I checked the recent threads and searched the archives, but it looks like this is a new development. Have we seen this before?
[ed]completed status list, cleaned grammar, and added bullets[/ed]
[edited by: mcguffin at 2:59 am (utc) on Dec. 12, 2003]
Status Max. CPC Clicks Impr
At risk £0.xx 2 82 2.4%
At risk £0.xx 1 96 1.0%
At risk £0.xx 109 7,457 1.4%
Disabled £0.xx 9 1,121 0.8%
At risk £0.xx 4 469 0.8%
Disabled £0.xx 0 4
At risk £0.xx 122 9,050 1.3%
Disabled £0.xx 1 61 1.6%
Disabled £0.xx 1 147 0.6%
At risk £0.xx 5 427 1.1%
I could go on and on....
Because of this, your invisable google-only CTR can shut down a keyword.
This explains so much! What used to appear to be a bug at least makes some sense now.
Disabled $0.xx 5,146 301,775 1.7%
If it's running well on Adwords' network, it would only seem fair to let it continue running.
Here are some more wacky ones:
At risk $0.xx 2,826 98,864 2.8%
Disabled $0.xx 265 16,966 1.5%
At risk $0.xx 53 3,845 1.3%
At risk $0.xx 1,984 28,076 7.0%
At risk $0.xx 3,797 123,971 3.0%
Disabled $0.xx 193 12,110 1.5%
At risk $0.xx 1 97 1.0%
I haven't missed the many posts reporting seemingly odd stats. And I am passing these comments on to the product development team for review. So if there are glitches, they'll be caught sooner rather than later, thanks to your posts.
I did want to chime in, however, with a bit of (what I hope will turn out to be) clarification.
Firstly, what the new 'Status' column amounts to is another way to view the same stats that have always been there. So please don't think that this feature has changed your stats in any way. It hasn't.
What the new feature is actually doing is showing you some information about your existing stats that you haven't had before: how your keywords are faring on Google alone, as opposed to on Google plus partner sites.
Now, it is entirely possible that a keywords will do extremely well on partner sites, and not well on Google. When you see a high CTR, and the keyword is shown as 'At risk', then this is the case.
The converse is just as possible. If you see a low CTR, and the keyword is marked as 'Strong' it means that the keyword has done well on Google, but not as well on partner sites.
Why? Well, it might be an audience thing, or a position thing as mentioned elsewhere by fidibidabah, or some other kind of thing.
In any case, think of the status column as a way to evaluate how your keywords are doing on Google, where their CTR is measured by the automated performance monitor that decides whether a keyword runs or is disabled.
I suspect that in a day or two, once it seems more familiar, you'll be really glad that column is there.
AWA
[edited by: AdWordsAdvisor at 11:40 pm (utc) on Dec. 12, 2003]
I guess the problem is that now that I understand the reason, I don't like the reason. Before I just blew it off as some weird inconsistency that would just take a lot of work to research and wouldn't produce anything productive for my clients.
Now that I know the reason, I wish Adwords would use the total network data as its basis for evaluating the effectiveness of a term.
Now that I know the reason, I wish Adwords would use the total network data as its basis for evaluating the effectiveness of a term.
Point well taken, cline. I'd like to very briefly explain why it's done this way.
Google has two groups of advertisers:
* Those who advertise on Google plus partners sites, and
* Those who choose to advertise on Google alone.
To be consistent, we need to hold all advertisers to the same CTR standard. This means evaluating their CTR on the one thing they have in common - which is the appearance of their ads on Google.
For this reason, CTR is evaluated on Google alone.
AWA
To be consistent, we need to hold all advertisers to the same CTR standard. This means evaluating their CTR on the one thing they have in common - which is the appearance of their ads on Google.
They have more than "one thing" in common: They have the appearance of their ads on the Google Adwords network.
Here's some more recent examples from a client who has lots of traffic and good Adwords conversion data. All of these terms are within the client's marketing allowable for conversions:
At risk 482 15,493 3.1% $0.05 $24.10 1.9 0.83%
At risk 3,594 142,329 2.5% $0.08 $257.50 2.5 0.70%
At risk 113 6,961 1.6% $0.05 $5.65 3.7 1.77%
At risk 440 17,408 2.5% $0.05 $22.00 2.3 0.45%
At risk 872 49,950 1.7% $0.06 $44.88 5.3 0.57%
We've got Google Adwords Network users out there that want to buy, and who have excellent CTRs (outside of Google proper). [In this case I'm pretty sure the difference is from AOL, as this client does particularly well with AOL's demographic relative to Google's.]
What's the most fair thing Google could do? I admit that it's not a perfectly clean choice, but from the perspective of the clients I represent (and their buying audiences) the most fair thing would be to be consistent across the network rather than basing it on the more narrow measure of just Google.com data.
Also, realize that if an Adwords customer wants to just run on Google they can, but they cannot run on just network search sites without also running on Google.
I am sorry to say, though, that I'm not really sure what you mean when you say:
They have more than "one thing" in common: They have the appearance of their ads on the Google Adwords network.
Those that appear on Google only do not appear on the partner network, unless I have misunderstood you.
Have I missed something?
AWA
Stated another way, the CTR data shown on our Adwords accounts is that from the Adwords search network. The advertiser may choose to do just Google or do the search network, but the data shown are from the search network regardless. Mathematically this could be either just Google or Google + a bunch of other sites (AOL...).
What I'm suggesting here is that it would be more fair and more consistent to use the search network numbers for CTR cutoff levels than to use just the data from google.com. The problem I'm pointing to would then not affect Adwords customers who were only running their ads on google.com, nor would it affect customers who were running their ads throughout the Adwords search network.
Heck, I'll go a bit further. Why not tell me what the CTR is on Google, and on the other network sites? Can I track them differently (to determine which have the highest conversion rates)?
Right now Google is acting like an affiliate network where all traffic is referred via Google so only G knows which affiliate is generating sales. That hardly seems fair.
This is how Overture's bar graph performance works. They compare how each ad is performing as a part of the total advertisers. If the ad behind you has a significantly higher CTR then you get a low mark.
To me it's a keyword where I am minimising spend and maximising ROI/conversion.
To Google it's a keyword where I am bidding high enough to maximise clickthrough and therefore maximise their income.
The fact that certain keywords are slowed / disabled is a GOOD THING if those keywords don't convert.
I can't help thinking this is a tactic aimed at panicking you into bidding higher to push poorly performing keywords into 'strong' positions.
Col
I'm sure most of us would like the exact G only CTR - but this is definately a nice step in that direction.
What is the reason for vague groupings like "Strong/Moderate/At Risk" instead of the exact (color coded) CTR anyway? As archie mentioned - can you at least tell us the corresponding ranges, AWA?
Hopefully today we start.
I do understand and agree with Google's policy of basing keyword ranking ONLY on Google's CTR, since this is what is most fair. But, I don't agree with Google's policy of disabling keywords based soley on Google CTR.
For example:
I have some keywords that nearly no other companies have bids on (either because they are poorly converting terms, or because it is impossible to keep an acceptable Google CTR). Regardless, I am one of the only people bidding on the term and I have an average network CTR of 3.7% on these terms so far this month. Some of these terms with a 3%+ CTR rate are getting disabled.
Now, if there are no other bidders on these terms and Google is disabling the keywords, Google is foolishly throwing money out of the window. This policy is losing money for Google, their search partners, and my company.
So, to recap. I think it is fair to base ranking on Google CTR only. But, when disabling keywords, Google should base it on overall CTR.
a -- Obviously it will cost me money, because I currently have the 3rd most productive (clicks & conversions) kw in one ad group labeled "At Risk" and about to be axed.
b -- The status metric is based on an invisible quantity, and therefore is not controllable without multivariate tensor calculus and a predictive fuzzy logic system.
Would all those who are not spending someone else's money please stand? Thank you.
If you used overall CTR - I'd immediately axe content ads. I get a few hits from them, but they'd bring down my CTR overall, and in some areas, by huge amounts.
I don't mind using a google only CTR - I just wish I knew the exact number instead of strong, moderate, etc..
If you used overall CTR - I'd immediately axe content ads. I get a few hits from them, but they'd bring down my CTR overall, and in some areas, by huge amounts.
You shouldn't have to: Google should do that.
Low CTR in content sites: axe ad on content sites.
Low CTR in Google: axe ad in Google.
Also let people know what the CTRs and conversions are for each, so they can run G only, content only or both.
Result? Less guessing, more relevant ads, more money for G, more money for advertisers.
Why on earth isn't this happenning? Why all the complicated rules?
Since the new feature's introduction, suddenly there are lots of them at risk. How can this be?
One possible explanation, and I'll apologise in advance if it seems a little cynical..
The Florida update nuked some big traffic sites, who still need the traffic, so they get an Adwords account rolling fast. They're used to being number one, so they bid top dollar. They are not alone, some of their competition got nuked too, so they get an Adwords .... yada yada
Net effect = higher overall bids, particularly for big money terms.
I'm not suggesting that this was Google's intention, but it may be an unfortunate side effect of the Florida update. Weird how this feature was introduced just after the end of Florida too?
Low CTR in content sites: axe ad on content sites.
The problem with this is that a lot of times we might generate 10k+ views because of a news article one day, and still maintain a consistant hit total per day, which we see on a daily basis.
This would make it so that if on any particular day, a news article was run, our content ads would be gone, but the next day they'd generate a decent CTR again - that doesn't make a lot of sense.
Since our content ads are often 0.5-0.8 position, this would end up axing all content ads in our area. I don't think that's a good solution either.
Just stopped by this forum briefly - has been a busy day here, and I've had very little time on WebmasterWorld since early this morning.
Anyway, I saw the flurry of activity on this thread, and wanted to say that I'll do my best to answer these questions later in the day - probably several hours from now.
Until then, one thing did catch my eye, and I wanted to comment on it in the few minutes that I have.
Not sure if I'm alone here, but before this new feature was introduced all my keywords were doing just fine. Some had been succeeding and achieving a high enough CTR for months without any bid changes.
Since the new feature's introduction, suddenly there are lots of them at risk. How can this be?
It is really important to note that nothing has changed with your actual statistics. The only thing that is different is that you now have an additional way to look at them.
So, if a keyword is showing as 'At risk' now, that is not a change. It was at risk before the new column was added, but you weren't able to see that. The new 'Status' column, however, specifically tells you.
So think of the 'Status' column as a new tool to see the exact same statistics in a different way.
Hope that clarifies at least one point from above.
More later. :)
AWA
Test my logic here, if you will:
- the new 'status' is the correct way
- new status not visibly related to old methods
- therefore, CTR is invalid measurement of kw value.
QED
You can't have two unrelated systems of measurement controlling one system from one control point. The result is chaos. (pardon me, my engineering background is showing...)
or in plain english...
If Google wants to install an arbitrary system of keyword valuation, then they should just do it. (already did it?) But do us a favor, and drop the "real-world" measurements (eg CTR). They just confuse the issue.
Point of order -- "new tool to see the exact same statistics in a different way"
Not "exact same" at all. Maybe same stats used *internally* but not the ones I can see. That makes ALL the difference.
AdWordsAdvisor, that's all good, but the new status thing has no relationship (that I can see) to the CTR, etc. we were using before. Does that mean that we have all been calculating the relative value of our keywords incorrectly before?
In brief, here are what the two columns show you:
* The CTR column in your Ad Group stats reflects each keyword's CTR for your ads appearance on Google, and all partner sites on which you have appeared.
* The Status column reflects each keyword's status on Google alone. This has been a frequently requested feature, as it is on Google alone that keywords are evaluated in terms of whether they continue to run, or are disabled.
Many have posted in this thread wondering why an 'Status' indicator is shown, rather than a 'simple' number.
The answer to that is that the number is actually not so simple. Underneath the hood, AdWords is more complex than you might imagine. The very short story is that the Minimum CTR required to keep running is actually slightly different for each position. It is 'normalized', and the minimum is actually slightly lower for each position moving down the page.
So 'just showing the number' is not straightforward, nor is interpreting that number relative to position. As a result, the decision was made to show a 'Status' instead.
One advantage of that decision, IMO, is that it allows you to quickly sort your keywords by 'Status', and interpret at a glance how they are performing.
Having said all of that, please know that I have heard your comments. I assure you I'll pass them on to the appropriate team at Google on Friday.
AWA
Point of order -- "new tool to see the exact same statistics in a different way"Not "exact same" at all. Maybe same stats used *internally* but not the ones I can see. That makes ALL the difference.
Ah, words are so hard to pick sometimes. ;)
What I had hoped to convey is that the new column does not change your statistics in any way. It is simply a way to look at existing stats, to interpret keyword performance on Google alone.
Hope this makes it more clear.
AWA
Test my logic here, if you will:
- the new 'status' is the correct way
- new status not visibly related to old methods
- therefore, CTR is invalid measurement of kw value.
Actually, I don't think the way you view them has changed. Would you rather have a moderate kw w/ a 10% CTR or a strong KW w/ a 5% CTR?
I think this lets us know when an ad has a 5% ctr, but could be disabled because of the G CTR, something we'd never have guessed based on total activity, and therefore can preemt the disabling, but overall, the way you treat your campaign hasn't changed that much with this new feature.
So, explain this: I have turned off the content ads for about a month now (ask me why). So, the CTR should equal the GCTR, right? No way. There is no corellation between the Status indicator and CTR.
Conclusion: either my assumptions are wrong, or GCTR does not determine status.
The point I'm trying to make is that I don't see any way for a user (advertiser) to control both the CTR and the Status in a rational way because the status indicator has no (known) rational basis (see above). It makes no sense that a keyword with a high double-digit CTR is now "at risk".
But doesn't that make things worse, though? Why hinge everything on GCTR? What part of my ad text do I tweak to increase GCTR without killing CTR? That is, how do you optimize your ad for Google and AOL, etc. separately? Because the ad doesn't do well on Google, then I can't run it on AOL? Huh?
Like one of my clients asked once, "What's goggle?"
