Quality Score Factors
Only historical CTR really matters

11 Flares Facebook 0 Twitter 1 Google+ 7 LinkedIn 2 11 Flares ×

Quality Score Factors: Only historical CTR make it soar!

Relevance holds it together, it has to be tight. Historical click-through-rate make it soar and landing page quality is the needle that can burst it in an instant. That’s the quick answer to the following question:

What are the factors that influence Quality Score?

Last post we talked about the impact of high quality scores and how it led to 5 times more conversions for the same keyword with no change to a landing page I had to deal with.  To get those kinds of results you need to know what influences it and focus on what matters – relentlessly – with no distractions.

Let’s talk about those distractions before we talk about what works.

Don’t waste your time on the following…

  • Bounce rate doesn’t impact QS (Here’s Proof)
    Although it makes a lot of sense that it would… it’s doesn’t. And Google sometimes floats the idea that it could… it doesn’t.  It may in the future, but as Dec 4 2012, it does NOT. Many of my LPs have high bounce rates and I don’t care so long as they convert.
  • Keyword density doesn’t impact QS
    Wether it’s on your LP or the ad itself, it doesn’t matter. Your ad and LP should talk about what you’re selling but it doesn’t matter that the keyword is there once, twice, a thousand times (unless it increases CTR). In fact the keyword  doesn’t even necessarily have to be in the ad nor the LP so long as everything relates to what you’re selling.
  • Having lots of content on the website does nothing to QS
    You can have one sentence and still have great QS.
  • SEO rankings, back links have no effect on QS
    It has no impact, it’s completely unrelated.
  • Anything that isn’t mentioned below
    You’re doing good, keep on reading…

If you ever hear someone somewhere on the web talk about the above factors having any influence on QS, tell them they’re wrong. If they ask you, tell them I said so. I don’t care if it’s someone at Google. I listened to everyone until I learned what works the hard way. I only trust my testing now.

Disregard my advice at your own peril.

On a practical level, here are the “ONLY” factors I ever worry about:

  1. Landing Page  Quality (mostly transparency and no shady stuff)
  2. Semantic Relevance (mostly ad grouping)
  3. Historical CTRs (ad copywriting)

That’s it. The order in which I put them is important. I always make sure my landing pages are good before I do any ad grouping. Then I take care of ad grouping correctly before focusing on ad writing. Then ultimately all I do is writing new ads — tirelessly. And I don’t ever stop writing new ads. Occasionally I will do some regrouping in the life of an ad group but 90% of the work is split-testing new ads, and I’ll teach you how to do that efficiently in the next post.

Here’s what happens invariably: my Quality Scores go up then I can focus on my conversion rates (which is another thing I’m obsessed about) and everything else that make me money.

Yes, there are other factors (like page load speed, etc) that influence QS but in most cases, the above 3 are all you need to care about. In fact the 1st one is a set it and forget it type of thing. So it’s really easier than you think: group your keywords tightly, then write irresistible ads.

Most people drop the ball at “write irrestible ads” part then complain that their quality scores suck. Luckily you’re a Tenscores reader and soon enough you’ll be performing better than your competitors consistently. Google will reward you for it.

I. Landing Page Quality

An on and off switch it is. Still is. The needle that can burst the balloon. Don’t ever spend too much time on landing page quality, you either have it or you don’t. If you don’t you’ll know it by how low your QS is, usually 1/10. I don’t want to spend too much time on this one, just follow these steps and you’ll be fine.

What about the recent update Google made that got everyone talking? I don’t care about it. Didn’t affect me nor the majority of our customers, and we have software that tracks all that stuff. (Actually, there’s a more recent update or bug that happened on Nov 7, mostly went under the radar).

I always optimize my landing pages for conversion rates, almost NEVER for Quality Score.

II. Semantic Relevance

Almost always, this is where I’ll start. Google looks at how relevant an ad is to the keywords in it’s ad group and how relevant a keyword is to a search query (keywords and search queries are not the same thing).

A lot can be said about relevance. Here’s the trick to having it right:

Rule of thumb #1: Group keywords by ad content not keywords.

There’s a reason Ad Groups are called “AD” groups, not “KEYWORD” groups. What you’re probably creating right now are keyword groups. That’s an imperfect way of doing things. You take a list of keywords, group them by themes then write ads for them. That’s a mistake. I’m often guilty of making this mistake myself.

The correct way to group keywords is:

Step 1: Pick a keyword
Step 2: Write the best possible ad you can write for that keyword
Step 3: Find other keywords that match the ad

(There’s actually an even better way to do this, but I’ll keep that for myself for now. Tenscores users will have it when we introduce it as an algorithm soon. We gotta show love to our users.)

Usually, you’ll end up with no more than 10 keywords in an ad group.

I consider ad grouping to be an art, it seems straight forward, but most people screw it up. Keep refining your ad groups over and over again. We’ll talk more about this, if you have specific questions about an ad group you want to create, write it in the comments and I’ll help you out.

III. Historical Click-Through-Rates

Bar far the most important factor that influences QS. In fact, when you’ve taken care of everything else, you can consider your quality score to be a direct measure of your CTR compared against your competitors’. Low QS will indicate that your CTRs are lower than those achieved by other advertisers in your space — even when you think they’re great. What you may consider to be good CTR may actually not be. Keep that in mind.

Google evaluates your CTR at all levels:

  • Individual keyword CTR
  • Display URL CTR
  • Ad group CTR
  • Account CTR
  • Keyword CTR by geographical areas
  • CTR by placement (display network website)
  • CTR by device you’re targeting
  • And more CTRs that we probably don’t know

In short, it’s all about CTR.  Anywhere there’s a CTR, it matters.

Anybody who tells you different doesn’t know what they’re talking about… or they just read some articles from people who live off writing articles. (Sorry if I hurt your feelings — deal with it!)

Google is always trying to predict what CTR you’ll achieve before even showing your ad.

They will use your recent historical CTR data of the keyword in your account, if there is any (whether it is paused or in another ad group if it’s there they’ll use it). If there’s no history for the keyword, they might use the ad groups history. If the ad group has no history, they might use the campaign’s history. If the campaign has no history, they might use your whole account’s history.

And if they don’t use any of those, they will most certainly use the average history that other advertisers in your space are getting. If it’s poor, you’ll get poor QS and you’ll need to prove to Google that you’re above the herd. That’s why some industries are said to have low scores across the board. If you believe to be in such an industry, I guarantee you someone in that industry is silently killing it, while you’ll be accepting your fate. Don’t do that, keep pushing. Until you rise above the herd.

So… if you’ve had great CTRs recently on a keyword, Google is quite confident that you’ll continue to have great CTRs and  you’ll be awarded high QS and lower CPCs. On the other hand, of you’ve proven to be a poor performer, Google will predict that you’ll continue to be a poor performer and give you low QS.

My definition of  ”recent history” (emphasis on my) is: within the last hundred impressions. That’s my second rule of thumbs:

Rule of thumb #2: Can’t get decent CTR within first 100 impressions on search ad? Pause ad, write new one.

Your QS will go down very fast if within a few couple hundred impressions your ads suddenly show poor CTRs.

What’s a good CTR?

Glad you asked…

It depends on your competitors. You need to outperform them. Which is sometimes not possible (example: it’s very hard to outperform a competitor on his brand terms). But, my third rule of thumb is:

Rule of thumb #3: For search ads, anything below 1% is unacceptable.

For display ads it’s another story (some people consider an 0.08% to be good enough). There’s currently no visible QS on display ads (there’s a hidden one) so we won’t dwell too much on them. What’s important is that CTR is as important.

Conclusion

Here’s what I would like you to get out of this post. Whenever you’re going to optimize QS, do this:

  1. Check that LP is fine (unless QS is 1/10, usually you’re fine). Then forget about it.
  2. Regroup your ad group. 90% of the time you’ll have this done half-right, usually it isn’t perfect. Do it once more. Then forget about it. (Having one keyword per ad group isn’t perfect either, keywords strengthen each other, so it seems)
  3.  Focus on CTR for the rest of the ad groups life. I guarantee you, you know nothing about the people clicking your ads, that’s what a low QS is telling you.

So now the question remains …

How the hell do I get better CTRs than my damned competitors?

Ahhhh… Now we get to the really fun stuff. I’ll tell you about a little trick that almost everyone in our industry says is unecessary… but that I use all the time… because it works.

You’re gonna LOVE it!

Stay tuned…

About Chris

Co-founder of Tenscores. Online advertising druid experimenting daily, learning by the minute.
This entry was posted in Quality Score. Bookmark the permalink.
Don't be a sucker!
Learn how to get more and cheaper AdWords clicks.
  • Andrew

    Hi Chris,

    Love the post. Some very interesting stuff to try out. Maybe times are changing but in the past we have had good success with single keyword Ad. Groups e.g. consistent double digit CTRs. Also using a technique you previously shared. Looking forward to the next installment.

    • http://tenscores.com/ Chris

      Hi Andrew, I used to use single keyword ad groups very often in the past. Was one of the very first thing I ever tried and I thought it to be the best way to do things for a long time. It got me lots of 7s. Having 3-5 tight keywords in the same ad group however, would take many of my keywords to 10s. Since then, it’s been my firm belief that keywords strengthen each other in an ad group. Great point. Thanks.

      • Asif

        Thanks Chris. But how do you use those keywords? Broad, phrase or exact?

        • http://tenscores.com/ Chris

          All three (four) match types with decreasing bids, exact having the highest bid and broad having the lowest. In new ad groups I always start with exact match then add phrase or broad later on when I’m happy with results. On low search volume keywords I might start immediately with modified broad match. I’ll soon write a blog post about all that.

  • mbogosi

    Chris – loved the post. Is there any reason to have 20+ keywords in an ad group in your opinion. After reading this post, I’m giving some real thought to changing my entire philosophy/approach to building out campaigns.

    • http://tenscores.com/ Chris

      You might have more than 20 keywords in an ad group if you’re an expert and you really know what you’re doing for a specific client. For example, you’re dealing with millions of keywords in a single account and you have a particular structure for managing them… The average advertiser never finds himself in such a situation.