Afraid of Click Fraud? Try Pay-For-Call
Back when the internet’s structure was created, it was set up with protocols focused upon linking distributed info together and making info easily accessible. Growth happened since then (understatement of the year), and people invented a lot of new systems on top of that existing structure which were never anticipated in the original internet design. Since the original system was built without these new paradigms in mind, all sorts of problems and weaknesses have arisen which were not addressed in those original architecture standards. Security, identity of users and site owners, traceability, authenticity — all these aspects have been cited as we have tried to control and limit things like spam, hacking, fraud, denial-of-service attacks, etc.
As online advertising evolved on the internet, and moved from a pay-for-impressions model to something closer to pay-for-performance models, the backbone internet architecture didn’t evolve in sync with it. So, all the attendant weaknesses or limitations are also impacting the online advertising industry.
So, where does that leave us with our PPC ads? Is there a solution for Click Fraud? Read on, and I’ll explain…
In the internet system’s process definitions, a “click” is not always proveably when a human moves a pointer over a link and clicks their mouse key on it. (I almost feel like I’m exposing some sort of industrial secret when I say that!) From the system perspective, that “click” is just a resource request, much like any request for a page or page content element (like an browser requests an image file from a server in order to display it on the page you’re looking at), and the system (and those of us hosting the system which receives the “click”) are essentially blind as to whether that request was made by a human or by a machine.
For Pay-Per-Click (”PPC”) ads, this means that the industry is sharply limited as to how it can tell whether a click should be counted as valid or fraudulent. The industry must use a whole lot of methods for estimating the trustworthiness of the clicks received — using things like user profiling, pattern recognition algorithms, heuristics, network traceroutes, and more. Yet, due to the current limits of the legacy internet backbone architecture, we can rarely know for certain if a “click” is truly a request initiated by a human or not. Any given click could actually be from a computer program written to initiate lots of requests in order to drive up PPC costs.
If you don’t have a big, mean competitor trying to drive up your advertising costs by sending your campaigns automated clicks, you still have reasons to fear fraudulent clicks. There are plenty of others in this giant Clue game who have Motive aside from the Butler. If your ad is running out on a PPC network of sites, it may be displayed on various pages which have some sort of thematic relation to your business or ad. Webmasters put up those ads so that they can receive a share of the revenue. Those guys have a definite interest in having more clicks happen through their sites, and some of them are clever enough to get around all the detection systems. Further, there could be an unethical ad agency or SEM agency which focuses on a particular market that your business is within, and perhaps they get paid by some of their clients based on some percentage of the total ad spends they represent. If they could get apparent click traffic to increase to ads running on that market vertical’s keywords, it could drive up the cost-per-clicks, giving them greater fees.
There’s also just randomly malicious individuals who could deploy code which spends up your campaign money, and even bots or data mining software which goes mad and accidentally deploys clicks against PPC ads.
Beyond the technical exploits and limits in detecting them, advertisers have also grown concerned about the user’s intent when they click on the ad. Is the user a bona fide, potential customer? Or, are they just curious about the ad or business when they click, and have no real intention of making a purchase? Or, worse, is the user malicious — are they perhaps your competitor just seeking to cost you dollars with each click? Even if the click comes from a human, it doesn’t guarantee that this action is going to relate directly to a conversion.
So, what’s the solution?
Assuming you can’t fix human behavior and assuming we won’t solve the endemic issues of the internet structure anytime soon, maybe it would save you some money in the near term to use Pay-For-Call (PFC) advertising instead of PPC.
PFC ads charge you when a user initiates a phone call to the number found in the ad. These online ads can be delivered based on keyword targeting, just like PPC ads.
Think about it: if you’re only paying for each call referred to you from an ad, I’d say that by nature the people calling you have to be more than just casually curious — they want to buy. Also, it’s pretty hard for a machine to immitate a human by making a phone call, so the percentage of fraud has to drop.
You’d only be paying for a real, interested human to be referred to your business — which is all you really wanted in the first place.
Quite a few of the leading internet advertising networks have introduced PFC advertising, so look them up and give it a try.
The cost for taking an order may increase a little over any online order transaction, but a good human call agent can sell to someone more effectively than any web landing page, so it may be worth it. You could introduce this type of advertising along with your regular PPC advertising and compare the conversion rates. You may find you want to ditch the PPC campaigns, or you may decide to use a mix of both types of ads in order to hit different demographics.
Popularity: 3% [?]
Possible Related Posts
Posted by Chris Silver Smith of Netconcepts on 03/31/2006 | Permalink | 
Email
| 
Print
Trackback | | Comments RSS | Filed under: Monetization of Search, Paid Search
i agree almost entirely.
thats why i built one of the first (if not “the first”) pay-for-call business models with various customizable tweaks.
i’m very flexible, so contact me if you’d like details.
Comment by lynn — 5/21/2007 @ 3:00 pm