Question about CTR, rankings, and anti-SEO algorithms

[broccoli]

So I keep seeing these graphs that say if you're in position one you should expect a CTR of about thirty percent, position two about fifteen percent, and position three about ten percent (give or take a couple of percent). And there are these SEOs that say that above average CTR should lead to increased rankings. Is that actually true?

I've been studying my CTR and it's actually much higher than that for the positions I've been given in the serps, on both mobile and desktop, though mobile is a fraction of a percent behind. I'm at about twenty percent CTR for position three, and I've increased to about thirty percent for position 2.5. For the pages I have at position one I'm getting a forty percent CTR. Even on position 4-9 I'm getting a much higher CTR than predicted. I think it's either brand related or that I managed to get my titles and meta descriptions spot on.

What might cause an anomaly like this? Is it niche specific? Would there be a reason for my high CTR to not translate into better rankings? My bounce rate is less than 30% on most pages (though I'm still working on getting my homepage bounce down), so I don't think it's as simple as users hitting the back button.

In fact I changed the title tag on a couple of pages a few days ago, saw a boost in CTR, and Google responded by dropping my rankings on those pages.

Am I getting subjected to anti-SEO algorithms? I've been updating my site every couple of weeks as I improve it. Could they actually be punishing me for doing that? Does anyone here have experience of dealing with this?

[justpassing]

30%, 15%, 10%... Is that actually true?

It might eventually be the average observed at global scale, but to me it doesn't make sense to think like that.

30% of people (humans) will not mechanically click on the first result, while 70% will ignore it.

Also, a given number of people clicking on the first result (or any of them), will return to the SERP and visit other links too.

So yes, of-course the CTR you can observe for a particular page / keyword combination is highly depending of the niche, of the number of ads, news, videos showing before the organic results, or the accuracy of the results returned (you can rank first for a given keyword, but if the site title and snippet sounds off-topic, people will not click on it).

In fact I changed the title tag on a couple of pages a few days ago, saw a boost in CTR, and Google responded by dropping my rankings on those pages.

This is possible yes. Changing lot of titles at once, can be seen an attempt to over optimize them, or an attempt at reverse engineering Google's algorithm. So Google might not like it, yes. Long ago, it was possible to change elements of a page, and see the result in the SERP almost in real time (few hours and days), so, some abused of this, as always.
It's also possible your drop in ranking be linked to other factors, like algorithm changes, etc...

Be sure the titles of your page are not misleading are in line with the content of the pages. Don't use tricks to over optimize titles, they need to remain natural looking.

[cr1m]

I wouldn't take these numbers as something to strive for. These percentages are probably average across all queries and niches.

I just opened my GSC account and even across one website I have very different CTRs, one page in 1st place has a CTR of 27%, another one in 3rd place has 33%, one other that has an average position of 2.7 has a CTR of 57%. I've also seen CTR of 90+ for branded searches.

Same with bounce rate - AFAIK, Google has mentioned they don't look at bounce rate as a ranking factor, and that would be logical from my point of view, because the same website with most keywords in TOP 10 has a bounce rate of 75%. The reason is because it usually has answers to visitors' queries on the same page and they don't spend too much time on the website visiting other pages. If it was an e-commerce website, I'd be worried, but I don't believe these numbers really matter for SERPs.

[broccoli]

I didn't - I just added a sentiment word to two page titles to see if it would improve CTR, which it did. One of the pages seems to be moving back up now, which makes me even more suspicious!

[broccoli]

Sorry - the reply above was for @justpassing.

@cr1m - does that mean all the stuff about Rank Brain using Chrome user experience data is just rumour?

I’m having a hard time distinguishing what’s real from what’s rumour in the SEO community.

My instincts have been telling me they can’t be using it as a important ranking factor.

[EditorialGuy]

Same with bounce rate - AFAIK, Google has mentioned they don't look at bounce rate as a ranking factor, and that would be logical from my point of view, because the same website with most keywords in TOP 10 has a bounce rate of 75%.

Ezoic (a company that blends an ad platform with multivariate testing) has an interesting take on bounce rate:

"If a user lands on a page of your site, spends 10 minutes reading the entire article and then navigates away, Google Analytics counts this as a bounce."

Ezoic (like some other companies that track Web statistics) has a different standard: If someone arrives on your site and stays 30 seconds or more before leaving, "they found something that is at least mildly interesting and shouldn't be counted as a bounce."

This page has more on Ezoic's stance, along with links to other resources on the subject of bounce rate:
[support.ezoic.com...]

[broccoli]

Yes, it’s very hard to determine what a real bounce is. Even back button clicks aren’t a reliable way of measuring user happiness, I often visit several sites when I do a search.

I measure engagement rates by raising an event when users make a couple of selections in a form and click a submit button. Almost all pages have less than 25% bounce and some are as low as 11%. My home page is a bit higher and that’s something I’ll be working on. Average pages viewed is 1.8.

I definitely don’t trust Analytics - before I set up the button event it reported an average 80% bounce and 1:30 mins on site, and I felt pretty ashamed of my site as a result.

After I set up the event tracker I could actually see user behaviour and Analytics doubled the reported average time on site to three minutes!