Why I Hate AdWords’ ‘Ad Relevance’ & ‘Other Ad Triggering’: Simple Solutions To Combat Flawed Ad Serving

I have a confession to make… I’m not fond of Google AdWords’ Quality Score calculation, on a good day. From what I gather from digital marketers the world over, I’m not alone in my belief.

The idea itself is a great one—a way to filter and rank advertisers for similar goods and services in Google’s AdWords auction—so that the advertiser with the highest bid doesn’t automatically win each auction. It’s intended to effectively eliminate a monopoly on ad space that would be waged between deep-pocketed corporations (although some would say, with good authority, it too often is a battle of the titans).

So, I have no problem with it existing. I do, however, have a problem with the way it seems to be—almost randomly—calculated by Big G. Specifically, for this blog post, the ‘Ad Relevance’ metric.

Generally the ElementIQ blog is filled with useful, actionable material to help you or your client’s businesses improve online and get a leg up on the competition. It’s all about improving your ROI in digital marketing efforts.

That being said, there isn’t any rule saying we can’t air the occasional ‘complaint’ blog post. So please, if you’ll bear with me, hear me out and perhaps we can gripe together (and come up with some solutions?) on this rather frustrating issue.

All About Ad Relevance

At its very core, Ad Relevance is supposed to measure how relevant your ad text is to the keyword in question, with 3 possible grades (like all Quality Score metrics): Above Average, Average, and Below Average.

As per Google:

“This status describes how well your keyword matches the message in your ads. For example, if someone searches for your keyword and your ad shows up, would your ad seem directly relevant to their search?”

Fair enough, right? One would think, then, that any mention(s) of your keyword in your ad text should sufficiently earn you at least an ‘Average’ grade, right?

After all, if your customer is looking to hire an immigration lawyer and they type “immigration lawyers” in to Google, wouldn’t the following ad be quite relevant to their query:

  KM_Law_Ad_Example_Blurred

So, in this case, why would the keyword “immigration lawyers” have a ‘Below Average’ grade for its Ad Relevance metric? Is the ad irrelevant, or not very relevant, to the keyword in question, “immigration lawyers”? I mean, the keyword is actually in the ad’s headline, written exactly the same way, and the term “immigration” is used again in Description Line 1. What gives? Isn’t this worth at least an ‘Average’ grade?

According to AdWords, ‘Below Average’ grades for your keyword’s Ad Relevance are caused by:

“…your ad or keyword [not being] specific enough or that your ad group may cover too many topics.”

And can be remedied by:

“…creating tightly-themed ad groups by making sure that your ads are closely related to a smaller group of keywords.”

Now, in the example I’ve used above, that Immigration Lawyer ad group contains 14 keywords (every one of which has “immigration” in it and some variation after or before it, ie. “attorney, law firm, find” etc.)

So, isn’t it safe to assume the ad group is tightly-themed, and the keywords are closely related to the 1 text ad I shared here? I’m not including ‘criminal lawyer,’ ‘immigration laws’ or other unrelated terms in here, so what’s the issue?

Wouldn’t the ad seem “directly relevant to [the customer’s] search” as per Google’s definition of Ad Relevance? I think it would to the vast majority of rational human beings on this planet, anyways.

What complicates matters with this supposed ‘Ad Relevance,’ is ‘Average’ or better ratings given to keywords that DO NOT appear in the text ad at all!

Another client has a keyword “roof cleaning cost” in their Roof Cleaning ad group, while the ad that’s served 99% of the time with the keyword looks like this:

 NSHS_Ad_Example_Blurred

 The problem with this is, the keyword’s Ad Relevance has an ‘Average’ rating. Really?! There is nothing in that ad copy that mentions the cost of roof cleaning, or anything along the lines of getting an estimate or quantifying the cost of such services.

With these 2 examples in mind, it seems like the first advertiser is getting robbed in the Quality Score department, doesn’t it? With a ‘Below Average’ grade in any Quality Score metric, there’s no chance of your ad getting a 10/10 rating, and very unlikely it even achieves an 8. So, despite the apparent relevancy to the rational human eye, this keyword will suffer with a lower Quality Score. It will, therefore, rank lower in the ad auction, and accrue a higher average cost per click than those with higher Quality Scores for the same or similar keywords.

I could go on with other examples like this, but this blog post must get to a discussion.

Clearly, ‘Ad Relevance’ isn’t always calculated on the description that Google puts forth about it, or else these aforementioned keywords would have different grades.

Maybe we can use Google’s Ad Preview & Diagnosis tool (which you can find at the bottom of the pop-out that appears after you hover over the little speech bubble beside each keyword in your ad group) to get some answers? This tool is supposed to show you which ad is currently running for a given keyword, in a geographical area and device of your choice.

Try this in your account. It will infuriate you. I performed this task multiple times across different ad groups in a client account, and while it would show me Ad ‘A’ in the preview, when we performed an actual search using the exact same keyword in the exact same geographical area, we would be served Ad ‘B.’ Or Ad ‘C.’ There was no rhyme or reason for this.

And just to add salt in the wound, Google says when you use this tool:

“You’ll see the exact same results as a Google search…”

Utter hogwash. Try it for yourself and see if Google is telling the truth or not…

Google’s ‘Defective’ Ad Serving

Ad_Triggering_Sad

As an aside, when we performed manual searchers ourselves in our browser, sometimes we were served ads that weren’t even IN the ad group in question! We’d be served ads from a completely unrelated ad group, connected to their own keywords. This is related to another object of my ire (and arguably an even worse offender): the dreaded “this keyword is triggering other ads with a similar keyword…” message in the keyword’s speech bubble in AdWords.

 Safe_Step_ad_triggered_by_other_keyword_example

 This is maddening. We create ads in a specific ad group for a reason, no? But in the case above, the keyword “walk in tubs” is triggering an ad from the ‘Safety Tubs’ ad group, which is associated with the keyword ‘safety tubs,’ a completely different term. Yes, ‘tubs’ is in each of them, but each keyword resides in its own, unique ad group that contains only similar keywords (ie. the word ‘safety’ does not appear in the Walk In Tubs ad group, and ‘walk in’ doesn’t appear in the Safety Tubs group).

Again, we create ads in a specific ad group for a reason, no? Apparently not… Google gets to decide this for you, judging by the following from their help page on this problem:

“This keyword is already associated with at least one other ad in your account.
You can only have one ad showing per keyword. If you have two or more ads eligible for the same keyword, the ad with the highest Quality Score will show.”

Thank you Google, for deciding which other ads (not in the same ad group) to match the keyword in question to. And, didn’t they just contradict themselves in the paragraph above? It says that the keyword is ‘associated’ with multiple ads, but then in the very next sentence, says that only 1 ad can show per keyword. Huh?

So why then, Google, did you decide to ‘associate’ another ad with the same keyword, despite saying only 1 ad can show for it? That’s why you made US make an ad(s) in the ad group for the specific keyword in question. Match it to any of those ads, not an ad that is supposed to be triggered by an entirely different keyword residing in another ad group!

What’s more, ads don’t actually have Quality Scores assigned to them: keywords do. So this just adds to the confusion. Worse still, I’ve seen an example of this in a client’s account—where the ad that ended up showing for the keyword in question, was attached to a keyword in a separate ad group—however, it had a LOWER Quality Score! It’s enough to make your head spin!

Take the above example, the keyword “walk in tubs.” When I went to look at the offending keyword that was ‘stealing’ the ad impression (safety tubs), that keyword was triggering other ads with a similar keyword TOO!

 Safe_Step_ad_triggered_by_other_keyword_example_2

 This time, it was “walk in bathtubs for seniors” that was stealing the ad impression; showing an ad related to the seniors term, instead of an ad related to the safety tubs term, which is what Google should have shown, for relevance.

 Safe_Step_ad_triggered_by_other_keyword_example_3

 Can you see the madness of this ‘cross-pollination’ of keywords and ads? Mind-numbing, if you ask me.

And no, I’m not done with the silliness… turns out, the keyword “walk in bathtubs for seniors” had a much lower Quality Score than the keyword “safety tubs,” so that flies right in the face of what Google told us in its multi-ad explanation. And I quote again:

“You can only have one ad showing per keyword. If you have two or more ads eligible for the same keyword, the ad with the highest Quality Score will show.”

So why would the keyword with the lower Quality Score trigger an ad in its ad group, despite it having a lower Quality Score and a less relevant ad, as well? Doesn’t AdWords pride itself on serving the most relevant ads at the right times? With all its incredible calculations and algorithms running in the background, how doesn’t it know to serve the most relevant ad to the searcher’s query, AND to follow its own rules, as per the quoted text above?

And finally, for the cherry on top of a very sour dessert, this ‘cross-pollination’ can change every single day. I checked the next day to see if this ‘other ad triggering’ was still occurring for the original keyword in question (“walk in tubs”), and indeed it was; but this time, completely different keywords were causing the wrong ad serving! At this point you just have to shake your head and realize you’re ‘playing the game.’

Sometimes there’s no justice in the AdWords world.

Are We Being Duped… By The Same System, We Have To Trust?

Is it also possible that the keyword with the lower Quality Score is being used in the ad auction itself (instead of the original keyword that should have done this)? If so, this could be real detrimental to our ‘correct’ keyword’s Ad Rank (which is calculated as Quality Score x Keyword Bid), since the lower the Quality Score, the lower the Ad Rank, and therefore, the lower the position of the ad on the search results page, and the HIGHER the cost we’d need to pay (keywords with higher Quality Scores ultimately receive ‘discounts’ in the auction, meaning the advertiser pays less and less with each numerical improvement). There’s an incredibly comprehensive look at the economics of Quality Score here.

As far as we know, the original, ‘correct’ keyword is the one receiving the impressions and other statistics for the ads being shown (even though the ad was from a different ad group which is supposed to be triggered from a completely different keyword). However, it does make you wonder, since the ad that showed is supposed to only be triggered by the keywords in its own ad group.

Sigh…

AdWord’s #1 Flaw That Humans (Not Algorithms) Can Solve

These 2 problems, the ‘Ad Relevance’ conundrum, and the ‘cross-pollination’ of keywords to ad groups (which no one can easily explain away), point to what I feel is a glaring weakness in the AdWords system:

Human Beings Are Not Allowed To Choose Which Ads Run With Which Keywords!

Again, I salute Google for creating an advertising vehicle that I couldn’t even begin to fathom how to create, and for the most part it is incredibly intelligent and useful. However, such a fundamental element of advertising is relevancy, and I don’t care what they say, if you have a marketing team or copywriter with at least half a brain, they’d be able to point out which ad copy would be the most relevant to the keywords triggered by a user’s search queries.

Google, why not let us decide which ads to run with which keywords? Most of us know our target audiences, and this way, we’ll avoid being at your mercy, and avoid potentially showing ads for “walk in bathtubs for seniors,” when “safety tubs” is what the search was originally for. Yes, maybe the ads are similar, but we made the ads in the “Safety Tubs” ad group for a reason, and the ads in “Seniors” for another.

In this way, we could still rely on AdWord’s ‘Ad Rotation’ settings to ensure the best ad in the ad group is still shown; simply choosing ‘optimize for clicks’ or ‘optimize for conversions’ will still allow Google to choose the appropriate ad—it will just choose between those in the given ad group… not from unrelated ad groups that potentially contain keywords with lower Quality Scores!

Is this asking too much? I’m sure Google likes the status quo, so it has control, and because they pride themselves on serving the most relevant ad, as per their electronic algorithm’s brain. Plus, it is possible that they’ll serve ads that may historically cost more than others (for all you conspiracy theorists out there). Again, we can’t see the inner workings, so we’re more or less at the algorithm’s mercy.

Common Sense Solutions To ‘Lower Your Risk’

What can we do to combat these 2 problems, Ad Relevancy and ‘cross-pollination’ of ads and keywords?

For Ad Relevance, love it or hate it, all we can do is:

  • Create tightly-themed ad groups like Google suggests. Keep it to 15 keywords or under if possible
  • Use a combination of broad, “phrase,” and [exact match] keywords, particularly [exact match] if possible, since these are the most specific and Google finds these super relevant to the searcher’s query, when it’s (obviously) what the searcher typed in
  • Diligently use negative keywords in your campaigns (and specifically at the ad group level)
  • Create ad copy that contains the keyword
  • Use {dynamic keyword insertion}. However, don’t overuse it!

To avoid ‘cross-pollinating’ ads and keywords:

  • Use [exact match] negative keywords at the ad group level. For my example above, we would use the keyword [walk in tubs] as an exact match, negative keyword at the ad group level of the “Safety Tubs” ad group.  Just like we would use the [safety tubs] exact match keyword as a negative in the “Seniors” ad group. So, wherever possible, use the ‘offending’ keyword stealing the ad impression as an [exact match negative keyword] in the ad group level of the correct ad group. Don’t forget to do the same for plurals and common misspellings as well! This can be very tricky with match type.
  • Create unique, tightly-themed ad groups, but don’t go so granular that AdWords sees little difference between them; ie. though it’s no guarantee, don’t create an ad group for ‘purple shoes’ and ‘flowery purple shoes’ just because of a 1 word difference, even if you do use [exact match negatives] and single word broad negatives at the ad group level.
  • Ensure your landing page is as closely related to the ad group’s keywords as possible. This doesn’t mean keyword-stuff your page, but be cognizant about what the overall theme is for the ad group, and ensure your landing page serves up relevant content around potential search queries that match to these keywords. Don’t forget, ‘Landing Page Experience’ is an important element of Quality Score as well!

Wrapping It All Up

Do you suffer from these same kinds of AdWords challenges? Is the algorithm going ‘rogue’ on your keywords and choosing what it thinks is the best ad for the searcher? If so, I hope you’re able to see some improvement in your accounts by following these guidelines.

AdWords will test your patience, but as a whole, it does do a good job of managing your campaigns. As long as you can recognize any problems and then know how to act accordingly, you’ll continue to see respectable results.

Do you have any suggestions to combat this unfortunately ‘inhuman’ aspect of the AdWords engine? Any personal success stories or valuable advice you can share with our readers? If so, please do tell us about them in the comments below!

I’ll see you there!