Keyword Elite AdSense Arbitrage Experiment Post-Mortem (Continued)

Today I'm continuing the AdSense arbitrage post-mortem by doing some detailed keyword analysis. The goal is to see what we can do to salvage the arbitrage campaign and turn it profitable.

Yesterday, though, Thunderbird crapped out on my email inbox, which was approaching 2.2GB in size. I'm a bit of a packrat and don't always delete the emails I should. So I've been without my email for a day or so as I've been trying to repair things. I've got the latest version of Thunderbird installed and have managed to compact things somewhat and free up enough disk space to get it working… unfortunately it's also re-downloading the last 5000 messages that were sent to me and it's not the first time this has happened during this recovery process… I'm just mentioning this to point out that if you've sent me a message and I haven't responded yet, it's probably because I haven't read it, though I've been able to read and respond to some from my BlackBerry. And if someone knows of a Thunderbird extension to delete duplicate emails, please let me know, I don't relish going through it all by hand!

Bid Gap Analysis

I've mentioned bid gap analysis several times before in this blog. I also have a few pages devoted to it in Uncommon AdSense. Basically it involves trying to figure out how quickly advertiser bids for a specific keyword drop off from the maximum cost-per-click value that almost every keyword tool can give you. Because the top value can be very deceiving. My favorite tool for bid gap analysis is AdSenseAccelerator, but you can do the same thing by hand by playing with the AdWords traffic estimator. It's just a royal pain in the butt to do it that way, and you certainly don't want to do it for many keywords.

Out of curiosity, I analyzed the keywords with the highest density on my content page: spyware, spyware removal, spyware remover and spyware removers. I did this for all three AdWords matching types. Here's what it looks like:

In case you're wondering, the data is sorted at the Bid 4 column, because I think the first four bids are the most important and can determine whether it makes sense to place one or two ad units on a page.

Now, before we go any further, let's sit back and remember one thing: these are AdWords bid prices for the search network that we're looking at. Not the content network. So the values you see here aren't representative of how much you'll make per click via AdSense — those numbers will be smaller, and not all advertisers will bid for placement on the content network. So no matter how much analysis you do, there's a lot of educated guesswork at play here. You'll only know if you're right when you start applying the results of the analysis.

Also, keep in mind that these values are snapshot in time of bid values. Bid values can change drastically over several days as new advertisers come into the market and others leave, etc.

Look at the difference between the Bid 3 and Bid 4 values. All the Bid 3 values are $2.72 or above — all but one are above $3. But the Bid 4 values go as low as $0.54 and no higher than $2.52.

Looking further down the bid ladder, notice how quickly the generic “spyware” term gets down below 10 cents.

What can we conclude from this? Probably that we should have no more than 4 ads on a page — and maybe even only one or two — to have a chance at making decent per-click values. In fact, two of the keywords we're analyzing — spyware and spyware remover — are in the AdWords ad group that is sending traffic to the site. Here are the relevant stats from my AdWords account:

With almost 40,000 impressions, the average rank of my ads was 4.8 and my average cost per click was 5 cents. So this data confirms that the 4th and subsequent bids aren't paying much at all, since the 5th spot is only costing me $0.05 as an advertiser.

As I mentioned last time, there are two things I need to do to resurrect this campaign: increase my advertising clickthrough rate (to get more traffic to the site) and increase my per-click earnings (to make money above the costs of the advertising). I should focus on the second aspect — increasing my earnings — before tackling the first one. So what I'm going to do first is change the ads on the page to take into account what we've just learned.

Stay tuned!

Sponsored Link: Learn advanced AdSense tips with Eric Giguere's Uncommon AdSense.

Eric Giguere wrote Make Easy Money with Google and Uncommon AdSense. His goal is to get his AdSense blog into Matt Cutts' blogroll.

Comments