homepage Welcome to WebmasterWorld Guest from 54.205.168.88
register, free tools, login, search, pro membership, help, library, announcements, recent posts, open posts,
Pubcon Platinum Sponsor 2014
Visit PubCon.com
Home / Forums Index / Google / Google AdWords
Forum Library, Charter, Moderators: buckworks & eWhisper & skibum

Google AdWords Forum

    
Is the top spot worth it?
Effect of position on click through rate
sunnylyon




msg:1152032
 12:42 pm on Jun 13, 2004 (gmt 0)

I need to boost my traffic. I've exhausted all the keywords my brain can think of. My texts are masterpieces of poetry and commercial logic (irony) - I don't want them to be any more attractive for fear of luring time wasters.

So I sit there looking at the statistcs, wondering whether it's worth raising my max bid in the hope of getting a higher position. A key element in my decision has to be the effect of position on clickthrough rate. Presumably a higher position will give a higher clickthrough rate and thus create a virtuous circle, so what I need is a formula

For clickthrough rate x in Position 1

Position 2 = x - a%
Position 3 = x - b%
etc. etc.

Anyone any realistic estimations? My own stats have just too many variables & the quantities aren't meaningful.

The relative conversion rate is a related issue. We all suspect that clickers on the number three spot are better customers than clickers on the number one spot because they've bothered to look a bit further. But by a factor of what?

Of course I appreciate that the ad text, the keywords, the market are all important factors, so we'd never come up with a definitive figure. I'm just trying to isolate the "google" variable in the equation.

Any suggestions?

 

Essex_boy




msg:1152033
 1:48 pm on Jun 13, 2004 (gmt 0)

Depends.

Not a very good answer!

I found that with one product I was selling I couldteh top spot for 53p or be third and final for 4p.

Guess which one I went for?

I think your better going with well written copy for the advert and being around half way to athird down the page.

All things being equal youll be ok.

But that doesnt come with a money back gurantee, so try it out.

nyet




msg:1152034
 2:04 pm on Jun 13, 2004 (gmt 0)

We all suspect that clickers on the number three spot are better customers than clickers on the number one spot because they've bothered to look a bit further.

I would not assume this.... ('better' how?)

clickers on position #3 might be better converters (as a percentage) because they are looking harder, but there might also be considerably *fewer* of them as well.

Clickers on position #1 may not keep looking because they found what they want and have bought it.

Clickers on pos #1 might also tend to not care as much about price.

it all depends.....

there is no set and fast rule which works for everyone.

The *amount* of traffic will undoubtedly go up in position 1. Weather or not that is *good* traffic is another thing.

Your best bet is to try it for a while and see.....

[slight rant] ROI is a good measure, but it is ROI of the whole business one should calculate. After all, isn't the *Bottom Line* the most important thing? I could tell my boss that I have structured his campaigns for maximum ROI (per word) alone, but when he sees only 2 sales per month, I am fired..... It has to be a *balance* between ROI and volume to make some money! [/end rant]

Robsp




msg:1152035
 2:23 pm on Jun 13, 2004 (gmt 0)

With sufficient conversion data you should be able to get a decent statistical relation between position, volume and conversions. Anyone did an analyses like that before?

sunnylyon




msg:1152036
 3:58 pm on Jun 13, 2004 (gmt 0)

I entirely agree with your rant Nyet, Volume x ROI is the key to the whole thing.

I'd happily pay to move up in position, if I knew it was going to boost turnover, nevermind a drop in ROI.

My problem is that in my niche market, "trying and seeing" takes weeks to create statistically important numbers, by which time factors such as "everyone's gone on holiday" mess up their value.

That's why I was curious to know if any of the big players out there had estimated out the effect of position on volume (notwithstanding the caveat that each ad is to a certain extent a unique case).

nyet




msg:1152037
 4:29 pm on Jun 13, 2004 (gmt 0)

My problem is that in my niche market, "trying and seeing" takes weeks to create statistically important numbers, by which time factors such as "everyone's gone on holiday" mess up their value.

Sounds like our business! Our service gets few conversions but costs thousands of dollars. So, like you, we have to wait months for real data. Since we do not 'sell' online (we close deals in person) we measure the average page views per user per word as our measure of 'conversions'. Our logic is that the more pages viewed, the more interested the potential customer is.

We have been bidding to have our ad #1 for high number of visits and also to 'show' the consumer we are a 'major' player. BUt that is just *our* solution.

That's why I was curious to know if any of the big players out there had estimated out the effect of position on volume (notwithstanding the caveat that each ad is to a certain extent a unique case).

I am sure many have done this for thier business model. The problem in such a question is like asking 'what is the average price for ALL items on Ebay' It can be calculated but it is a number which has little meaning. To make it meaningful, you would have to ask "what is the average price of 2 year old used camcorders" (for example)

eWhisper




msg:1152038
 5:00 pm on Jun 13, 2004 (gmt 0)

You might want to pick a few keywords first, and slowly increase their max CPC to see what effect it has on position, CTR, and conversions.

If you have a lot of keywords, I wouldn't necessarily jump up all the bids at once, but pick a few good keywords that you are very famaliar with and have excellent historical data on, and experiment a bit to see if raising your bids leads to higher profits.

Here's an old post about an experiment I ran:
[webmasterworld.com...]

skibum




msg:1152039
 10:28 pm on Jun 13, 2004 (gmt 0)

IMO, the main reason not usually to go for the top spots is because the clickable area is waaaaaaaaaaaaaaaay to large.

In spots 3+, the only thing clickable is the title, for the premium positions the whole "banner" area across the top of the page is clickable.

That's not to say it always advisable to stay out of the top spots, but the number of accidental clicks is likely to be much higher on the top spots.

On the other hand, that increased CTR may lower the CPC so that in the end it comes out to be a wash.

blaze




msg:1152040
 4:58 am on Jun 14, 2004 (gmt 0)

Another thing to consider is .. do you compete on price or features?

If you compete on price, a lower spot will be fine because you will sell well to comparison shoppers.

However, if you compete on features the top spot can be great becase you'll wow the impulse buyer.

AdWordsAdvisor




msg:1152041
 6:09 pm on Jun 14, 2004 (gmt 0)

You might want to pick a few keywords first, and slowly increase their max CPC to see what effect it has on position, CTR, and conversions.

Very smart plan, eWhisper.

IMO, the main reason not usually to go for the top spots is because the clickable area is waaaaaaaaaaaaaaaay too large...

Excellent point, skibum, which I'll pass on. I agree that it'd be good to be consistent between the top two ads and the ones on the right.

So I sit there looking at the statistcs, wondering whether it's worth raising my max bid in the hope of getting a higher position.

Welcome to WebmasterWorld, sunnylyon!

Not in this thread in particular so much as in general, I'm not really certain what we mean when we refer to the 'top' positions. Do we mean the (up to two) ads at the top of the page, or the top ads on the right hand side? This is important, as the two locations are different in some ways worth noting.

Position on the right hand side is based on a slightly different algo that position on the top.

* For the ads on the right your Max CPC and your CTR are equally weighted.

Ads on the right can be sent to the top spots (maximum of two) if the meet an additional performance standard that weights CTR more heavily.

* For the ads on the top, your actual CPC is more important than your Max CPC, and your CTR is more than 50% of the equation.

So, you can't really send your ads from the right to the top by simply paying more. The best way to get up there is to really target your keywords/ads so your get a higher CTR.

One other note: When, in your stats, you see 'Avg. Pos.' of 1.0, this refers to the first ad on the right, rather than the first ad on the top.

AWA

nyet




msg:1152042
 8:34 pm on Jun 14, 2004 (gmt 0)

When, in your stats, you see 'Avg. Pos.' of 1.0, this refers to the first ad on the right, rather than the first ad on the top.

! gasp!

Major failure to not report *actual* position...Doesn't G *hope* that we (some of us) want that top position?

it should be somehow reportable! (x1? -1?)

!

AdWordsAdvisor




msg:1152043
 1:58 am on Jun 15, 2004 (gmt 0)

! gasp!

Well said, succinct, and to the point!

I'll pass your comments on, nyet...

;) AWA

eWhisper




msg:1152044
 2:05 am on Jun 15, 2004 (gmt 0)

One other note: When, in your stats, you see 'Avg. Pos.' of 1.0, this refers to the first ad on the right, rather than the first ad on the top.

I got an ad with a pos of 1.3 thats been in one of the first two spots on the top for over a year. It's never been on the right since the old premium ads were removed and ads bumped up.

Are the top ads suppose to have like .7 again or?

eWhisper




msg:1152045
 9:59 pm on Jun 16, 2004 (gmt 0)

*bump*

hoping for an AWA clarification on how premium posistions are being displayed.

AdWordsAdvisor




msg:1152046
 10:35 pm on Jun 16, 2004 (gmt 0)

*bump*

Thanks eWhisper. This one fell off of my radar.

I'll post again when I have more specific details. Have to go find one of the Panel of Experts.

AWA

UpDown




msg:1152047
 10:48 pm on Jun 16, 2004 (gmt 0)

In my experience the first ad on the page is position 1. If there are two premium placements they are 1 and 2 and the first ad on the right is reported as 3. But if there are no premium ads then the first ad on the right is repoted as position 1.

I have an ad which is consistently at 3.0. It is the top ad on the right.

Average position 9 or 10 is confusing as it could be bottom of page 1 or top of page 2 depending on whether any ad qualifies for the top slot.

ngsmith




msg:1152048
 9:03 am on Jun 17, 2004 (gmt 0)

I wanted to follow up on some of the earlier posts on this topic as I have done some of the mathematical analyses that give you a "better than hunch" answer to the question.

Given good enough input data it is possible to model the correlation between bid position and the impressions, clicks and conversions you get. A straight line seems to be a good enough approximation.

Knowing which values you attach to each of these things you can then build a (not insignificant) non-linear simultaneous equation and solve to get the best combination of bid positions to maximize your Profit for a particular budget(it's true that ROI is irrelevant as long as it is > 1).

It's no small amount of maths so don't expect to do it on your calculator - or even in a spreadsheet. The results can be very illuminating though.

AdWordsAdvisor




msg:1152049
 7:52 pm on Jun 17, 2004 (gmt 0)

hoping for an AWA clarification on how premium posistions are being displayed.

I haven't forgotten this (again). It's just taking a while to hook up with the expert I have in mind!

AWA

nyet




msg:1152050
 8:02 pm on Jun 17, 2004 (gmt 0)

build a (not insignificant) non-linear simultaneous equation and solve to get the best combination of bid positions to maximize your Profit for a particular budget

ouch..ouch..my head hurts!

AdWordsAdvisor




msg:1152051
 8:06 pm on Jun 17, 2004 (gmt 0)

ouch..ouch..my head hurts!

Mine too, actually.

nyet, are you getting funnier by the day, or is it just me?

AWA

nyet




msg:1152052
 8:33 pm on Jun 17, 2004 (gmt 0)

AWA, my son said that the other day...but i think he had a different meaning in mind

Robsp




msg:1152053
 9:08 pm on Jun 17, 2004 (gmt 0)

ngsmith,

Can you be a little more specific as to how this can be done in e.g. excel? As far as I know there are quite a few statistical functions in there that should make it work (but my math is kinda rusty after 20 years :) ). If you do not want to post just sticky me some details.

AdWordsAdvisor




msg:1152054
 10:24 pm on Jun 18, 2004 (gmt 0)

hoping for an AWA clarification on how premium posistions are being displayed.

Ok, finally.

Well, folks. I learned a lot on this one. For example, how, after two years of working daily with a product it is still possible to have a misconception.

So, I was mistaken (or only partially correct, anyway) way up there in post 10 when I said:

One other note: When, in your stats, you see 'Avg. Pos.' of 1.0, this refers to the first ad on the right, rather than the first ad on the top.

eWhisper was correct in his observation in post 13 and UpDown was right on the money in reporting:

In my experience the first ad on the page is position 1. If there are two premium placements they are 1 and 2 and the first ad on the right is reported as 3. But if there are no premium ads then the first ad on the right is repoted as position 1. I have an ad which is consistently at 3.0. It is the top ad on the right.

Average position 9 or 10 is confusing as it could be bottom of page 1 or top of page 2 depending on whether any ad qualifies for the top slot.

So, thanks for pushing back a bit, and sending me off for a little research.

BTW, eWhiseper asked:

Are the top ads suppose to have like .7 again or?

This was caused by a bug, quite a while ago - and long since fixed. I hope no one is seeing that in their current stats! ;)

AWA

eWhisper




msg:1152055
 12:00 am on Jun 19, 2004 (gmt 0)

Thanks for the clarification.

Does G track what page numbers, or premuim vs non premium ads were clicked?

Running reports where one could see the average page number, and premium vs side position would be useful.

Something like ad position: 13.2. Where 13 is the ad number and 2 is the page number could clarify some issues.

running scared




msg:1152056
 12:44 pm on Jun 19, 2004 (gmt 0)

How about some nice stats that would show you what average CTR you actually get in different positions. No that really would be a reporting headache.

What I love about adwords is how such a simple concept has become so complicated that we need to be solving simultaneous equations to be getting the most out of it. So much potential to find competitive advantage. Every little helps

AdWordsAdvisor




msg:1152057
 7:48 pm on Jun 21, 2004 (gmt 0)

Does G track what page numbers, or premuim vs non premium ads were clicked?

Not in the sense that I think you mean, eWhisper. Although the information is of course incorporated in the Avg. Pos. that you see in your stats, for whatever date range you are looking at.

Running reports where one could see the average page number, and premium vs side position would be useful.

Something like ad position: 13.2. Where 13 is the ad number and 2 is the page number could clarify some issues.

I'll pass this on later in the week.

AWA

sunnylyon




msg:1152058
 7:08 am on Jun 22, 2004 (gmt 0)

What would be ideal would be to have a report available by "real position" as opposed to average position. Quadratic equations aside, the difficulty is knowing whether in any given case "av. position 3.7" means "half the time at position 1 + half the time at position 7", or "all the time around the 3 / 4 spot" or any other variation on that theme. If we were to have a report "Ad Position Data" with Collated statistics

Position 1 Impressions CTR Conversion Data
Position 2 Impressions CTR Conversion Data
etc etc

collated accross all campaigns, or adgroup by adgroup,
that would be an enormously powerful set of statistics...

AdWordsAdvisor




msg:1152059
 10:58 pm on Jun 22, 2004 (gmt 0)

What would be ideal would be to have a report available by "real position" as opposed to average position.

Interesting point, sunnylyon.

When I think of reporting of actual position, per keyword, per impression, though, I start getting the shakes.

I think of accounts with tens of thousands of keywords, and with a lot of those keywords getting tens of thousands of impressions daily.

Each one of those impressions is likely to be in a different positon, since it position can literally change for every search.

That is an awful lot of data to store in a database, or for an advertiser to look at in a report!

Welcome to WebmasterWorld, btw. You'll pick up a ton of info here.

AWA

ngsmith




msg:1152060
 7:45 am on Jul 2, 2004 (gmt 0)

Following up w.r.t. Maths

Given sufficient input data LINEST in Excel is what you could look at to give you correlation information between bid position and cpc / impressions / clicks / conversions.

You get a different equation for each keyword and then if you assign a $ value to each impression, click and conversion you can calculate your total estimated Return for any position for a specific keyword.

Assuming that you have a limited budget, and/or don't want to pay for unprofitable advertising the challenge is then to vary all the bid positions for each of the keywords so that you get the maximum profit (Return - Ad Cost).

This is the non-linear equation bit - a useful F.A.Q. can be found by Googling "nonlinear programming FAQ".

Solving these equations by brute force human effort for 5 or 6 keywords is feasible, but more than this and you need some software. I have a user for the software I wrote who solves the equation for 1,782 keywords. It gives him an estimated profit improvement of 54% over the campaigns he had been optimizing manually.

I'm afraid that I'm not prepared to go into details on the maths I use here - that's my software's secret.

ngsmith




msg:1152061
 7:50 am on Jul 2, 2004 (gmt 0)

Follow-up w.r.t. detailed position reporting.

From the graphs that I have plotted, the average position that Adwords reports are sufficient provided you keep your bid constant for at least a full 24 hour period and go out of your way to change bid positions only at the start of a new Google reporting day.

Given sufficient volume of well-collected data any "funnies" get rinsed out in the statistical wash.

Neil.

Global Options:
 top home search open messages active posts  
 

Home / Forums Index / Google / Google AdWords
rss feed

All trademarks and copyrights held by respective owners. Member comments are owned by the poster.
Home ¦ Free Tools ¦ Terms of Service ¦ Privacy Policy ¦ Report Problem ¦ About ¦ Library ¦ Newsletter
WebmasterWorld is a Developer Shed Community owned by Jim Boykin.
© Webmaster World 1996-2014 all rights reserved