AdWords is now Google Ads. Our new name reflects the full range of advertising options we offer across Search, Display, YouTube, and more. Learn more

Ads
4.1K members online now
4.1K members online now
Improve your Google Ads performance and boost your ROI, CTR, and Quality Score
Guide Me
star_border
Reply

A hypothesis about why my QS numbers are bad.

Follower ✭ ✭ ☆
# 1
Follower ✭ ✭ ☆

A week or two ago, I posted a question about why my quality scores were dropping.  Lots of people made suggestions, and I must have spent a man-month trying them all, but none of them had any effect.  I was at a loss.  Since then, I think I've figured it out.

Now, everything in this post is hypothesis.  But I've been staring at this thing for a month, so I think I might be onto something.

Now, Imagine for a moment that you have a store that sells paint.  You can mix any color.  You use only one keyword: "paint."  That leads to an ad that says "paint! We sell paint, in any color."  Imagine that the landing page is all about "paint."

Now, imagine for a moment that google users *never* type the word "paint" by itself.  Instead, they type:

"burnt sienna paint"
"sunset orange paint"
"lilac lavender paint"

Every single "paint" query contains some esoteric color name.  Of course, you don't care what color they're searching for.  You've got one of those paint-mixing machines, you can synthesize any color at all, so it obviously doesn't matter.  So you just match the word "paint," and you think you're good.  But you're not.

Here's what I think happens.  Somebody types the query:

"burnt sienna paint"

Google sees this query and decides that the most interesting, distinctive keyword in the query is "sienna."  People rarely use that word, it's a very distinctive word, so therefore, google thinks it must be very important.  It then looks at your ad: "paint!  We sell paint, in any color."  It sees that your ad doesn't contain the word "sienna."  It doesn't contain the word "burnt" either.  So it concludes that your ad is not relevant to this particular query.

Then, somebody types the query:

"sunset orange paint"

This time, google sees the word "sunset" and decides that's the most important word in the query.  It looks at your ad: "paint! We sell paint, in any color."  And it says to itself, no, the ad doesn't contain the word "sunset."  Once again, it concludes that your ad is not relevant to this particular query.

So over and over, google compares your ad to somebody's query:

"midnight black paint": poor keyword match on "midnight"
"marigold amber paint": poor keyword match on "marigold"

Each time it fails to match, your QS drifts downward a little, and eventually, the Keyword Relevance: Poor light turns on in the dashboard.

Now, as I said: this is only hypothesis.  I've only been doing this a month.  But I've been staring at this thing constantly for a month this is the only thing that makes any sense.

4 Expert replyverified_user

Re: A hypothesis about why my QS numbers are bad.

Top Contributor
# 2
Top Contributor

jyelon;

You are a little bit under estimating AdWords capabilities....Smiley Wink

There is a solution, that any color the user types into the query, will be displayed in the ad text. It's called: "dynamic keywords insertion". Read more on the technic in the help center.

 

-Moshe

Moshe, AdWords Top Contributor , Twitter | Linkedin | Community Profile | Ad-Globe
Did you find any helpful responses or answers to your query? If yes, please mark it as the ‘Best Answer’

Re: A hypothesis about why my QS numbers are bad.

Top Contributor
# 3
Top Contributor

Hi,

 

So I just went back an read the other thread you mentioned.  There are a lot of superstars in there that gave you some very sound advise.

 

So on the hypothesis, let me get this straight...

 

X-company sells Y-Product

Y-Product comes in infinite variations of color

Y-Product is the only keyword in the campaign

Y-Product as a keyword is never searched

Honestly, Who would do that?

 

To think you've deciphered an algorithm because you've "been at this a month" is reaching my friend.  At a month, I was still stumbling around with conversion code and match types.

 

If you have questions, by all means post them.  As you saw in your last post, there are many willing and able participants here that take quite a bit of time to share proven and tested methodology along with the general practices of Adwords.

 

Take that advise that was presented to you, implement a strategy and give it some time to run its course. Search behaviors change from hour to hour, day to day, month to month...etc. If I'm not mistaken, you only have a month or two total history so far. In my experience and opinion, when starting out, you should really let a campaign run for a few weeks or more, then analyze, make a some changes and let it go again for another few weeks. A strategic step by step process will get you to where you want to be.  It just takes time...and while you're waiting, you can read up on potential new strategies and more advanced tools available in adwords.

 

Why time? You need to be able to compare what you've done to the results and a few days or even a week is just not long enough to determine very much. 

 

Your passion and willingness to dig in are great, Adwords is a great tool to learn.    

 

Good Luck!

 

-Tom 

Tommy Sands, AdWords Top Contributor | Community Profile | Twitter | Philly Marketing Labs
Did you find any helpful responses or answers to your query? If yes, please mark it as the ‘Best Answer.’

Re: A hypothesis about why my QS numbers are bad.

Follower ✭ ✭ ☆
# 4
Follower ✭ ✭ ☆

Moshe: I considered dynamic keyword insertion.  If I understand it correctly, the way to use it would be to put a whole bunch of keywords like this

 

+burnt +sienna +paint

+lilac +lavender +paint

 

Now the question is: does that work when the individual colors are low search volume?  Eg, let's suppose that only a tiny, tiny percentage of people searching for paint are searching for burnt sienna paint.  In that case, the keyword +paint by itself would have plenty of search volume, but +burnt +sienna +paint would be low search volume.  Can I still insert the keywords +burnt +sienna +paint back into the ad, *even though* that keyword was inactive because of low search volume?

 

 

 

 

Re: A hypothesis about why my QS numbers are bad.

[ Edited ]
Follower ✭ ✭ ☆
# 5
Follower ✭ ✭ ☆

No, you've misunderstood the problem:

 

X-company sells Y-Product

Y-Product comes in infinite variations of color

Y-Product is the only keyword in the campaign

Y-Product as a keyword is never searched BY ITSELF - the query is always Y-Product (eg, "paint") plus some other words (eg, "burnt sienna paint").

 

So in the case of my cake shop, yes, there are a small percentage of people who actually search for the word "custom cake" all by itself, but the vast majority search for:

 

custom piano keyboard cake

custom big ben clock tower cake

custom animal muppet cake

 

The DKI could work as long as I can guess in advance that somebody might want a sculpture of big ben the clock tower, and I create a keyword for that.

 

Re: A hypothesis about why my QS numbers are bad.

Top Contributor
# 6
Top Contributor

Hi jyelon, as Moshe has said, you are under-estimating the complexities and capabilities of the Adwords system (and Google search in general).  Google doesn't just match words dumbly, it attempts to understand the query being made and present the most likely relevant results.  Your hypothesis suggests that for "burnt sienna paint", Google would consider the word "sienna" to be the most important - did you actually try searching for this phrase?

 

As soon as you get as far as "burnt si" Google is suggesting "burnt sienna" as the most likely match and burnt sienna as a colour is in the results.  Go to include the word paint and all the results are for the colour and paint manufacturers/suppliers.

 

OK, let's move the words around for no apparent reason and start with "sienna".   As I expected, when you put this word in, the top results are for the British actress Sienna Miller, but as soon as you add the word "paint", the results change to paint content and indeed there's even an Adwords ad for Dulux Sienna paint on the top right.

 

In short, as long as your key product - whatever it's shape, size or colour, is a well known commodity/service, Google will pick up on that, not the shape/size/colour.  Try it yourself!  I've just typed "piano" and got results for the instrument.  I add the word "cake" and I get results for piano shaped cakes.

 

So how can you organise your keywords and get to use DKI?  For cakes I'd probably start with +custom +cake and variations thereof (personalised, designer, etc.).  Monitor the results over time and I'm sure you'd start to see key designs (piano, muppet, etc.) appearing frequently.  You can then decide to use key phrase or even [exact] matches in a separate group that uses DKI based ads.

 

This can't happen overnight, it'll take time, which is why I always recommend starting with a smallish budget and monitoring closely over a period to tune before increasing the budget.

 

Finally, the bottom line is your Return on Investment, rather than individual metrics like QS.  If Adwords makes you a profit - or even a small loss that's acceptable because of referred business - then you can be happy and work over a longer period to increase that net profit by tuning the account.   I'd also say you should always bear in mind that it's possible Adwords simply won't work well for your business.  There's no magic solution for all products or services and for some it simply isn't cost effective.

 

I hope this helps.

 

Jon

AdWords Top Contributor Google+ Profile | Partner Profile | AdWords Audits

Re: A hypothesis about why my QS numbers are bad.

Top Contributor
# 7
Top Contributor

jyelon;

I started to type my answer, while noticing that Jon has explained in detailed the complexity of both Google search and AdWords. I would just add that, if those were that simple, they would not be so popular among users and advertisers.

I'm all "me too" with Jon's answer (+ a star Smiley Happy)

 

 

-Moshe

 

Moshe, AdWords Top Contributor , Twitter | Linkedin | Community Profile | Ad-Globe
Did you find any helpful responses or answers to your query? If yes, please mark it as the ‘Best Answer’

Re: A hypothesis about why my QS numbers are bad.

Follower ✭ ✭ ☆
# 8
Follower ✭ ✭ ☆

Jon: you're saying that google is very good at determining keyword relevance.  If that's true, then perhaps when google says "Keyword Relevance: Poor," then perhaps it's correct?  Maybe there's some sense in which

 

+pittsburgh +cake

 

and this ad:

 

Custom Cakes - Pittsburgh

Display a Beautiful Cake Sculpture.

Unique Cakes in any Shape or Style.

 

Is not relevant to the queries that users are actually typing?  I've seriously considered this possibility.  For example, consider this query:

 

"tickets to cake in pittsburgh"

 

Cake is a band!  For all I know, they could be touring, the could be coming to Pittsburgh!  My ad *isnt* relevant to that!  This insight just came to me spontaneously a few weeks ago.  As soon as it did, I put in negative keywords for band, tickets, concert, etc.

 

But maybe, just maybe, there's some *third* meaning of cake - not the dessert, not the band, but something else that I haven't thought of.  Maybe cake is a brand of electric razor?  Maybe people are trying to find a sharpening service for cake razors in pittsburgh?

 

In the end, I'm a little skeptical of this explanation.  I really do think this ad is relevant to most of the queries out there --- as demonstrated by the fact that people are clicking on it like crazy.  So when google says "keyword relevance: poor," I have to be a little skeptical at this point.

 

Re: A hypothesis about why my QS numbers are bad.

Follower ✭ ✭ ☆
# 9
Follower ✭ ✭ ☆

Actually, as long as I'm on the subject...

 

the other day, somebody came to my custom cake website after typing this query:

 

chocolate cake pittsburgh

 

That is probably not somebody looking for a custom cake sculpture.  That is probably somebody looking for dessert.   As soon as I saw this, I put in negative keywords for chocolate, pound cake, angel food cake, etc.  It's not that we don't make those types of cakes - we do.  It's just that somebody who is asking first about flavor is probably just looking for a meal, not for a sculpture.  So that really is a "keyword relevance: poor" problem.  My ad *really* doesn't match that query.

 

Same thing happened with this query:

 

gluten-free cake pittsburgh

 

Once again, that's somebody searching for something to eat, not for a custom cake sculpture.  So that, too is a "keyword relevance: poor" problem.

 

But the interesting thing is: my campaign was running for a month before some chocolate-cake searcher accidentally clicked on my ad.  If he hadn't accidentally done that, my ad would still be showing my ad, inappropriately, to the chocolate-cake and gluten-free cake searchers. 

 

How do you guys go about catching these mistakes?  For all I know, my campaign could still have an actual relevance problem - not with cake the band, or cake brand razors, but with actual cake.

Re: A hypothesis about why my QS numbers are bad.

Top Contributor
# 10
Top Contributor

@jyelon;

We do that, exactly as you described: We run a query report (i.e figuring which keywords trigger which ad), analysing them, and add negative keywords, based on queries that include keywords that are irrelevant to our business.

 

-Moshe

Moshe, AdWords Top Contributor , Twitter | Linkedin | Community Profile | Ad-Globe
Did you find any helpful responses or answers to your query? If yes, please mark it as the ‘Best Answer’