Adsense mystery

I started a magazine-type website in the beginning of January. I had adsense on it, getting a cool 2% CTR on desktop and 1.6% on tablets.

Then one day, overnight, something strange happened: CTR on desktop tanked to 0.5% where it still is to this day. This was two weeks ago. Happened overnight. My only explanation was that a seemingly minor change introduced a bug that crashed lots of desktop browsers. However if this were the right explanation, then the average number of pages per session should have tanked at the same time, which didn't happen.

So at my completely lost. Any suggestions on why this might've happened and how to fix it?

Did you add new ads, or change the placement of the old ones? You'd be surprised how simply adding something else audiences want to click near an ad will result in far less ad clicks. Figure out every change you've made since then, and you might figure out why. Also adding poorly placed additional adsense ads will bring your overall ctr down.

back2basics said:

Did you add new ads, or change the placement of the old ones? You'd be surprised how simply adding something else audiences want to click near an ad will result in far less ad clicks. Figure out every change you've made since then, and you might figure out why. Also adding poorly placed additional adsense ads will bring your overall ctr down.

Click to expand...

I concur with back2basics. I usually experiment with my ad placements on my websites, and during one experiment where I added an extra link ad per page (a link ad unit) and moved 1 of the ad units to the sidebar, my CTR dipped down to 0.2%! However, after I tweaked the ad placements a bit more and removed the link ad unit, my CTR jumped up to 0.98%.

Just run tests with the ad placements and start using the placement that results in the highest CTR for your website.

Change your theme or optimize your files as their may be a bug within the code you are choosing to utilize.

Once you do work it out I'd suggest only making one measured change per week in the future, any fluctuation is then easily remedied. Good luck

Damn it. I know you guys must be right. I religiously keep track of every single tweak I make to my website and I've tried reversing them all. Unfortunately, so far no luck. I came here looking to see if someone would point out something I have overlooked.

What you guys pointed out is most likely the correct explanation, I just haven't been able to pinpoint what the critical tweak was that changed everything...

Oh well, back to trying.

ankame said:

Once you do work it out I'd suggest only making one measured change per week in the future, any fluctuation is then easily remedied. Good luck

Click to expand...

Why one per week? I've been trying one per two days. I would even do one per day, but the reports don't all come from the same timezone, so I effectively need two days to guarantee that all reports refer to the same change.

oneloop said:

Why one per week? I've been trying one per two days. I would even do one per day, but the reports don't all come from the same timezone, so I effectively need two days to guarantee that all reports refer to the same change.

Click to expand...

Just so you have time to take in every little fluctuation, bit by bit. You need to be able to highlight the magic formula and doing too much over a shorter timeframe means you'll still be guessing a bit when a change (good or bad) happens