I sell widgets. And i have 27,000 widgets with particular names/model numbers and parts that people search for.
I have been advertising these widgets by very exact terms - model numbers / skus that are commonly looked for and detailed product specs. For example I sell a famous brand widget model number 11-1111-sk-v. I get a horrible quality score if i phrase match that (because i'm not interested in having someone click on an ad for another widget by the same brand i may not carry). For the life of me i trained myself to "target target target"
Google/Adwords suppor has basically said "We can't tell you WHY your QS is low but we can tell you that going to broad match we will increase your QS" The gist of poor quality score is THEY don't generate enough earnings per click - irregardless of MY EPC.
SO basically i need to start paying for irrelevant ads in order to keep my campaigns going on google OR start paying 1-2.00 per click to meet quality score.
Isn't that extortion?
For example i my be able to sell "consumer brand x" but i can't sell "consumer brand x elite" - am i to now have to put in broad match and come up with 1000s of negative keywords to try and re-target how my exact match worked from the get-go so i don't end up buying traffic for "sucks" "Reviews" "repair" "hacks" and every other whacky variation i don't want to look/pay for?
Broad match on products is poison and a waste of $$ - is google basically telling me to hit the road or pay to play?
"We can't tell you WHY your QS is low but we can tell you that going to broad match we will increase your QS"
i haven't heard this one before. However, some broad matches in my accounts get incredibly good quality scores and some long tails get killed - Thats a fact (i hate it).
Quality score is so out of whack it's not even funny.
1)Try DKI if you haven't yet. May be better CTR will help bring your min cpc's down.
2)Try running broad match for a little while and if you get good quality score add the phrase match in that adgroup and next day pause the "broad match" keyword (this way you might get broad match qs passed to your phrase match).
3)In my experience min bids of $1+ are products of google not liking your landing page. (try isolating a few specific keywords with those bids and make manual ads (without DKI) put the keywords in the title and description, also put that keyword on the landing page and see what happens.
I can't guarantee that this steps will help, but in any case good luck to you. I feel your pain.
There are a number of write-ups on the topic of how Google encourages use of broad match to enhance their revenues:
David Rodnitzky, aka Blogation: [blogation.net...]
MediaPost SearchInsider:
[blogs.mediapost.com...]
SearchQuant (me):
[searchquant.blogspot.com...]
WebmasterWorld:
[webmasterworld.com...]
I was gonna say 'Hope that helps', but it sounds like you already know what you need to do.
-Shorebreak
Have you looked to see what percentage of your searches are by part number, as opposed to a keyword?
If nothing else, running broad match for a small bit and then running some reports might give you some healthy strings of negatives you can add in. Sure, it's a pain, but it's relatively easy with the AdWords Editor. And it's a HECK of a lot easier than the limited negative capabilities of MSN and Yahoo.
No warranty for this claim.
I think you've been told to do something simply because your conversation party had nothing else to say. I had cases where I would say something I know for sure and I get "Huh" from other side. They never heard about it.
I believe that you need to change the content of your lading pages and wait for adboot, no matter how good you think they are. You've been marked and that's it. They let the system decide. You may be able to ask for manual review, if you are getting hit by high minimum bids while providing good content and products that match search queries. Keep in mind that most of replies are that a specialist took a look into your site and agreed with what the system has determined.
In other words they're very DETAILED landing pages with full product specs, pictures, detailed pictures and tons of data about every product and related product. If its the landing page then any etailer using standardized product pages would be SOL.
I spoke two 5 different CS reps in the history of emails and ALL of them stated i should try broad match.
I guess in the short term broad match at the lowest bid will help me stem out negative keywords and still be cheaper than a single click on QS penalty by my oh my, what a terrible way to do business!
like telling a mountain dew addict, try drinking coffee, the caffeine in that soda ain't good for you...
something's wrong in this story to me, like maybe the reps misunderstood some point or something.
or something's wrong with the algorithm, like maybe it's built with a bias towards wide net casting over spear fishing... which would make G's claims about the importance of relevancy a big lie. I can't see that being the case, G risks everything if their ppc results are skewed away from max relevancy...
i hope to hear more on this, if it comes.
Anyway, there's one major issue with using model numbers like "11-1111-sk-v" as keywords: that model number is usually listed only once or twice on the product page. As a result, your landing page relevancy is assumed by the algorithm to be low, and you get a low QS.
If this campaign is important enough for you and you're able to tweak the content of your product pages, then focus on landing page optimization. The goal would be to make the algorithm realize that your page is relevant to the model number keyword. Make a bunch of test landing pages for a few products and see what kind of changes can improve the QS. There are the landing page guidelines [adwords.google.com], of course, but try some basic on-page SEO principles too -- sprinkle the keyword in a few more appropriate places in the body of the page, header tags, and/or the title and URL. The trick will be to balance the landing page tweaks with the user experience and the general business goals.
Good luck!
In my specific case the very targeted keyword is in 7-10 KEY places.
1. Keyword is in url string - url string is derived from the product_id
2. Keyword is in the product image alt text
3. Keyword is in multiple h tags of multiple levels - top heading, short description, long description, advanced features, model name and vendor link.
4. keyword is often in the ul areas multiple times.
I'm still working with support to see what i can do to increase the QS.. they said to look at my landing page this morning but they just seem to copy and paste info that you find in the adwords help vs offering any credible/logical fixes to my specific case.
The "product_id" which translates into the widget is just one of the keyword phrases i use.
For example i also use "brand widget type" type keywords where widget type is in dozens of places on the landing page and it too has a poor quality score.
As for the landing page itself, its pretty simple html/css design - nothing failing w3c checks, not terribly heavy in any javascript and whatever javascript there is, it doesn't impact the core description/content of the page but more enhance it for the end user (such as more images, reviews and 3rd party services that aren't relevant to my content but an added service (warranty/disclaimers.. blah blah)
Is QS derived from pagerank in any fashion? these products i'm pushing are new pages with 0 pr at the moment.
If you think about it that way, it becomes much easier to understand all the instances where their QS decision doesn't jibe with our own subjective (but reasonable) view of landing page quality.
Is QS derived from pagerank in any fashion?
Of course it is flawed. It was (I think) primarily a measure to stop arbitraging..We all know how successful that has been.
Legitimate ecomm sites get hit because sometimes, it may not be possible to fill a page with 'high quality content which is also unique'.. Might as well start a information site.
QS is also very very very unstable. If yo get a high QS today..The same keywords may get a 'poor' rating a few days down the line (even if you have a good history). You see,, they may and do just change the 'weights' in the algo any time they want.
This is how adWords is currently. Gotta live with it and find work arounds.
My 2 cents..
Try deleting the campaigns and start again. You may get lucky.
As others say..Try fewer keywords and try to add 'privacy policy', 'sample pages', 'contact us' links to your landing page. Also try putting some links to 'information sites' ( like wikipedia') to give the surfer a 'high quality experience' ;-)
Let broad match run with a low budget for atleast a month and a half. I find, the algo takes this long to finally start decresing bid prices and giving some decent traffic.
For example the product spec sheet for widget may have 3-5 large paragraphs with marketing details and then a very detailed spec sheet below it listing the widget parameters.
If QS is based on "unique" filter then adwords has no merit in ecommerce unless you sell a unique product that no one else has or custom develop your own descriptions for #*$!,#*$! products.
Oddly enough a product that hasn't been "Spec'd" that has a basic 5-6 UL list of the features and says "product will be updated but is available for purchase" gets a decent QS.. I just tested a batch of new "Cheap" stuff that i don't want to run PPC because there isn't any ROI in doing so but wanted to make sure i wasn't getting whacked for some reason.
So for widget Z that i JUST imported that is an el-cheapo product that is very common, hardly unique and has a terrible description with no values but 5 <ul>s describing basic features i get a QS of "good".
So i'm guessing el-cheapo widget is so commonly searched that the QS is "good" since a mediocre CTR yields enough $$ to google to "ALLOW" me to run an ad.
i'll try deleting everything and starting fresh and see if that helps. May just give up on the specific product targeting and work on brands where i carry the entire product line - at least with google since i don't want to handhold QS & expensive cpc's as much as it appears i'll have to being that targeted.
As RhinoFish said, if all is truth and you are really selling that stuff, something seems to be wrong here.
My understanding is that if a site is selling something in legitimate way, it should be let go through QS thing. There may be different minimum bids applied under Great and OK, but still, it should not really get stack by QS.
When the BIG QS “attack” happened back in summer 2006, I had a talk with my rep and asked something like (hard to recall exactly):
“If I have the worst and ugliest web site in the world which still technically functions perfectly right, and I sell the cheapest very popular blue widgets and have never given bad experience to a customer, why bother about QS, who cares about it, why stopping people getting to such site, if they are looking for what I have in exact match?”
From this, it comes out that the whole QS thing was created in order to expel those that Google did not like (for some reason), not really to focus on some (still undefined) quality.
Many experienced keywords being hit by $10 minimum bids while having long great history and CTR of over 10%. It’s neither CTR nor history, it’s that they did not want you there.
Google – Foogle (from fog)
Time to buy new spring shoes. [youtube.com]
I have approved Google Checkout buttons in adwords and google base. I do about 10k/month through google checkout as well.
Phone, Toll free, Email addresses & fax are on every page. An easy to find contact us for email & postal address is linked on every page in the menu and footer.
I'll ask about the QS in relation to ecommerce. Oddly enough the first time i complained about it they said i was under "manual review". Perhaps a competitor complained and got away with something? i dunno..
I wish you the best and hope you’ll get over, with or without their direct help.
Keyword Relevance:This keyword isn't highly relevant. Based on the keyword's relevance to the associated ad text, CTR, historical keyword usage, and other performance factors.
Landing Page: No problems found.
So my landing pages seem fine, but now its finally stating keyword relevance. I have several ad variants utilizing the keyword in the ad as well so i'm going to see if i can find something that won't trigger the QS.
i guess the ultimate test is to steal amazon or ebays cheasy ads saying "Buy now free shipping" with no mention of the keyword/product/variants and work from there.
leaving them active at low bids to see if they just get kicked in after the "Random" displays that they say happen to try and re-calculate qs.
Is QS derived from pagerank in any fashion?
No, not at all.
Rehan: what makes you say that? I'm quite convinced pagerank is an important factor in determining 'relevance' of the landing page. Just as it is important in the natural listings...
Sounds like a site quality issue. Have someone familiar with running ppc for multiple clients review your site for issues.
I'm quite convinced it's not. I have a site that I alone built and that I've been ppc'ing for some time and it's either some weird exception to your rule, proof your theory is wrong or shows my PR (and nearly non-existant seo positions) are being misunderstood by me. I know seo well enough to rule out the third case and I've got nothing to base a weird exception (case 1) on... so I'm going with your theory is wrong. :-)
However, if a site has good PR, it's got to be more likely to be a higher quality site than average, or seo is more out of touch that I believe. So it wouldn't surprise me to observe that high PR likely means no quality score whacking happens, but that doesn't mean a low PR means you will get whacked or even that PR is an important factor in QS.
Sounds like a site quality issue. Have someone familiar with running ppc for multiple clients review your site for issues.
The site always passes the QS review.. its the keywords getting poor QS.
Just for grins i created an add to a competitor site and its the same thing. I even copied a running ad verbatim and same thing. I think its just a penalty on a new account hitting me to tell you the truth. (mind you, i quickly deleted the ad, but it was the best test-case i could do.. Would amazon.com really pay 1.00 MIN CPC on an item that sells for 11.99?)
Rehan: what makes you say that? I'm quite convinced pagerank is an important factor in determining 'relevance' of the landing page. Just as it is important in the natural listings...
Byron wrote:
Just for grins i created an add to a competitor site and its the same thing. I even copied a running ad verbatim and same thing. I think its just a penalty on a new account hitting me to tell you the truth. (mind you, i quickly deleted the ad, but it was the best test-case i could do.. Would amazon.com really pay 1.00 MIN CPC on an item that sells for 11.99?)
However, if a site has good PR, it's got to be more likely to be a higher quality site than average, or seo is more out of touch that I believe. So it wouldn't surprise me to observe that high PR likely means no quality score whacking happens, but that doesn't mean a low PR means you will get whacked or even that PR is an important factor in QS.
that makes sense...
the reason though i said PR matters is because we've tested ads and keywords by pointing to various domains (not our own) which all had similar content on their site. I.e. we compiled a list of sites which all had a landing page for 'widgets'. Some of these sites were white labels, which assumedly are optimised in SEO terms.
What we found is that min CPC of our 'widget' keywords were strongly correlated with the pagerank of the domain we were sending the traffic to. Of course it could be that there's a dependent variable responsible for the result which made it appear as if pagerank was the main variable.
Perhaps the white labels were 'punished' by other means than through pagerank?
Kind of like Johnny failed both math and english. Two teachers graded him as performing poorly, each for their own reasons though.
And I do know for sure that a site with a very low PR can have a perfectly fine QS, so even if there was a direct relationship, it's got to be a very weak one.
I don't see how in the world that is possible. Setting keywords to broadmatch enables the "expanded broad match" to wreak havoc and show your ads for the most ridiculous of search queries...Just check your search query reports. Adding phrase and exact matching absolutely targets your ads more effectively...ESPECIALLY when you consider how "well" the expanded broad match works.
I don't think there is any other conclusion to make, other than this is a clear case of Google acting in their own interest to increase revenue, while burning us advertisers.
btw....Adwords Advisor usually does a great job helping us....but anyone find it strange (s)he sits these ones out?
btw....Adwords Advisor usually does a great job helping us....but anyone find it strange (s)he sits these ones out?
The reason for that is actually quite straightforward, BDuns.
Suggesting broad match to improve QS makes no sense to me - and I simply lacked the time to truly absorb this thread, find out who is offering that advice and why, and then compose a reply that made sense. ;)
Some may have noticed that I have been rather scarce in virtually all threads, not just this one: I'm still paying for having taken time off at the end of last year.
I spoke two 5 different CS reps in the history of emails and ALL of them stated i should try broad match.
I find that to be very surprising, ByronM - so much so that I was even tempted to sticky and ask to see those emails so that I could see who sent them & ask them why - or hope that the context of the emails might at least shed some light. In fact, time (and other responsibilities) have not allowed me to do this.
i believe you're telling the truth of what happened, but the reps replies smoke my brain's core... try broad to fix your qs issue. [...]something's wrong in this story to me, like maybe the reps misunderstood some point or something.
My head was nodding in agreement here, RhinoFish.
Maybe the Google rep suggested broad match so you could build up a history of impressions faster than if you used phrase or exact match. (Whether it would be a good history is debatable, though.)
My first thoughts were along these same lines, too, Rehan.
Bottom line, in my experience, switching to broad match will not increase Quality Score. And if an advertiser doesn't find that broad matched traffic is valuable to them, then I would certainly recommend against using broad match.
AWA
I'm going to take up on an offer to try out some ads on an aged account to see if they get the poor QS.
I didn't have a single broad match term prior to working with support so maybe support made that recommendation in haste - but it was still repeated in multiple emails to try.
Needless to save i've got some campaigns with a reasonable .50 cpc running that have anywhere from 2% to 20% ctr to see if that triggers something on the account as well.
