Benchmark Your CPA

When it comes to cost per action (CPA), it’s better to be like the folks from Missouri. They’re the ones who say, “Show me.”

Most CPA deals waste publishers’ time because they aren’t tested. Instead of offering a deal that has merit, people throw together an ad and a landing page, then try to convince publishers to test.

We ran a CPA deal from a prominent offline training company. Great brand, great reputation, decent ad copy. The CPA margins were excellent, and we were psyched. Because of the company’s size, we allowed it to test on our network, a choice we later regretted.

When we were ready to go, here’s we learned about the real CPA deal:

• The company sent us a 92K HTML email graphic that resembled a huge print ad. After users opened the email and waited minutes on a dial-up connection, they saw a big picture of a woman’s head… I’m talking huge. If you want a woman with a big head, this is your ad. 
• The call to action was buried below the fold, and the size was ridiculous. The company had conducted campaigns on the Internet before, but it is primarily skilled in offline marketing. 
• The landing page was cluttered with too many choices and was too big. Big graphics, fancy pictures — great for the gang who might be pushed from a TV ad, but a poor choice for Web-driven traffic. 
• Its tracking was done through WebTrends, amateurish for a company of this size and unreliable.

It took two weeks of meetings and phone calls to clear up all the problems. In the end, we made some money, but the company wasted our time. What should have taken days took weeks.

Remember: Publishers pay with their time. Most CPA deals make the publisher do the dirty work.

Does this scenario sound familiar? It should, because it happens everywhere on the Net. CPA deals are notorious for wasting time. It doesn’t have to be that way.

Good advertising works when it has been tested. Online advertisers make critical mistakes by sending out untested creative. They waste the time of the publisher and, ultimately, the customers we are all targeting.

Bad CPA programs undermine the results of other CPA programs. Bad advertising weighs down the whole industry.

Which is why you need to benchmark your CPA. Prove to yourself that your offer works before convincing others to test your ad.

Benchmarking CPA deals is more common offline than online. Companies will spend a bit of money up front to prove an ad is pulling the right response and ultimately generating orders.

This practice is much easier online, yet few do it. Here’s how you can set up a simple CPA benchmark to prove the value of your offer. Before beginning, keep in mind that all the various tracking systems on the Internet confuse the issue of benchmarking. If you compare your numbers to someone else’s, you will almost always be off by 10 to 15 percent.

In my network, we benchmark deals with specific rules. Pop-unders are served only once daily to users so we don’t annoy them. Therefore, we benchmark how many unique visitors see our ad, how many click, and how many convert.

This is a much smaller number than we would get if we bought pop-unders by total impressions, with users seeing the ad multiple times a day. Yet this number — and rule — works for this specific offer. But comparing my numbers to ones generated by someone using DoubleClick (which serves multiple pops per user), we’ll show entirely different results.

Understand that as an advertiser, your scenarios are based on certain assumptions. It is up to you to prove those assumptions. Here’s how:

1. Set up your CPA campaign creative, and test it on a variety of Web sites or email lists by paying for them. That’s right, pay to test your CPA. Why should publishers pay to test your ad? 
2. Buy ads in small sections of sites, or buy clicks from Overture to test your offer. It is amazing how much testing you can do for $500 — money well spent if you are serious about your campaign. 
3. Test the conversion of your offer and your landing page. Developing good emails and banners is not as hard as developing a great page that converts. 
4. Keep your landing page focused on your specific offer and aim for a 2 percent conversion. That is a lofty goal, and you’ll likely get closer to 1 percent, but if you can prove a fairly consistent 2 percent conversion, you have a winner. 
5. If you can prove an adequate response, make sure you note the specific conditions you need that ad to run under. You may be targeting ads by geographic location — make sure you include that as part of the written request for your CPA deals. Let the publisher understand how your ad works. Don’t leave this to chance. 
6. Establish open rates, click-throughs, conversions, and other applicable data you need to measure success. Once you get a test scenario that seems to work, go to CPA publishers with a proven deal. 
7. Take the time and effort to develop a good sales process that is tested. When you call publishers with your next CPA deal, they will value you as someone who is prepared to back up his bluster and numbers with tested results.

Cynical marketers who like to poke holes in any solution will say benchmarking is impossible, comparing sites is crazy, and the only real test is in the delivery of your ad to that audience.

That may work for you, but it doesn’t work for the publisher. Benchmark your CPA deals, and help us all save time and money.

Show me a deal that can work, and I’ll act on it. Show me a deal you haven’t tested, and you won’t get another chance.

Publishers generally make more money from CPA deals that work than from cost-per-thousand (CPM) deals. But it’s up to you to prove it — to show me, the publisher, the money.