basically the client is usually in the top 3 for their main keyword all day long with an average of around 2.5 position on the day.
sunday I got up at 9 am and he's sat in 8th to 10th place, I couldn't figure it out but there were 4 newbies now in the skyline which is bizzare as it's a decent spend on that one keyword as it at around $800 a day so they'd have to be brave to wade in like that and have very high bids
bizzare as it was, the only explanation I could think of (quality score was great) was that 4 newbies had indeed waded in on the same day with skyhigh bids and knocked my guy and the other guys who were in the top 3 off.
I upped the bid and nothing happened, I upped the bid again later and nothing.. and so on.. so eventually I whacked it up by 50% of the original bid (that usually has us welded to the top slots) and it moved from 8th/10th to 5th and never moved.
later that day I logged into the office PC via vpn and checked positions and it was in first place..
I then did 20 searches on each pc and found :
home gave 20 number 5 positions
office gave 20 number 1 positions
neither ad moved at all.. there were welded in 5th or 1st depending which PC I used (both use different ISP's.
it would seem that google are serving different results to different ISPs, the end result is.. I have a position average of 2.5 and a CTR that is about 60% less than it should be and now I'm getting charged more per click to maintain that 2.5 average than it was before.. about 20% more in fact
Anyone else noticed anything similar? seems to work in googles favour as we have to jack up bids or suffer a position 3 or 10 .. or pay more and get a 1 or 5 position giving a 2.5 average.
It'd be nice if someone from adwords could comment on this.. is it an anomaly or multi ad testing in play (hope not.. really do or we are looking at an aggregated top 10 .. can't see how that would get any cheaper to play in either) I know there is a slight variance sometimes in positions but fixed in first compared to fixed in fifth isn't a slight variance at all ..
In any case, in my opinion, it's best to focus on overall stats to gauge ad performance - rather than on the answer to the rather limited question of "where do I see my ad right now?"
Also, I think advertisers would benefit from using the 'ad preview' tool to look at their ad placement - since it allows you to see what most other users see, and also does not cause you to accrue an impression while looking.
More info on the ad preview tool (which I think is under-utilized by most advertisers) may be found here:
What is an ad preview page?
[adwords.google.com...]
Since I think that the ads preview tool is not as well known as it could be, I'll also excerpt much of the info from that page:
An ad preview page enables you to view ads and search results as they would appear on a regular Google search results page, without accruing extra impressions for your ad.To perform a test search, follow these steps:
1. Go to www.google.com/adpreview.
2. Enter your search term. Click Go or hit Enter on your keyboard.
3. The search results and ads appear as they would on Google, except the links on the preview page aren't clickable.You can also use the ad preview page to see if your ad is appearing in other geographical locations...
Some other thoughts or observations:
I then did 20 searches on each pc...
Please recall that when you repeatedly search on your keywords, you are negatively effecting your own CTR in the short term - which could potentially explain what you mention in your quote below, ShaneRbn21:
I have a position average of 2.5 and a CTR that is about 60% less than it should be and now I'm getting charged more per click to maintain that 2.5 average than it was before.. about 20% more in fact
btw, my budget is unlimited in essence 2K a day or something like that....
Please note that your budget is not directly related to your ads position, ecommerceprofit. Daily budget limits how often one's ad will appear, rather than in what position.
AWA
<edit> Fixed link, and several other small changes: I hit Submit rather than Preview, before I was finished composing. ;) </edit>
[edited by: AdWordsAdvisor at 10:54 pm (utc) on June 4, 2007]
Thanks for your response, well it's google's baby so they are well within their rights to test however they want, but this is causing some pain for sure and I'm not sure how it could be spun into something positive for advertisers to be involved in unless you are one of the newbies who suddenly goes from 10th to first by doing nothing.
I take your points on board and agree in principle with them but in practice I'm afraid that after weeks of being in the same spots day in day out, hour after hour then coming online to see that the ad that *should (all things being normal) be in position 2/3 is suddenly down in 8/9/10th place it's natural human behavour to think oh my god what's happening especially when the stats back up that there is a problem, reduced CTR, lower traffic volume etc.
if on 364 days of the year the ad is where it *should be then on day 365 it isn't then I am going to rub my eyes and search again and again until the painful truth is clear and I know I really am showing in that position I *shouldn't be.
Yes we should all use the ad preview tool .. but how many of us do?.. would it not be far simpler to log the IP we use to log in to adwords accounts with and then ignore the searches (and any clicks within the same sector) on any adwords from those IP's too? seems a simple logical step to me than trying to get us to use a preview url
Anyway back to the problem in hand ..
"Please recall that when you do this, you are negatively effecting your own CTR in the short term"
Yes I agree 20 searches on each PC would detrimentally affect CTR by a very small factor BUT.. and it's a two part but ..
1) as long as no ad was clicked at all then everyone would be exactly the same as they were before the searches but with like a 0.0001% less overall ctr.. but still all relative to each other... everyone would have 20 impressions and no clicks.
2) which would you consider is more detrimental? 20 searches to establish the current position (based on the other 364 days of seeing your ad where you expect to see it) or the hundreds of genuine searches steaming through that we are now missing out of clicks from as we are down in 8/9/10th place despite having a bid that would guarantee us first place the day before and have some spare change from? just how badly is that affecting our CTR compared to 20 test searches to establish where we are when CTR, ad position and traffic volume is slipping
Not sure if this is testing but it smacks of increased bid costs for the major players for two reasons..
1) you'd do a search and see yourself low down and yet the positon says 2.5 average yet you simply do not see that .. you see yourself in position 5 etc... and see the clicks are down, the CTR is down yet the quality score is ok and 4 new players above you.. so what would you do? sit and watch the CTR spiral down more knowing it's a case of pay it now of pay double later to build it back up so you bid up a bit and wait for a reaction, which doesn't come?
2) the newer ads that wouldn't normally get a look in are now building a nice CTR so what , as artificially dumped down the page as we already are, hope have we of getting your CTR up to a position to challenge them without bidding up?
sure hope it's a testing anomaly and this "agregated" version of the top 10 disappears soon as the only party in the equation that benefits is the newbie advertiser who get a play up top where he wouldn't normaly end up based on their current bid and CTR... (well those newbies and then google benefiting from the increased bid hikes made by the previous slot 1/2/3 residents to recover lost traffic and positions)
but I can't help being a natural born cynic as thesedays on competitive terms the price doesn't ever go down when a new feature is introduced it's only upwards and onwards as of late.
[edited by: ShaneRbn21 at 11:33 pm (utc) on June 4, 2007]
for many of us, G is our home page and we lazily open a browser to look at ads.
think i'll change my home page, on my main work computer, to adpreview for a little while to see if i can train myself to develop better habits.
basically it's supposed to work along the lines of :
user searches for torch
user searches again for torch, algo realises he didn't click anything the first time so serves some different results second time
user searches for torch again.. algo kicks in and shows another set of results as first and second time didn't yeild any clicks
valid logical argument.. yes.. but in practice it's not happening this way
when I searched the ad was fixed in position not moving downards at all.. 20 number 1's 20 number 5's so something wasn't right
hey ho.. adpreview tool it is.. be better if you could input a list of "account IP's" to be excluded from having an effect on google.. far better overall solution if it was IP based methinks.
Also - Search results are often different if you're using google.com -- as compared to the google tool bar. I'm not sure why.
