| This 41 message thread spans 2 pages: 41 (  2 ) > > || |
|On hold Garbage|
To expand on what is being said
First of all, I want to say I understand that "Search Total" clicks do not all come from Google, and only Google CTR matters for determining status. I also want to reiterate that Google states it only uses the 0.5% minimum for the top spot, and all subsequent spots are at an even lower threshold for being "Normal" ads.
That being said, here is just a segment of keywords I have that have been put "On Hold" just today. These statistics span the past 7 days combined.
Clck Imp CTR Pos
53 2,842 1.8% 3.7
45 1,595 2.8% 1.7
15 1,028 1.4% 2
14 1,407 0.9% 1.2
13 421 3.0% 2.6
12 938 1.2% 1.3
6 643 0.9% 1.1
6 432 1.3% 3.1
6 182 3.2% 6.1
5 537 0.9% 1.7
5 143 3.4% 3.3
5 47 10.6% 4.7
4 269 1.4% 2.9
4 158 2.5% 3.4
4 109 3.6% 4.9
4 97 4.1% 4.1
4 23 17.3% 8.4
3 199 1.5% 12.3
3 65 4.6% 11.8
2 106 1.8% 4.3
2 79 2.5% 5.5
2 69 2.8% 2.7
2 49 4.0% 4.9
2 49 4.0% 6.4
2 34 5.8% 2
2 11 18.1% 2.5
1 69 1.4% 5.6
1 67 1.4% 3.3
1 39 2.5% 12.2
1 23 4.3% 9.4
1 14 7.1% 2.3
1 13 7.6% 4.4
1 11 9.0% 8.8
1 9 11.1% 3.8
1 9 11.1% 4
1 8 12.5% 4.3
1 8 12.5% 8.2
I find it very hard to believe that ALL of these keywords (and more in my account) fail to meet Google's minimum CTR requirements.
Now some observations. Most of these keywords are niche keywords. They receive relatively few impressions per day (most <100), but they are extremely targeted, mostly 2-3 word matches. These keywords are fairly new (Created Oct. 30), but many of them already average 5%+ total CTR.
It seems that almost every low impression count keyword has been placed on hold amongst 'newer' ads in my account. I have other groups that are almost identical in form to this one but have been around for months, thankfully, those seem unaffected.
Now, assuming what I've observed is of meaning, hasn't Google shot themselves in the foot? Much of the idea of adwords is to be relevant by being specific and using well honed in keywords. If they immediately kill ads for terms that receive a relatively small number of searches per day, they lose a huge overall segment of ads that are of peak relevance.
I remember Shak mentioning a possible sandbox effect for adwords. With christmas around the corner, I was excited and wanting to ramp up adwords usage this season. Now this apparent rule change on new ads gets thrown in our faces. I'm afraid now the season will be filled with frustration and heart-ache. And all for what... if anything I thought google should be trying to INCREASE ad display this season, not killing them off without a chance at all. Sometimes 1 impression and poof -- the ad is gone :(
same here. google just wants us to increase bids. if you chance bids from 0.05 to 0.06, on-hold is gone.
All of the terms I listed above are still on hold in my account, despite performing well at "Moderate" for the 2 or so weeks prior to yesterday. Changing my CPC did nothing, and it should not be neccessary anyway as the terms were already performing well. Something is seriously wrong with the new ranking system. I still think it has something to do with low impression count terms.
Reading your other post on 7.5% CTR makes me more certain that this is now a widespread problem. It may take others a longer time to notice.
BTW - My account is over 2 1/2 years old (Since almost the beginning), has spent well over 6 figures in total, and averages a healthy 2.6% CTR. These keywords should not be On Hold, it is ridiculous and frustrating.
Furthermore, look at these. These were the all time stats for some new keywords in my account, which are now ON HOLD:
Clk Imp CTR Pos
1 1 100.0% 1.0
1 1 100.0% 4.0
2 3 66.6% 2.4
1 2 50.0% 1.0
How can an ad be 1 for 1 and then be shut off. Absolutely ludicrous.
Same thing happened to me. Google's pretty much killed one of my more sucessfull campaigns.
Hey! Just wanted to follow up. After many days of these high CTR keywords being on hold, they are now all Normal status again in my account. This is a big relief, and thanks google for addressing this.
Lucky for you, many are still affected.
I must agree with you. This is really annoying. How can a well targeted key phrase with 25.0% CTR be put "on Hold". It only had 4 impressions! One of many keywords.
I guess I will just have to delete 'em and re-add them daily. What a pain.
Same here. Some keywords started out "on trial" and were disabled after 2 (two!) impressions.
We have some of those wierd statuses too. Mostly keywords that are in trial or disabled after less than 100 impressions. I think the new algo has some bugs :(
The low impression keywords that are being put on hold in my account are keywords that are used in dynamic keyword ads. It appears as though those keyword impressions are being aggregated, so that if i have 50 key phrases with 20 impressions each and no clicks, all 50 are being put on hold, even if one keword has 900 impressions and the rest had only 2-3 impressions each. I rarely use dymanic keywords , so I am curious to see if the problems people are seeing are used in dynamic ads or not.
By the way, I personally think there is merit to the idea of lumping keywords in a dynamic ad together. That might cut down on some of the "silly ads" that seem to appear.
|I must agree with you. This is really annoying. How can a well targeted key phrase with 25.0% CTR be put "on Hold". It only had 4 impressions! |
|We have some of those wierd statuses too. Mostly keywords that are in trial or disabled after less than 100 impressions. I think the new algo has some bugs :( |
Actually, a keyword with zero impressions can be put on hold. It really isn't a matter of impressions at all.
Here are some basics, that I hope will provide some clarity:
* Keywords that are certain to perform well, based on similar usage of the same keywords for the past 2 plus years, go immediately to 'normal' status.
* Keywords that don't immediately go to 'normal' will either go to 'in trial' or 'on hold'.
* An account has a limited number of spots available for 'in trial' keywords, and those spots are filled by the keywords which (while not high enough to be marked as 'normal') have the highest CTRs of the remaining pool of keywords.
* Once the 'in trial' spots are filled, the rest of the keyword list then goes to 'in trial' status.
* Please note that 'on hold' is not the same thing as 'disabled'. This is really key.
* As impressions are accrued, 'in trial' keywords will either go to 'normal' or be 'disabled', depending on how they perform.
* When an 'in trial' keyword goes to 'normal' or is 'disabled', then the 'on hold' keyword with the highest CTR moves into that spot.
* Now, here is another key point that many folks don't understand: if you add a bunch of new keywords to an account, this may cause a re-shuffling of the 'in trial' and 'on hold' keywords. This is because all keywords with 'in trial' and 'on hold' status are sorted according to CTR - and the system fills the limited number of 'in trial' spots with the keywords most likely to succeed, whether old or new, and 'pushes the other words down'.
|I guess I will just have to delete 'em and re-add them daily. What a pain. |
This really shouldn't buy you anything, madman21. With all of the above in mind, you'll see that deleting 'on hold' keywords, and then re-adding them, would just put them back into 'on hold' status. (This is assuming that you have not also deleted any 'in trial' keywords at the same time.)
|Some keywords started out "on trial" and were disabled after 2 (two!) impressions. |
It may be that these keywords are still active, but 'disabled' elsewhere in your account.
|The low impression keywords that are being put on hold in my account are keywords that are used in dynamic keyword ads. It appears as though those keyword impressions are being aggregated, so that if i have 50 key phrases with 20 impressions each and no clicks, all 50 are being put on hold, even if one keword has 900 impressions and the rest had only 2-3 impressions each. I rarely use dymanic keywords , so I am curious to see if the problems people are seeing are used in dynamic ads or not. |
jim2003, this is really not the case. If you're noticing this, it'll be purely coincidental.
I hope this helps everyone to make better sense of the new system. And it is worth noting that well chosen, specific (as opposed to general) keywords are very likely to start out as 'normal', from the word go. And that the keywords that are in 'in trial' and 'on hold' status are ones that have historicaly not performed well, for all other advertisers, over a substantial period of time.
Thanks AWA, that was probably the most definitive post on this subject so far.
Ok, so if all of this is true and google automatically trials keywords and then either accepts or rejects them, then why the need for account slowing? Does it simply give you a chance to refine your ad before keywords get disabled? It seems like more of a pain than anything.
When does a keyword move into "Normal" status for good? Is it after 1000 impressions?
Also, when an account gets slowed, does it only slow the at risk keywords or does it slow all keywords? Every time my account has been slowed so far it seems like I don't get anymore clicks at all until I reset the account.
Perhaps if google published their algorithm it may help us all to understand how the system works.
|* As impressions are accrued, 'in trial' keywords will either go to 'normal' or be 'disabled', depending on how they perform. |
* When an 'in trial' keyword goes to 'normal' or is 'disabled', then the 'on hold' keyword with the highest CTR moves into that spot.
What is unclear to me in this otherwise very clear explanation is how much impressions are used to determine whether an 'in trial' word goes to either normal or disabled.
I especially find the link to "how words did for other advertisers influences how it will work for you" a tricky one. One can only determine that based on a decent amount of impressions and this will be very different from one ad text to another.
I understand the concept as a way to queue up trial words and give them some subsequent "air time" but think that the old system was much easier to grasp and manage. This brings up the overall management cost of a campaign.
My 2 cts
|Thanks AWA, that was probably the most definitive post on this subject so far. |
Thanks, gtrdude - glad it seems to have cleared up a few things. And welcome to WebmasterWorld, by the way!
|Ok, so if all of this is true and google automatically trials keywords and then either accepts or rejects them, then why the need for account slowing? Does it simply give you a chance to refine your ad before keywords get disabled? It seems like more of a pain than anything. |
The concept behind what AdWords does in this arena, including with 'Smarter Keyword Evaluation', has to do with 'Ads Quality'. By this I mean the quality of ads that are delivered to Google's users, as judged by those users, with the unit of measure being CTR.
Account slowing continues as a reminder to advertisers that, in order to improve ads quality for Google's users (also known as your customers), the advertiser will need to re-visit their keywords/ads to improve targeting, and perhaps get rid of very general keywords.
|When does a keyword move into "Normal" status for good? Is it after 1000 impressions? |
I guess the truest answer to this is that a keyword is never really guaranteed to retain 'normal' status. If it performs on the edge of 'normal' for an extended period of time, and then has a really bad stretch, it is possible that it could move to 'disabled'. By the way, this is the way it's always been: in the past, 'strong' keywords could move from 'strong' to 'moderate' to 'at risk', etc, entirely dependent on performance.
|Also, when an account gets slowed, does it only slow the at risk keywords or does it slow all keywords? Every time my account has been slowed so far it seems like I don't get anymore clicks at all until I reset the account. |
In your quote, you mention 'at risk' keywords, and these don't really exist anymore. I'll bet that you meant 'in trial' keywords, right? In any case, it is only 'in trial' keywords that are slowed. 'Normal' keywords continue to run as normal, and 'on hold' keywords were not running yet anyway.
|What is unclear to me in this otherwise very clear explanation is how much impressions are used to determine whether an 'in trial' word goes to either normal or disabled. |
This is a really excellent question, Robsp, but one for which a 'single, always correct' answer does not exist. It is not based on a fixed number of impressions at all.
The true answer is that 'in trial' keywords go to 'normal' or become 'disabled' when the system is entirely sure that CTR will not meet the minimum performance standard. One thing that many folks have not realized is that the new system may well give the keyword a lot more impressions than the previously used standard of 1000. Basically, the keyword keeps running, until the system is sure, one way or the other. This could be after many thousands of impressions, or it may be fewer.
|I especially find the link to "how words did for other advertisers influences how it will work for you" a tricky one. One can only determine that based on a decent amount of impressions and this will be very different from one ad text to another. |
Behind this all, is a whale of a lot of data - imagine data collected for every single variation of every single keywords, as ever used by any advertiser, for every single ad creative, since the inception of AdWords. After 2 plus years, we have a pretty good idea of what will work. And, of course, the system also looks at your history in particular, as well.
|I understand the concept as a way to queue up trial words and give them some subsequent "air time" but think that the old system was much easier to grasp and manage. This brings up the overall management cost of a campaign. |
Understood, and I'll pass on that feedback later in the week, Robsp. And again as mentioned above, the concept behind the whole system is really to improve our ability to show relevant ads to our users - as judged by those users - so that they will continue to trust AdWords, and click on your ads for a long time to come.
I have to say the new Adwords algo really sucks, especially when you start a new campaign and do broad match.
Several days ago I had most of my newly added keywords got in trial after only 10s of impressions. One even was put in trial after 3 searches! That's incredible! Yesterday I moved one of my well performed campaigns from one account to another. The consequence was half of the keywords got in trial and two were put on hold! I ended up calling my friends to click on those slowed ads to artificially turn them to normal.
With the new algo, the performance of the same keywords in other advertisers' accounts will be considered when estimating the performance of your keywords. What's the idea in that? The level of experience in managing Adwords ads is different among advertisers. Others doing poorly does not necessarily mean I'll do it poorly too. Let alone the keywords I moved from one account to another was doing great before!
I don't know if Google is really serious about making things more complicated and harder to control for advertisers. I miss the old days when you submit keywords, you don't have to worry about keywords being slowed after only a couple of searches.
Ok, well I've just noticed a typo in my post #12 above, which could lead to actual confusion. Oops. Let me correct myself...
In post 12, in the fourth bullet point down, I said:
|* Once the 'in trial' spots are filled, the rest of the keyword list then goes to 'in trial' status. |
Uh oh. Of course, I really meant to say:
|* Once the 'in trial' spots are filled, the rest of the keyword list then goes to 'on hold' status. |
Oh, those pesky brains. Sometimes you just can't trust 'em. ;)
|With the new algo, the performance of the same keywords in other advertisers' accounts will be considered when estimating the performance of your keywords. What's the idea in that? |
Yes, this really dilutes individual advertiser 'skill' so to speak. If you write a really good ad, it still may not be enough to overcome the previous history of the KW, and on the other end if you write a really bad ad, it still might make it in based on previous success.
Although I must say - this will only be the case for a very small percentage of ads.
On the positive side, if you can nab a KW now with newly increased barrier to entry from the algo change, you could really use that to your benefit.
Just thinking about this system some more...
Some keywords are strictly seasonal, or even only of mild interest for a few days or a couple weeks. How can these be distinguished? If too many advertisers try these terms during the off peak times of the year, will anybody be able to use them when the time is right? Will the 'early birds' to the new season be greeted by diasbled ads, acting only as 'water breakers' for when the system catches up to realizing the keyword is converting again?
Also what about an up and coming term. A dead term today could be a vital one tommorrow. (e.g. say a movie comes out containing a dead term or a band names themselves after one) Adwords will be behind the curve picking it up, no?
Ok i'm tired. I don't know what i'm saying anymore hopefully this post makes sense.
I would like to express my extreme displeasure with the new On Hold/In Trial system. Among a host of concerns, there are two major problems:
1. I have to now go through and edit thousands of keywords, in many ad groups, across various campaigns.
Early in my Google experience (1.5 years ago), I attempted to use broad match for the term widget. The problem was that Google would only show my ad when it thought widget was placed with relevant terms (for example, Google would not show my ad when someone searched for widget pencil or widget dog).
However, my company provides a general information service for the widget industry so almost anyone typing in the term widget (in any combination) should find the ad relevant. Widget in this case is a rather specific term which only people in the industry or people interested in the industry would use.
To counter this problem, I came up with long lists of keywords paired with the term widget. Though it required much more work on my part, this was working fine. However, the new On Hold/In Trial "feature" has thrown a wrench into the entire system. Many new keywords or keywords for other campaigns are being placed On Hold and cannot even make it to In Trial status since I already have a number of keywords from these long lists put In Trial. Because of the number of keywords I have, my In Trial basket is always full.
So now I need to sort through all of my keywords so I can have a relatively empty basket for any new keywords. This is a daunting task to say the least and one I do not look forward to.
2. Google's "predictive" technology hurts innovation and cuts their revenue.
My new keywords are punished before they start due to the performance of other advertisers. Why not let my ads run and fail? Is 1000 impressions or so really that much? Further, I have a suspicion that the system is putting keywords On Hold that haven't been used or have been used by very few advertisers. Just because other successful advertisers have not used that keyword does not mean that it is an unsuccessful keyword, it could be underused. For a number of keywords, I see no ad running on the page. Even if Google shows ads that are slightly less relevant, it should be using that space. At the least, it should run the ad and let it fail.
Google has lots of excess inventory it is not using. The On Hold/In Trial system only seems to exacerbate the problem.
I hope AWA reads this and can pass along to folks. These days Google has me scratching my head and wondering who is making the business decisions over there.
|However, my company provides a general information service for the widget industry so almost anyone typing in the term widget (in any combination) should find the ad relevant. |
That's the thing, it's like google now automatically does negative matching for you, assuming it knows best. Now for the average advertiser this may be true, because they are too lazy to add the proper negative keywords, but for advanced users it becomes a major impediment. I can see why google defaults to this system as overall it probably leads to less frustrated advertisers and greater relevancy, but the option should be open to use true broad match and do negative matching yourself if you so choose. Your example is one of the best reasons why.
A general rule, if most of the related broad match terms are relevant, it is generally better to have the user in control. If most of the related broad match terms are not relevant, it makes more sense for google to be in control. Unfortunately, we don't have a choice.
The only override is huge unruly keyword lists as you mention, which when you approach being relevant for almost everything, is flawed as you can never in reality pick up all the terms you seek, and under the current system google will never 'grant you license' to all of them either.
|since I already have a number of keywords from these long lists put In Trial. |
I think this is what happened to my account when the changes were first made. Suddenly tons of keywords were put on trial because the new system questioned them, and it took days to run through them. Afterwards, over 95% returned to normal - but losing the traffic for those many days was not a fun experience. I assume similar things happen now when you create new campaigns with large keyword lists. You might not get use from them for many days if the system decides to check them out one-by-one first.
Overall, I'm very displeased. I think I'm going to go ahead and email that over to Google as well.
|Overall, I'm very displeased. I think I'm going to go ahead and email that over to Google as well. |
JBrown, last evening I passed the link to this entire thread on to well over a hundred folks, in an Advertiser Feedback report.
So please know that your feedback, along with that of others, will be heard by lots of manager/decision-maker types at Google in general, and within AdWords in particular.
Thanks AWA! I definitely appreciate your efforts.
I just want to add my voice to the 100+ people at Google AdWords who will be reading this thread that I find this on-trial/on-hold system very detrimental to the entire AdWords experience.
For tens of thousands of advertisers who find new keywords with 5, 7 or even 25% clickthrough rates being put "on trial" the experience will be very beffudling unless they were one of the lucky few to stumble across the explanation in ths thread...
For the sake of advertisers, your AdSense clients and your shareholders, I think it makes the most sense to let advertisers try something and fail themselves, as opposed to pigeonholing them based on the past failures of others with particular keywords...
Every month new Websites, new technologies, new services, new products are being launched that could benefit from keywords which to other clients, in the past, may not have had success with. The Internet is always changing - as they say:
"Past performance is not an indicator of future results"
Add my name to those displeased with this whole "on hold/ In trial" mess.
I've just had conversations regarding an "underperforming keyword" in a client account. The underperforming keyword is the 4 word name of their company. Apparently, others using that name get better click thru than the owners of the name?
The rep was nice enough, actually both reps, giving me good advice (although I already have been successfully using the system since inception) but the system still refuses to allow me to keep an ad on the company's full and true name. If not one but TWO GA reps cannot help create an ad that clicks past the threshhold for a 100% relevant term (company name) then I don't know what to say. Perhaps based on its "history". I dunno.
I just have a real, real bad feeling about a system that allows competitors to steal visits from users who actually know and search for the complete legal company name of my client, while not even allowing us to do battle and bid against them. Lucky for us, I guess, that we are #1 in the organic side.
Other than that, spending over $500k with GAW has been pretty good. Google still rocks, just not as vigorously, IMHO.
>> Google has lots of excess inventory it is not using. The On Hold/In Trial system only seems to exacerbate the problem
Very Very True!
A side note on this (as far as inventory/broadmatch) -
Recently I needed to create a new ad & url in a different campaign for example keyword -big red bike-, (no quotes broadmatch). By instinct, trying to get more impressions, I created the new ad with keyword just -big bike- without much thought.
The result was I lost 90% of impressions. For the term, more people actually search on -big red bike- than -big bike-, but for whatever reason google decided to not match me to -big red bike- anymore. When I noticed this, I readded the -big red bike- keyword and regained all the lost impressions.
Inventory not being filled? For certain.
Has there been any change on this?
Been frustrated with the on hold thing for days...
also, how do we figure out how many in trials slots we have?
It won't give an exact number for slots, but you can go to Tools > Find and Edit Max CPCs to get a list of all your keywords currently "In Trial".
| This 41 message thread spans 2 pages: 41 (  2 ) > > |