While I was in Miami at Affiliate Summit this July did I had some discussion via email with Andreas Reiffen from Marketing2null.de, about a post he did at his blog (in German) that looks into the pro and cons of changing Google’s paid search model from pay-per-click to CPA, also known as cost-per-action.
It caught my interest for several reasons.
Background
I am an affiliate marketer, who uses paid search among other marketing methods and I am a blogger who was also writing about the subject earlier. In my June post about Google’s worldwide expansion of their Pay-Per-Action beta program, did I elaborate a bit on the performance based CPA compensation method, which is old school affiliate marketing and about the potential opportunities, Google has with their program.
I also debunked some myths about the real or felt threat the Google Pay-Per-Action might or might not cause for the affiliate marketing industry. Andreas Reiffen, who is also a German like me is not a newbie about the subject Google and performance marketing. We wrote (in English) the study for Clicks2Customers / Incubeta last year, called “Profit Sharing: The Future Business Model in performance-based Search Marketing”.
He just did recently translate the German post from July into English and posted it on his new blog at Marketing2null.de.
Summary
I do agree with the majority of statements made in the post, as I did with the German post already in July. As with the German version do I also disagree on a few assessments and conclusions drawn from the analysis and prognosis of the potential impact of CPA as the preferred revenue model for Google’s AdWords paid search application.
Solving the Click Fraud Problem
I agree that the performance based pricing model will solve the problem with click-fraud that remains an unresolved problem in paid search to this day. The problems are the same as the problems affiliate marketing faced 5-8 years ago when the pay-per-click commission structure was at its height to fade away due to exactly the problems with abuse as search engines are facing today. It is the same thing. Only on a much larger scale, since Google revolutionized the model in 2003 when they introduced Google AdSense. Although Googles mighty technology does provide Google with an advantage over the affiliate advertisers and networks of the pre-dot-com crash era will all the technology in the world, not be able to solve the inherit fraud problem of pay-per-click finally.
Shift of Risk / Risk Distribution
Advertisers paying instead for eyeballs or traffic, paying only if a user performs a desired action, which the advertiser specified, shifts the risk almost entirely to the publisher who promotes the advertisers offer. This includes the search engine as a publisher just like any other publisher of the search engines content network.
Risk of Disintermediation
Andreas concludes that this shift in risk will motivate more advertisers to do their search marketing in-house rather than outsource it to a specialized search-engine-marketing agency and/or to their affiliates. I believe though that this would affect search-engine-marketing agencies much more than it would affiliate marketers. I do not believe that the much more advertisers would take the job in-house rather than outsource with CPA as pricing model.
The exception are advertisers who try it in-house first and “get burned” today with the PPC pricing model, if they don’t know what they are doing. I do not know how much percent of advertisers today end up outsourcing search because of a failed in-house attempt. I will revise my statement, if the number is much higher than I think it is.
Impact on SEM Agencies
SEM agencies are usually paid based on advertising spend and not per performance. They bear today virtually none of the risk for the advertiser. That changes with CPA. I am sure that they prefer keeping it that way. A shift to CPA based prices would have significant impact on SEM agencies.
Impact on Affiliates
For affiliates would not change much and it might even be a good thing from their point of view.
It is a challenge for affiliates today to track their campaign effectiveness based on the provided CPC metrics provided by Google AdWords and the virtually nonexistent metrics with conversion data that merchants would have to provide, but to 99.9% do not.
The affiliate’s role in the paid search game changed over the recent years. The low hanging fruits, such as bidding on well-known trademark and brand terms become scarcer and often prohibited, via the affiliate program terms and agreements between the affiliate and the advertiser.
It shifted more towards brand protection, coverage of additional slots on the same search results page and tapping into the “long tail”, covering keywords combinations the advertiser did not even think of.
The problem affiliates will face is much more technical in nature.
What can they specify as conversion goal and report-back to Google, if they do not have access to the exact conversion data on the merchant’s website? If they do not, one option would be to fall back and use CPC metrics derived from their existing tracking and reporting and specify as the action for the CPA campaign in Google, the click-through to the merchant’s website. Other alternatives would depend on the technical details and available options provided by Google.
Change in Ad Ranking Logic
The ranking algorithm for CPA-ads would be a more challenging calculation for the search engines, but doable. They could still use CPC metrics internally or alternatively EPC (Earnings per (Hundred) Clicks), which is an important metrics from the affiliate marketing space.
I believe that the risk of abuse by Advertisers to get free exposure via non-converting ads is negligible.
Low traffic keywords with no competition mean also little exposure for the advertiser. In addition, Google can always implement thresholds that will drop ads that do not convert enough (or at all). They can also do a comparison with other Ad campaigns for the same industry or across their whole network and make assumptions about the expected performance of a new Ad campaign.
I agree that it will take them a bit more time and more data to do any quality assessment of a new campaign to be able to rank ads efficiently.
Decreasing Reach and Other Problems
The barrier of entry for advertisers is higher with CPA, because changes to the advertiser’s website are necessary for CPA tracking. The current CPC model does not require any chances to the existing site, although search engines and SEM consultants recommend it. Efficient and effective paid search campaigns are virtually impossible without some sort of conversion and ROI tracking on the advertisers end.
A much bigger concern will probably be the fact that CPA requires advertisers to share conversion data with Google. See also the debate about using Google Analytics in combination with AdWords. This is today optional with CPC and opinions about it vary. There is no consensus if you should share the information with Google or not. Google would certainly like the access to that information and could possibly use it to the disadvantage of advertisers and cause the further decrease of profit margins.
Not discouraging the increase of costs for campaigns to come close to a $0 ROI would be in their favor and increase their profits.
I think that enough advertisers provide Google with this type of data to get a very good idea of this type of information, by either using the AdWords conversion tracking and/or using Google Analytics (with or without using AdWords). The shift to CPA for paid search would only extend that and allow Google to make more accurate assumptions and predictions than they are able to do today.
Conclusion
Once again, make sure that you check out Andreas Reiffens conclusions to this subject in his blog post titled: “Will Google Extend the CPA-Model to Search?”. I did not repeat a number of points he made where I completely agree with him.
I personally think that this is where the industry is eventually heading. It aids the blend of the different types of advertising together and allows easy management of all of them via a unified interface. CPA will not replace CPC or CPM, but will become another option for advertisers to choose. I also believe that CPS (Cost per Sale, revenue sharing) will eventually become an option in paid search some day in the future as well.
CPM is currently only supported in the content network by Google, but I would not be surprised, if it becomes also an option for their search-engine-result-pages (SERPS) one day. Speculations about the positive branding effect of paid search ads showing in SERPS were already done in the past, but actual data to confirm or disprove the actual effectiveness.
Each model has its purpose is the whole mix that we call advertising and marketing. A blend of all three of them, plus adding the forth option of paying a flat fee for exposure will also blend-together advertising and marketing into something that has no name yet. You might call it “Promotioning” or something like that.
Other Indicators and Further Research
If you do not believe me, look at the information that are out there and accessible to anybody who wants to see them. Display Advertising today is a mix of campaigns with 47% of the campaigns using CPA or performance based pricing and 48% with CPM based pricing, where CPA increased steadily since its introduction to display advertising as early as 1998. See the Ad-Revenue reports published by the IAB, the Interactive Advertising Bureau, on their own website.
Based on the AffStat 2007 report is the share of CPA in affiliate marketing is about 19%, compared to 81% of programs who use revenue share as compensation model.
The improvement of web analytics and tracking of metrics like cost per acquisition of a new customer and average customer lifetime value over the recent years allow merchants to explore new territories.
You can see retailers, who are classic revenue sharing model candidates, starting to explore the CPA option either via the classic affiliate-marketing channel or through other channels, such as Co-Registration (CoReg) and Lead-Generation (LeadGen).
This space is growing rapidly. Just the number of CPA networks today compared to the number of CPA networks in 2005 is incredible. Here is an incomplete list of networks on my site. I recently gave up the goal to keep it a “complete” list, because it became impossible to keep up with it.
The recent acquisitions of classic display advertising companies by search engines are another indicator for this blending process that is taking place. Another one is the overlapping of paid search with affiliate marketing via contextual or keyword targeted advertising via Google AdSense, Yahoo! Publisher Network (YPN), Microsoft Content Ads (non-public beta), eBay’s AdContext, Amazon’s Omakase Links, Miva Monetization Center and companies like BidVertiser, AuctionAds, Chitika and numerous others.
I am interested to hear what other marketers think about this subject and invite anybody who has his own thoughts about where the industry is heading in the comment box below.
Carsten Cumbrowski
Carsten is an internet marketer with over 6 years experience in the affiliate and internet marketing space and over 5 years of experience in organic and paid search advertising. One of the projects he is involved with is the internet marketing resources portal at Cumbrowski.com among others.
