| 7:11 pm on Aug 29, 2007 (gmt 0)|
thanks for dropping by.
|The new "Ads by Google" logo style, for example, received a very negative response at first, but improved publishers' CTRs and earnings. If we had only tested it on publishers who liked it, we wouldn't have heard from all of you |
Here's how we solved beta testing at (insert name of huge Internet provider here) a while back:
1. Build a large enough sample group from users who agree to perform tests.
2. Explain the tests to them.
3. Do the tests and work with the results from the closed user group.
The fact that the user group is "closed" does not mean that it will contain only users who "like" the product. It will mean that it contains only users who AGREE to testing and to provide honest feedback (and maybe also additional information).
There is no need to run semi-secret tests against live sites.
| 7:59 pm on Aug 29, 2007 (gmt 0)|
First, thank you martinibuster.
|we do everything with your best interests in mind |
ASA, if you're interested in specific feedback, changing the font size randomly by a factor of two or three is certainly not in our best interests. It looks displeasing and unprofessional. We need a set predictable font size so that we can be sure it always matches our site design in a tasteful way.
It may be in Google's short-term financial interests to vary the font size, but it's not in the publisher's long-term interests as far as the user experience is concerned (certainly not for us -- we want our visitors to like our site and to want to get back, and Google should want the same).
We've repeatedly asked Google to provide an option to opt out of that, to no avail. And that's just one of the reasons why we'll have to leave AdSense.
PS - I know you do listen to feedback. I was one of the people who asked for the "Gooooooooogle" word to disappear. I appreciate it did. Alas, it was the only thing we've achieved I believe. The vast majority of other presumably reasonable suggestions seem to have been, more or less, ignored. Anyway, all the best.
| 9:21 pm on Aug 29, 2007 (gmt 0)|
|The tricky thing about notifying publishers ahead of time about tests |
Hold on, why are you running tests on peoples live sites without informing them. As people have said before conduct your tests with closed, willing beta testers. I'm sure you can find one participant very easily in this thread.
Work out what works for both the publisher and Google and then give written 30 day notice to publishers detailing the changes via email and in their stats area (there is a space for optimisation tips, so I can't see this being a problem) so they can decide whether to continue to allow their sites to be represented by Google.
| 9:25 pm on Aug 29, 2007 (gmt 0)|
Clicking on one of those extra little links brings you to a "search results" type of page so what exactly is this?
I HIGHLY doubt adwords users got any warning. The ads that show up when you follow through on one of those added new links ARE NEITHER SEARCH NOR CONTEXTUAL since nobody who clicks on them actually searched for the term (hence its not a true search results page) and they aren't in page context either (since theyre no longer on the same site even).
I think Google broke some rules on this one.
| 1:24 am on Aug 30, 2007 (gmt 0)|
I have no trouble giving an impression here and there for testing ways to improve CTR and revenue.
In fact, I'd suggest Google going a bit more creative in their ways to test new approaches and doing more of these tests.
Also, I'd be even more for these Google tests if they were for ways to improve direct, one-click CTR and revenue, instead of adding adlinks to the traditional adsense box.
What is a bit troubling for this particular ongoing test is that it has been going on for a good two weeks now, and Google should have enough data to make initial conclusions, given that there are so many publishers to provide to the sample size.
| 1:43 am on Aug 30, 2007 (gmt 0)|
sorry guys, i find most of your reactions a bit diffuse and immoderate. but i understand the general issue.
since adsense participants only seem to notice the new ad models one after another in a remarkably stretched period (look at the date of the original post), let's resume once again what google is testing:
1) "show more" link in the bottom corner of an ad block. when clicked, leads to a result page of theme related google ads. just as ad links.
2) several keyword links in the bottom corner of an ad block. when clicked, reload a new ad inside the iframe with keyword related ads.
whereas i find the keyword links (2) highly innovative, the issue i and others may have with the "show more" (1) link is, that it's nothing else than ad links.
many webmasters have deliberately not implemented google ad link boxes, because they think it's a bad user experience when a user is directed to a wall of ads with no content. some people (including me) argue, that this is nothing else than google profiting from an own legalized mfa practice, since these links direct to pages consisting of nothing but ads.
so a few of the guinea pigs have strong reservations against this form of advertising. some publishers (including me) waive ad links in favor of good user experience even if they would bring in some extra money. so googles' argument, that they only experiment for our benefit to optimize earnings doesn't apply in this case.
| 2:01 am on Aug 30, 2007 (gmt 0)|
|1. Build a large enough sample group from users who agree to perform tests. |
You've already biased the same when asking for volunteers. Like it or not, Google is doing the right thing and has our best interests in mind.
| 3:04 am on Aug 30, 2007 (gmt 0)|
"Like it or not, Google is doing the right thing and has our best interests in mind."
Yeah, because it couldn't just be PR bull#*$! from ASA, right? Google representative says that it's for our good, so it must be true! Why, of course!
Personally, I find the pure concept of AdSense great. However, it is extremely poorly executed. Sadly, there is no real alternative, so NO, I cannot "switch to something else". OK?
I want MUCH more control in AdSense. But Google is not interested in anything but maximizing profits. It's so transparent I'm almost ashamed of having to tell you this. I don't know about you, but MY best isn't the maximum amount of money, but to earn as much money as possible WHILE sticking to my highly set ideals. These are NOT compatible with the things Google are "testing" on us without asking first.
The only way I can have my way is to make my own solution. This is impossible because I could never make deals with all these little companies. It's impossible, so I'm stuck with AdSense and at Google's mercy.
We need some sort of open source AdSense alternative.
| 5:48 am on Aug 30, 2007 (gmt 0)|
|You've already biased the same when asking for volunteers. |
Just send a friendly email to 10,000 randomly selected Adsense publishers asking whether they would be willing to let Google run tests against their live sites, explaining in a very sober way the scope of the tests ("new content/layout within existing ad blocks"). Then collect the data required and ask for additional feedback from the sample group.
The sample group will be probably be just biased with regards to the desire to improve the product (Adsense). This, however, is not what I call "biased".
You get a "biased sample group" when you restrict the admission to the group by knowingly applying a filter through pre-selection (e.g. only experienced Adsense sites, only big sites, only US sites, only sites dealing with "widgets", only MFA sites, only parked sites) as such clusters are in itself biased and may have different objectives than the complete user group.
|Like it or not, Google is doing the right thing and has our best interests in mind. |
I highly doubt that.
| 7:41 am on Aug 30, 2007 (gmt 0)|
<removed> has had a carte blanche adsense account for over a year now. You can see google ads that look like forum posts on his main site. The ones on his personal blog (top right corner) don't even look like ads at all since the "ads by gooooogle" isn't there.
Let them keep testing their stuff with guys like <removed> who run 30 sites of tests. The majority of his google ads, and he has them everywhere, aren't available to us all.
Everywhere you look on his entire network of PR 5 and PR 6 sites (even though some are 100% amazon affiliate?!?) is google test ads.
I personaly don't want to pay extra for the testing and dislike not having a choice.
[edited by: martinibuster at 9:54 pm (utc) on Aug. 30, 2007]
[edit reason] No Outing of specific sites, please. [/edit]
| 7:50 am on Aug 30, 2007 (gmt 0)|
Actually you know what? It bugs me a little that <removed> is even still in the program for testing new adsense stuff. He's one of the biggest link sellers out there, I respect the effort he's put into his network but many of the sites have 30+ links on every page, using multiple buy/sell link services with nofollow.
Someone at adsense needs to level the playing field, it might be that <removed> is only 17 and hence not a legal threat or whatever... who knows.
I propose that if our personal adwords campaigns are used as guinee pigs at any point in time that we also get to test adsense in return at that point. It would be fair.
[edited by: martinibuster at 9:55 pm (utc) on Aug. 30, 2007]
[edit reason] Removed specifics. [/edit]
| 8:24 am on Aug 30, 2007 (gmt 0)|
|Personally, I find the pure concept of AdSense great. However, it is extremely poorly executed. |
No offense but that's a real stretch of reality when AdSense is in fact the leading contextual network. That aside I think that google could go about this in such a way as to avoid asking permission at all. Why not have the option of picking say 10 different types of "ads by Goooooogle" or "ads by Goooooooooooooooooooooooooooooooooooooooooooooogle" when setting up the adsense code. We'd all inevitably choose different ones which they could subsequently test by our choice of using it. Forward that opt in idea to the new 728x90 that now randomly has a link unit built into it. Then after time when google concludes (with cold hard stats) that "ads by Gooooooooooooooooooooooooooooooooooooooooooooooogle" doesn't work as well as "ads by Google" then they could remove it from being an option and notify webmasters still using "ads by Goooooooooooooooooooooooooooooooogle" that stats have proven it reduces their CTR by X% on average. They wouldn't even have to send out emails, just have one of those big banners with the new info when you login to your account next, any half brained webmaster would seriously consider regenerating old code (or better yet checkboxes in the account that don't require ad regeneration) to have the new and proven "ads by Google" logo instead of "ads by Goooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooogle." Though I have to admit by doing it the way I've mentioned they would greatly reduce their ability to adequately test their ideas as they do now.
Having said that I personally haven't minded too much the way they've been testing the adsense layout and the google logo and such over the years, just like anything else Google needs a solution that fits the masses like the dumbed down version of adwords you get now when you open a new account, or a more complex version of adwords for more advanced users.
One last thing...I would be mighty big of google to give us referral credit when some one clicks on the "ads by google" logo and ends up signing up for an adwords or adsense account (which is where they end up nowadays when they click on it.) I can only wonder how many thousands of clicks I've dished to that "ads by google" logo by now.
| 2:53 pm on Sep 3, 2007 (gmt 0)|
I just noticed the six additional text links under our single 728 x 90 block. We never run more than one Adsense block per page, nor do we run any other third party advertising.
In my personal opinion, the text links look spammy. I don't know whether they will help or hurt our overall earnings, the point is we didn't sign up for them.
| 2:59 pm on Sep 3, 2007 (gmt 0)|
Are the additional links within the 728 x 90 space or outside of it? To me, that's the real issue.
| 3:20 pm on Sep 3, 2007 (gmt 0)|
|Yeah, because it couldn't just be PR bull#*$! from ASA, right? Google representative says that it's for our good, so it must be true! Why, of course! |
I'm not big into consipiracy theories.
|You get a "biased sample group" when you restrict the admission to the group by knowingly applying a filter through pre-selection (e.g. only experienced Adsense sites, only big sites, only US sites, only sites dealing with "widgets", only MFA sites, only parked sites) as such clusters are in itself biased and may have different objectives than the complete user group. |
No, what some might call a stratified random sample.
Look - can Adsense do a better job? Sure, this is a great forum for feedback. But let's not kill the messenger. There is no good reason for Google to purposefully alienate its partners.
| 12:18 am on Sep 4, 2007 (gmt 0)|
|Are the additional links within the 728 x 90 space or outside of it? To me, that's the real issue |
Didn't think of checking. Pasted a screen shot into Paint, clipped off the white space, and it came out 728x88. So it's within the space, but I don't agree that size matters. That's like saying it would be OK for them to show image ads on our site, when we are signed up for text only. If we wanted Adlinks, we could add them ourselves.
| 3:29 am on Sep 4, 2007 (gmt 0)|
|So it's within the space, but I don't agree that size matters. That's like saying it would be OK for them to show image ads on our site, when we are signed up for text only. |
I think it's probably a grey area, but I wouldn't compare it to using an image ad when a site has opted for text ads only. I suspect that, if link units within ad units were deployed for real, and not just in a beta test, publishers would be able to opt out through a checkbox or by contacting AdSense Support (as publishers can with site-targeted CPM ads, which some publishers find objectionable because of headline size or other reasons).
| 5:49 pm on Sep 4, 2007 (gmt 0)|
|Are the additional links within the 728 x 90 space or outside of it? To me, that's the real issue. |
| 5:54 am on Sep 9, 2007 (gmt 0)|
I think there is a huge difference between modifying the ad format and completely changing it.
| This 49 message thread spans 2 pages: < < 49 ( 1  ) |