Forum Moderators: martinibuster
(For the purposes of this thread I'll call our CTR before Aug 15 "x" and the eCPM "y".)
Obviously such a big explosion of traffic caused some problems. Our server costs went from low $x-- to $xx,---. We had a couple days of major server issues and lost some of our previous regular visitors. Dealing with all of these issues has required that we spend less time on some of our other sites and we've lost money on them.
We also noticed that it appears we were getting smart priced. Our CTR went from x to 2x yet our eCPM went from y to 3/4y Now today our CTR is 3x and our eCPM is 1/2y. Obviously I'm not thrilled about that but at least I can make plans and react to that data. The problem is that we got a letter from Google asking us to justified the traffic increase and saying that they may need to adjust our earning lower because of poor ad conversion. We sent back an email telling them exactly the source of the traffic, which again is extremely easy for them to comfirm and 5 days later have no response.
Now I'm stuck. I'm watching costs skyrocket and eCPM drop and I can't even trust my earnings reports from Adsense at this point. How am I suppose make business decisions when I can't even be sure of how much money I'm making? If Google was to adjust my earnings wouldn't that basically be Smart Pricing twice?
There's a legit chance that I may be spending more than I'm making right now if Adsense decides to adjust my earnings. If that's the case I'd quickly take action but at this point there's just no way to know. I don't want to have to wait until Sept 25th to find out that I really only made half of what appears in my reports. However hard I way hit with the first round of Smart pricing as least it's the enemy I know. I'm finding it almost impossible to run my site when I have no idea how much of earnings Google might decide to keep.
Again this new traffic is not spammy or blackhat in anyway. In fact it's coming from a site Google owns.
BTW, according to what you wrote, your CTR doubled and then tripled? So now you must be getting maybe 60 times your previous number of clicks, but are now earning LESS than before? That does seem pretty extreme. I understand you can't be more specific, but personally I'd need more details about what your website is and how the traffic gets there before I could form an opinion.
Is it possible that you're displaying ads in a way that confuses clueless users into clicking? (E.g., through blending?) If so, making it more obvious that the ads are ads might help to reduce casual clicks and improve your conversion rate. Another option is to find a source of CPM skyscrapers or banners that pays enough to make up for the decline in AdSense eCPM. (There's no rule that says you can't have both types of ads on the same page.)
I feel like all we did is build a nice clean site that became extremely popular and now we're worried about G adjusting earnings to the point that we were better off without the traffic boost. This site was very profitable before the new traffic, now with the added expenses and time it might not even be worth it.
EFV- There's no issue with blending, the ad blocks are clearly defined. There could be an issue with accidental clicks because of a language barrior but I'm not sure how that could be researched.
As far as the banner ads, we have considered that options. But it's tough because it's contrary to the business model we developed for this site. Many Social Network Resource sites have Adsense, pop ups, pop unders, multiple banner ads and everything else. Our plan was to offer quality product on a site design totally for the user with limited advertising to enhance the user experience. We earn far less per visitor than most Social Network Resourse sites but the plan was to make up for that with greater traffic from happy visitors. Maybe we need to adjust that plan, but it was sure working well until two weeks ago.
I can think of several main issues surrounding this...
1. As mentioned in the other thread, Google may be showing ads in the wrong language. This MIGHT be a factor, but it's probably impossible to know how big this effect is without signing in a on a Portuguese-language Windows installation from a Brazilian IP address. Setting up a Portuguese-language mirror site, including a meta tag telling spiders what language each page is in, might help.
2. Even if Google can technically show ads in the correct language, it may not be able to target relevant ads properly because there is no Portuguese on the page. Again, mirroring the site might help.
3. Advertisers can control where their ads show geographically, and I'd bet most do not want their ads showing in foreign-language or Third World countries. Sorry Brazil, no offense intended. Anyway there's not much you can do about this.
4. Google may also include a discount factor for traffic from Third World countries. I'm not specifically aware they do this, in other words they might leave the whole pricing in the hands of their competitive bidding algorithm, but it seems like a plausible thing to do. It's a simple fact that, for most purposes, traffic from some regions of the world is more valuable than from others. And again, there's not much you can do about this.
So it's kind of a mixed bag. The bottom line according to me is: it's worth experimenting with things, especially if you think this traffic boom will be permanent, but if you end up with EVEN MORE Brazilian traffic but no better CPC then you may end up exacerbating your problem.
P.S. An interesting experiment would be to use PHP to do some simple geotargeting to put all your Brazilian traffic into a different channel. A first stab might be to take all traffic from the 200.*.*.* and 201.*.*.* IP blocks, and stick them in a special "South America" channel.
We were thinking along these same lines excect we went ahead and launched a full on Portuguese language site servering similar product but all Portuguese language. 24 hours later the results are not pretty. Traffic is through the roof but eCPM for the first day is lower than I thought possible given the CTR. You'd think someone out there would be able to figure out a way to monetize a ton of Brazilian traffic. So far we haven't figured out a way to do it.
And through all this, still no word from Google about adjusting our earnings. With the increased server costs and seriously increased man hours this traffic has causes, our profit margin is slim to begin with. I'd really like to know if they are going to adjust anything as soon as possible so we can plan our next move. Right now we are flying blind. How can we test anything if we can't trust the results? Maybe that low eCPM on the Portuguese version is the best we can hope for, if so I'd like to know now so we can start moving the Brazilian traffic to that site and hopefully get our original site unSmartPriced and back to making money and growing.
I just can't bring myself to block that traffic. I have to feel that at some point having that much traffic will be profiable even if it isn't so much right now. I just wish it wasn't interferring with the growth of an already very profiable site.
But the Google Adsense engineers should program the ad server to isolate a bad source of hits, blocking off ads, if necessary.
The serving structure should have at least three or four levels to determine types of ads to show if any, incl., but not limited to:
1) first-time visitors;
2) first pages visited;
3) social networking links;
4) search engine links.
Ads should then be based on the value of the leads.
Last year I had suspected Google is sensitive about traffic spikes not just wrt click fraud but also for pricing. I get a seasonal spike that yields 10x earnings over a few months on one page. But the smartpricing kicked in, so earnings could have been much higher.
Google engineering is primitive. It should not be asking you to explain spikes from social networking sites it owns--or discuss possible smartpricing.
Unfortunately there's not too much you can do just now. Google Adsense is like a secret society. G wants absolute transparency from us but is very secretive.
p/g
As for Google's being secretive, of course Google is secretive. Wouldn't you be, if you had thousands of petty and not-so-petty crooks (none of whom practice any form of "transparency") trying to cheat your network and your advertisers every day?
