So I changed the structure of my site. I took a list of information that was once on a single page, added a lot more, and cut it down into a state-by-state list. My thinking was, that doing this would create more pages to be indexed because each had unique information on it. However, my adsense ads are set to display on every page and since I've done this my page impressions have tripled which has caused my CTR to go down and now I'm earning much less per click. About half of what I was earning before I made the switch. So is google thinking that because I'm getting so many impressions with so few clicks my site isn't worth as much so they're paying less? I'm actually getting more clicks than I was before the change but because my ads are being shown 3-4 times rather than twice my CTR is down. So should I go back to a single page? It wouldn't be that hard and I could just have linked sections so people don't have to scroll through the entire list just to see their state info. Here are my stats for this month to show you the difference: Date # clicks Earnings EPC 1-Sep 7 9.67 1.38 2-Sep 13 8.87 0.68 3-Sep 16 11.47 0.72 4-Sep 26 16.77 0.65 5-Sep 12 12.98 1.08 6-Sep 19 13.92 0.73 7-Sep 39 17.86 0.46 8-Sep 39 42.05 1.08 9-Sep 66 47.79 0.72 10-Sep 66 49.41 0.75 11-Sep 65 34.84 0.54 I changed the layout on the 9th. Thanks for your thoughts.
I don't think a 2 day period is a significant sample size to judge whether or not the layout change is to blame for your revenue loss. I would certainly leave the site in it's current state for another 5-7 days minimum, this way you can establish whether or not your overall earnings remain affected since the layout change over a longer time frame.
Ha! You're right. Thanks for the reality check. I'm just impatient because I came so close to breaking $50 yesterday and today I'll be lucky if I break $40. But, in the grand scheme of things that's pretty darn good so I shouldn't complain. I'll give it until next weekend and if the EPC hasn't gone back up I'll switch things up a bit. Thanks.
I would give it a while longer, because you only added the new pages a few days ago, Google's bot they use for adsense may not have crawled all the pages yet, therefore may not be displaying ads as relevant (and high paying) as the older page. It may also be that because there is a lot more text now, there may be ads showing up for other keywords that do not pay as well. You could try ad sectioning (Section targeting) if you don't already, that may help.. Again it will take a few days before Section targeting starts to make any differance. Cheers James
Since we do not know which ads were being shown or even how much each of the advertisers are paying for the ads and to make things worst, we cannot influence the visitors to click the ads or even to select which ads to click, so it is pretty hard to predict anything. Just make sure that when you set up a website, you do not do it for the sole purpose of earning Adsense income. Do it because you enjoy doing what you are doing.
I agree that you should wait a bit longer. sometime because you change the layout, google could not find the right ads for your site yet. do it at least a week, if it still not improve try to change back for a week and compare again.
You have not screwed your self. These days Google CPC is quite low. The layout of the site effects the earning. Once layout changed Google bot takes time to crawl the site. You should for some time if the earnings increase its oK , if they don't you should revert to the old layout. Tehmina
I pretty much agree with PokerVillain on your sample size, Vagabondette. However I think you need even more than a week. I'm roughly in the range of earnings that you've been accumulating over the last few days, and for me it takes probably around two weeks for a sample to be close in rate stats (CTR, CPC, eCPM) to the overall. That's even though mine is probably a little more "steady" than yours as it has more clicks but less value per click. Because of that, my "high" or "low" results are not as weighted as yours would be, so in theory your results would be prone to more fluctuation (if you get a few less clicks or a few low-valued clicks, your rates could all be down a lot, or a few high-valued ones or extra clicks could bring them up a lot). If your recent rate continues it would take about two weeks to get 1000 clicks. That's still a somewhat low sample size, but it's much more meaningful than 60-some-odd. One other thing to consider is the new pages that have ads. First of all, their content might not be as valuable to advertisers as that which you had been using at first. On the other hand, they could just not have enough information indexed by Google to give you the best ads yet. Either way, since you said the impressions have tripled, that means the new pages have roughly twice the weight of the old pages. If they are giving you low rate stats, it's going to drag down the overall averages by quite a bit. Have you thought about tracking the new vs. old by channels?
Thanks josk for your feedback. I'm going to quit being neurotic and just give it some time. The ads that are showing up are coded into the header and footer so to track them by channel I'd have to create new headers/footers and with everything else I have going on I just don't have time for that. The impressions have tripled because before the navigation was site -> list and now it's site -> menu -> list. And many people look at more than one state list so I think the menu -> list step is being viewed quite often. Patience is not my best virtue but I think my time is better spent getting the new blogs up and running rather than worrying about this. Thanks again! I think I just needed some people to tell me to chill. My friends have no idea what I'm talking about when I talk about this stuff so they just look at me with a blank expression when I talk about impressions, CTR and EPC.
Give it sometime (at least a week) before you can judge on the effeciency of your procedure. I have done similar change on one of my sites and I had the desirable results. More impressions and more clicks means more money.
I try to do month-long testing if I make any changes. Why? Because advertisers vary their campaigns throughout the course of a month. Spend end of month excess, or skimp if they don't have money at the end of the month. By doing a full month, there is a _better_ chance of hitting all the dynamics. Besides, "30" is usually the minimum sample size N that should be used for anything where you are trying to prove statistical significance.