Should You Buy Ads on a "Cost-Per-Action" Search Engine?
A search engine called Snap.com has been grabbing a lot of headlines recently. There are a couple of reasons for this: Snap.com is a creation of Bill Gross, a famous Silicon Valley entrepreneur whose company Idealab was one of the biggest incubators of the dot-com era. Bill also happens to be one of the people responsible for CPC ("cost-per-click") advertising, the ad serving model used by the major engines today.
But the biggest reason that Snap.com is getting lots of press has to do with its business model, which departs significantly from that of the other engines. Unlike Google, Yahoo, MSN or Ask, the company provides advertisers running paid search campaigns the option to buy ads on a "cost-per-action" basis. Which means they won't be charged unless a clicker actually performs the action desired by the advertiser (e.g. filling out a lead form, subscribing to a service or purchasing a good).
The cost-per-action model isn't exactly new (Snap began offering this option back in 2004), but it does have renewed appeal right now, given how much the problem of click fraud has been weighing on marketers' minds recently. By removing the incentive for fraudsters to rack up phony clicks (either to pump their own publishing revenue or to use invalid click costs as a weapon against competitors), cost-per-action seems to be inherently superior to CPC, which Snap's boosters have characterized as a "risky" ad model that can "click and dime marketers to death."
Unfortunately, a cost-per-action marketplace comes with its own set of risks. The biggest risk for marketers is that they will fail to realize their goals because the campaigns they run will not be as well thought-out or as well executed as they would be in a PPC-driven environment. Here's why: on Google, Yahoo and other engines running CPC auctions, a marketer is solely responsible for all the variables that make or break a given CPC campaign: factors such as quality of copy, click-through rate, landing page design, etc.
Because top positions are allocated in a competitive, performance-based environment, which rewards excellent, relevant campaigns and penalizes poorly executed ones, there is enormous pressure for marketers to squeeze out inefficiencies from their campaigns. They do this by using every tool in their toolboxes, including expanded keyword discovery, geo-targeting, day-parting and ROI-based campaign management technologies. The result is a win-win scenario for all constituents, including users, marketers and the engines themselves.
A cost-per-action ad model, however, removes this incentive from the mix. There is no reward system for excellent marketers, nor is there a penalty for poor ones. After all, what incentive is there for marketers to do a great job, when somebody else (the engine) will be footing the bill? The whole system is set up in a way that rewards (or at least, does not penalize) mediocrity and irrelevancy in results, which depresses clickthrough and conversion rates, diminishes ad revenue to the engine, and worst of all, creates a poor experience for the user.
It's far too early to declare that Snap.com or the cost-per-action model are necessarily doomed. Many broadband users will likely find Snap.com's unique "split-pane" interface, which lets users preview Web sites, appealing, so Snap.com may find itself a niche. I am much less impressed by Snap's blending of sponsored and organic results on its SERPs because I believe that doing so will confuse users and possibly erode trust in the neutrality of Snap.com's search algorithm.
I do have major doubts about whether the Web really needs yet another general-purpose search engine, but I've been surprised before and if a critical mass of users decides that Snap.com provides more relevance and a better user experience than offered by the major engines, it might have a future. Right now, Snap.com has a long way to go toward being considered a serious contender to Google, Yahoo, MSN or even Ask.com (according to Alexa.com, Snap.com's traffic is about one-eighth of fourth-ranking Ask.com).
Whether marketers decide that Snap's model provides a superior ad channel is another matter. The PPC ad model may have its flaws, but it's still the most efficient ad channel available to marketers, and these flaws will lessen as the engines continue to refine it through improved anti-fraud measures.
Should you advertise on Snap.com? Well, it's your search marketing nickel, and I'm not going to tell you how to spend it. But you should take any talk you hear about cost-per-action being a "risk free ad medium" with a major grain of salt.