New Adwords Traffic Estimator

[PPCPro]

Thanks Google for a useful New Traffic Estimator. However, the numbers seem a bit too high for certain categories. Can u kindly check AdwordsRep?

Cheers

[Nacho]

An impressions or search volume estimate would be nice too. It's hard to understand what is being used for average CTR like this.

[realgames]

I find the clicks estimates to be too low. Also the Suggested Max CPC and CPC given by the estimator seem flawed. For example, if I sell bespoke diamond widgets and I try the kw phrase "bespoke diamond widgets" into Estimator to get the Max CPC required to Rank 1 85% of the time it will come back with 4p. But this is wrong because Estimator is failing to take into account that fact that people searching for 'bespoke diamond widgets' will also be matched against other bidders who have the braod match kw widgets, phrase match "diamond widgets" etc. So the suggested max CPC is often useless.

[AdWordsRep]

Thanks Google for a useful New Traffic Estimator. However, the numbers seem a bit too high for certain categories. Can u kindly check AdwordsRep?

I find the clicks estimates to be too low. Also the Suggested Max CPC and CPC given by the estimator seem flawed.

I'd like to weigh-in with a few thoughts about the Traffic Estimator, if I may. Thanks for the invite PPCPro.

First, at the bottom line, the Traffic Estimator is a tool that is designed to give you a ball-park estimate sufficient to allow you to proceed with your decisions.

But it is worth noting that it is doing an exceedingly difficult job, in that it is attempting to give you a 'snapshot' in time, accurate in the moment in which you check, of a system in which the pertinent information is changing with incredible speed.

What is changing? Well, among other things, the number of competitors you have, how often their ads show (which is a function of their daily budget which they may change from moment to moment) the keywords they are using, the 'matching options' of those keywords, the Max CPC for each keyword, the current CTR of each of the keywords, and so forth.

For this reason, it is wise to treat the estimates as just that: estimates.

Really my best advice is to take the estimates for what they're worth (i.e. a jumping-off place), and then proceed.

By this I mean, use the keywords that you wish to use, set your Max CPC at a level with which you are truly comfortable, set a daily budget no higher than you are comfortable with, and then set the ads to running. At this point, it is of course very important to monitor your real world stats (which are an exact record of what really happened) and then make adjustments.

With all that said, please know that it is one of our more important ongoing goals to increase the accuracy, reliability, and usefulness of the Traffic Estimator.

And a final note: I'll be happy to pass on any feedback you all have about this tool. So have at it!

AWR

[AdWordsRep]

An impressions or search volume estimate would be nice too. It's hard to understand what is being used for average CTR like this.

I'll pass this feedback on later in the week, Nacho.

In the meantime, here is a 'trick' that'll get you a ball park figure for impressions:

The number of estimated 'clicks per day' is based on a CTR in the neighborhood of 2%. This means that if you are estimated to get 2 click per day, the system is predicting 100 impressions. And so on.

Still, IMO, concentrating on search volumes alone can lead one down an unhappy path. For example, if you use high volume keywords that aren't relevant to your ad (and your product or service) you're likely to get a low CTR, not to mention poorly qualified leads.

AWR

[Mikkel deMib Svendsen]

An impressions or search volume estimate would be nice too. It's hard to understand what is being used for average CTR like this.

I'll pass this feedback on later in the week, Nacho.

I know that this has been presented to Google several times over at least a year, so I think what we need now is more some feedback on why Google don't want to give us the impression figures. After all, as you say, impressions are what is being used as part the calculation you already make, so how hard can it be to turn it on so we can se it?

I understand all the limitations of estimates - and there are even more than you mentioned, but please at least give us - the "pros" - access to the raw numbers so we can make out own calculations. No offence, but I believe I do better calculations based on the knowladge I have of any market I work in that Google does ... If only you would give us the raw data.

[AdWordsRep]

I understand all the limitations of estimates - and there are even more than you mentioned, but please at least give us - the "pros" - access to the raw numbers so we can make out own calculations. No offence, but I believe I do better calculations based on the knowledge I have of any market I work in that Google does ... If only you would give us the raw data.

Your point is well taken - and I'll pass your well-stated comment on verbatim. In fact, I've just pasted it into the report I send out each week.

...so I think what we need now is more some feedback on why Google don't want to give us the impression figures.

I am not able to speak to the 'Why', Mikkel deMib Svendsen. I'm not really aware of why this decision, or other business decisions, have been made - as these are things which are outside of my arena here at AdWords.

With that said, though, I think one of my real values here lies is the fact that I am both very willing, and also able, to pass your feedback on directly to those decision makers.

AWR

[Mikkel deMib Svendsen]

I am not specifically talking about you, AWR, to tell us why we can't get access to impressions. The best thing would just be to get it However, as a large spending client I just don't understand why nobody at Google apparently seems to be able to get a clear answer on such a simple question.

For a lot fo AdWords advertisers I am sure it will be more usefull with the current click estimates but for me, and I know others have the same wish, it would be so much better to get the raw impression numbers.

[AussieWebmaster]

Originally Posted by Mikkel deMib Svendsen

I am not specifically talking about you, AWR, to tell us why we can't get access to impressions. The best thing would just be to get it However, as a large spending client I just don't understand why nobody at Google apparently seems to be able to get a clear answer on such a simple question.

For a lot fo AdWords advertisers I am sure it will be more usefull with the current click estimates but for me, and I know others have the same wish, it would be so much better to get the raw impression numbers.

I agree these numbers being held back hinder a thorough insight into Google search.
Not only on the PPC side but also on the organic. If we knew the impression numbers we would be able to better determine the habits of searchers - what percentage of people click where etc.
As it stands we have to conclude from other engines - and there are subtle differences between the demographics.
A Yahoo searcher is not the same as a Google or MSN searcher etc.

[PPCPro]

Pardon me if I sound harsh...

I was hoping that the Traffic Estimator would be help reduce "shooting in the dark", but alas..

With due respect to Google's attempt on offering a Traffic Estimator, I feel its throwing up numbers that are ... well.. weird.

Why can't the advertiser get data as basic as impressions estimate - referential data rather than dynamic data,as Mikkel pointed out? Who wants a "snapshot"?

Also, I am perturbed by :
1. how some sites never go off from the top position
2. impressions suddenly shoot up and since there is a 4-5 hours lag time between the results, things go out of control and CTR drops radically

I appreciate Google's attempt at being innovative and trying to introduce new features, but whats the use if the user is still groping in the dark?

Thanks

[realgames]

AWA -

My biggest issue with Estimator is that the suggested Max CPC it spews out is downright misleading. Here is a fictional example of how:

Someone selling bespoke diamond widgets who uses Estimator will be given a suggested maximum CPC (to appear in Rank 1 85% of the time) of 4p by estimator when he puts in his target Keyword Phrase:

"bespoke diamond widgets"

But if he uses this suggested CPC he will actually appear half way down Page 3 rather than in first position for searches for "bespoke diamond widgets".

The reason for this is that Estimator is only looking at other bidders bidding on the EXACT SAME Keyword Phrase "bespoke diamond widgets" but ignoring the many bidders bidding on broad match keywords such as "widgets" etc.

So the suggested Max CPC is dangerously misleading becuase it does not do what it says on the tin and, in my opinion, Google should either fix it or not show it at all.

[AussieWebmaster]

Unfortunately the new estimator is not much better than the old one... there just seems to be too much filtering... those curtains are getting in the way.


If you used results for a 4 week period with a buffer of a week say or however long it takes to make the data accurate. The estimator must have access to exact numbers. It does not have to be predictive by guessing... it can be predictive by using previous events and informing people based on that what they can hope to achieve.

Keeping too much hidden creates poor filtering.

[AdWordsRep]

Hi all - I'll be including the comments in this thread in the Advertiser Feedback report that I send out each week - now on Thursday evenings.

So if you have more thoughts on how the Traffic Estimator is, or is not, working for you - or how it could be improved, this is the time and place to post them for maximum impact.

Unfortunately the new estimator is not much better than the old one... there just seems to be too much filtering....

Ahhh. I should have picked up on this before. My mistake.

The subject of this thread is not actually a new Traffic Estimator. It is the same Traffic Estimator, which is now available in a more easily accessible location.

As I mentioned earlier in this thread, the Traffic Estimator is certainly understood to be a very important tool to advertisers - and work is always underway to improve it's accuracy and value to you.

AWR

[NFFC]

>So if you have more thoughts on how the Traffic Estimator is, or is not, working for you - or how it could be improved, this is the time and place to post them for maximum impact.

I would just print out Mikkel's comments and tell them to action them asap.

"I understand all the limitations of estimates - and there are even more than you mentioned, but please at least give us - the "pros" - access to the raw numbers so we can make out own calculations. No offence, but I believe I do better calculations based on the knowladge I have of any market I work in that Google does"

[AussieWebmaster]

Okay.


Since it is generally agreed that the numbers are skewed. What should be used are real past numbers... forget about futures.

From archived data you should be able to say what has happened previously.
Total impressions for a term... number of impressions seen by various positions... number of clicks gotten by various positions (yes we know this is impacted by the creative - but being given a general idea allows us to make better decisions)...
As opposed to what is there now... if bid changes could be tied to the percentage of the total impressions for a term and the position in the ten PPC results would be good.

[halfacat]

i agree the info given is too sparse. With Overture I look at the impressions more than clicks because my CTR always fluctuates depending on creative and relevancy. If the CTR here is based on 2% then I can figure out what the impressions are but its just making things into a pain the rear asking me to do this outside of your tool.

I do not look at these tools as some type of Oracle. They are a basic starting point and allow me to create budget estimates for clients and VP's.

[halfacat]

Once thing this doesnt address at all is Content results, any plans for that?

[OptimizeOnline]

Please count my vote also for a request to include impressions. It's useful to know that the clickthrough figures are based on an average 2% though!

I also find that the estimate of clicks produces spurious figures very often. I appreciate they are, after all, estimates but why when I add some additional keywords in AdGroups that typically lift clickthroughs by 20% (on average) the estimates that were reported at the time I added then to Google AdWords are actually *lower* by an equivalent amount. This doesn't happen happen on all occasions but with enough frequency to make me question how useful these estimates are.

[OptimizeOnline]

I forgot to mention that I also find figures estimated by Google for 'All Countries' are around 25% higher when compared to 'UK' figures for the same keywords. Now I can't say that this is very scientific testing at all but I've done this a few times for some clients' campaigns and this 'magic 25%' figure seems to pop up with some regularity - could it be that figures across countries are pro-rata'd somehow? Or are country-specific estimates' figures obtained from their respective regional Google 'databases'?

Originally Posted by realgames

AWA -

My biggest issue with Estimator is that the suggested Max CPC it spews out is downright misleading. Here is a fictional example of how:

Someone selling bespoke diamond widgets who uses Estimator will be given a suggested maximum CPC (to appear in Rank 1 85% of the time) of 4p by estimator when he puts in his target Keyword Phrase:

"bespoke diamond widgets"

But if he uses this suggested CPC he will actually appear half way down Page 3 rather than in first position for searches for "bespoke diamond widgets".

The reason for this is that Estimator is only looking at other bidders bidding on the EXACT SAME Keyword Phrase "bespoke diamond widgets" but ignoring the many bidders bidding on broad match keywords such as "widgets" etc.

So the suggested Max CPC is dangerously misleading becuase it does not do what it says on the tin and, in my opinion, Google should either fix it or not show it at all.

I haven't yet seen a response to this entry. It's extremely important to know if this is actually how the Traffic Estimator works; and, if so, if there're any plans on changing it to actually reflect the way searches are matched against keywords.

AWR, any comments?