Tweak AdSense Project: Every Statistic Increasing

Just a quick update for you folks at home (or work, as it may be). I wondered exactly which statistic was on the rise. Why was I getting 2x my normal revenue? Was it page impressions, clicks, CTR, or eCPM that was the wonderful culprit?

I was wondering about it because I noticed in my other web stats that I gained a few hundred visitors recently too. Nothing out of the ordinary, but an increase none the less.

So, I went looking at the last 3 days averages, just to see the difference from the prior averages.

Page Impressions

In the past three days, my page impressions went up about 1-200 pages. That’s not insignificant, and I definitely would love more to generate traffic. But, if that’s the only reason I’m making more money, I’ll need to work on my Adsense strategy a bit more. That’s only one variable, and in my opinion, the hardest one to fix.

Raw Clicks and CTR

The number of people clicking through the ads also went up recently, surprise surprise (remember, I put a lot more ads on my pages). So, this goes along with traffic slightly. But it’s not perfectly correlated because my CTR also went up. For the first two weeks of May, I averaged only a measly 1.74%. The last 3 days have been averaging 2.12%. I’d like to dedicate this delta to Rick Schwartz, and his post on the invisible cost of losing website sales. I’m doing my part Rick.

CPM (eCPM)

My eCPM jumped like crazy in these past few days. It’s now averaging $5.23, where the first 2 weeks of May it averaged $2.94. That’s pretty significant for averages. I’ve had higher individual days (upwards of $8), but not 3 days running.

After about a week, though, I’m going to test removing the sheer volume of ads on this blog again, just to confirm the differences between eCPMs from this week to next week. I introduced an error into my tweaking methods by doing two things at once (putting more ads up and doing the competitive ad filter at the same time). That’s a debugging no-no. Bad Nate. Bad.

The boring part of stats

So, I know this is a little boring, but hey, if you’re going to be doing it, be a stats whore. Some people might say 3 days is too little to tell. There might be some whacked out anomaly black whole in the earth, Google might have changed something, sun spots, whatever. Although that’s possible, statistically speaking, 3 days is a pretty good number to test with (at least with my traffic counts). You only need about 150 subjects/impressions to statistically have confidence in the trends you’re seeing. So, with over 2200 impressions in 3 days, I think I can be statistically confident.

End of boring stats: FUN!

So, with all that jazz, what does this mean? Simply this, I’m one step closer to living in Europe, and maybe New Zealand.

If you have any suggestions as I’m on this journey, by all means, leave them in the comments. I’m all ears, and I really want to learn how to do this right.

[tags]adsense, tweaks, cpm, ctr, impressions, advertising[/tags]

Nate Ritter lives in the Pacific Northwest (U.S.), popularized the #hashtag and creates web applications for a living. He also does miles and point hacking to enable cheap travel for his family. More here →

Comments are closed.

 
More in Business, General, Projects
And the “Best Compliment Ever” Award Goes To…

[photo courtesy brainware3000 on flickr] I think I've just received the best compliment of my 30 years of life so...

Close