and “sole” is now “primary.”
When I first read the new policy in June, my first reaction was to wonder where Google would draw the line. From one angle, network television shows, fashion magazines and even The New York Times are, from a business model and monetization standpoint, “content designed for showing ads.”
No matter the quality of the content, or the passion behind producing it, content as a vehicle for advertising is the very basis of what we call media.
As paranoid as search marketers can be, and as open to interpretation as the new policy is, no one that is producing high-quality original content which is monetized via advertising is particularly afraid Google will be kicking them out of AdWords.
Surely, Google is talking about that very obvious arbitrage such as where Ask.com shows on searches for everything and sends the user right back to a page of Google paid results on its site.
After SEL founder Danny Sullivan calling out Ask.com’s practice in 2008, editor Pamela Parker, responding to a reader inquiry, wrote about the practice again in 2012 asking, “Is Ask.com Continuing To Play The Google Arbitrage Game?”
Ask.com CEO Doug Leeds responded in the comments of that piece, saying that users liked the ads, that they drove high retention rates, and pointed out Ask.com’s purchase of about.com as proof of their investment in quality content.
Here is a screenshot from a search this morning:
There is Ask.com with the familiar ad format it’s been using for years. It is joined by About.com, the company it acquired to invest in quality content. The landing pages of these two ads are almost identical except for the site logo. It appears that not only is Ask.com still playing the arbitrage game, it is also double-serving, another prohibited policy.
The obvious next question is how are these sites making money off this. Several factors contribute to the possibility that they are squeaking out a margin:
Having said all that, it’s also very likely that Ask and About are not actually profiting directly on this arbitrage. If the simple click arbitrage isn’t profitable for them, why would they do it?
The answer: To collect first-party data.
Ask.com (and About.com) are owned by the massive, public internet giant IAC. IAC’s network of sites includes many other top web properties like Match.com, Investopedia.com, collegehumor.com and dozens of others in basically every category.
IAC also owns the recently launched Conversant Media, a conglomerate of previous smaller media brands and ad networks. Here’s a snippet from the Conversant website
“We don’t rely on commoditized third-party data, instead we leverage vast amounts of exclusive browsing, shopping, interaction and purchase data not found anywhere else”
Basically, the likely play is that Ask.com is collecting first-party search behavior as a signal that powers its ad network via AdWords ads and subsidizing most of the costs of that data collection with the search partner ads presented on its site.
It’s a much more complex kind of arbitrage than the simple click-in, click-out model. It also seems even more in opposition to Google’s policies on privacy and quality sites. It’s not exactly against the arbitrage policy that forbids “Content that is designed for the primary purpose of showing ads” – it’s content designed for gathering first-party data and selling ads elsewhere.
We think of arbitrage as Ask.com and About.com making a shady play at some easy small margin revenue with giant scale, but it’s easy to forget that they are only getting around (probably) 65% of the cost of that click on their sites. The other 35% goes to Google and happens immediately after Google’s first, 100% of revenue, albeit lower CPC, click from the original search.
If Ask.com paid 50 cents for that first click to its site from Google.com and then sends it to a result that drives a 68-cent click, if they get a 75% revenue share on the 68-cent click, Ask squeaks out a penny on the whole deal.
Google, which gets the full first 50 cents and its 25% share of the second click, makes about 67 cents on the deal – 67 times more than Ask.com. This math assumes that Ask.com is making a margin on those Google clicks, which may or may not be true.
Also interesting to consider is that in the Panda 4.0 update in the spring, Ask.com was identified as one of the biggest losers. Some estimated that Ask.com lost 50% of its organic traffic.
The organic Ask.com listings are just as packed with pre-loaded search partner results as the paid search landing pages. Google lost a big chuck of its own search partner revenue when it bumped Ask.com out of the SERPs.
Is Google trying to recoup that revenue? In 2010 Google banned 30,000 low-quality advertisers from AdWords, confident that the net effect would be positive. That was the tail end of the always-increasing CPCs for Google, though.
Google had eight consecutive quarters of lower average year-over-year CPCs and huge pressure from stockholders to continue growing quarterly revenue at a rapid pace. It is grappling with declining desktop impressions and lower CPCs on mobile.
One way to squeeze out more revenue is by turning a blind eye to arbitrage players that facilitate Google’s getting paid for users, especially those on lower-quality searches, clicking twice and getting 2x the revenue.
It’s been several years since Google went after arbitrage sites in any aggressive or highly-publicized manner, so perhaps the signals Google uses to identify low-quality or arbitrage sites have weakened relative to current practices, or have simply been de-prioritized.
Is Google purposely turning a blind eye to arbitrage as it creeps back in scale? Is it directly colluding with these sites to recoup its own lost revenues from the Panda update? Or to profit off its data when it cannot monetize the data effectively internally?
Right as the EU, pushed particularly by France, put the “Right To Be Forgotten” in place in May, Ask.com dramatically increased its AdWords spend in France (and the rest of the EU).
Or is this just different silos of Google chasing departmental goals (quality results and higher quarterly revenues) in ways that appear related but really are not?
It’s clear that Ask.com (and a few other players, as well) is getting special treatment; at least in getting past the policy that has kept others from participating in this kind of advertising for seven years.
Aside from this level of unfairness, it’s likely that Ask is using the platform to collect first-party data to re-sell via its ad networks and using Google’s partner revenue share to subsidize the costs of that data.
The net result is almost as if Google were selling its internal data to a third party. This certainly goes against Google’s public stance on quality and privacy as being primary principles of the company.
Ask.com has had a good run. It’s time for Google to bite the bullet and enforce its arbitrage policy fairly and across the board.