The Smoking Gun: Deleted Google Author Photos Boost Ad CTR | Wordstream

You are here

The Smoking Gun: Deleted Google Author Photos Boost Ad CTR

July 10, 2014
27

If you’ve been following developments in the search industry lately, you’ll have undoubtedly have come across discussions about Google’s recent decision to remove authorship photos from search results. We reported on this soon after the change went into effect, and other industry thought leaders like Rand Fishkin wasted no time weighing in, either.

Removal of Google authorship photos impacts paid search ad CTR

However, despite overwhelming support for the idea that Google’s decision was based on authorship photos’ impact on the CTR of paid search ads, some naysayers insisted that this couldn’t be the case. I wasn’t content to leave things at that, so I set about looking for evidence to support my theory – and I’ve found it.

Google authorship photos removed negative keywords serp

In the figure above, you can see that in a search for the term “negative keywords”, our paid search ad is the only ad displayed. This eliminates any possible ambiguity resulting from the variance in CTR based on ad position. You’ll also notice that two articles authored by me are included prominently in the Google search page, as well as another article authored by Darian Schouten over at TechWyse. These three results would have been accompanied by author photos prior to their removal on June 25.

We examined the CTR of the ad above both before and after Google’s announcement. We found hard evidence that the CTR of the ad improved significantly when author photos were no longer being displayed in the SERP.

Removal of Google authorship photos impacts CTR of paid search ads graph

Obviously, the peaks and valleys shown in the figure above are normal. CTR varies from one day to the next, so these fluctuations should come as no surprise.

As is clearly evident from this data, the CTR of the ad after the removal of author photos from the SERPs is much higher. In fact, the CTR of the ad is 44.8% higher than beforehand within this ad group. We tested this data rigorously, and the difference we observed is statistically significant with 99% confidence due to the high number of daily ad impressions (thousands) for this keyword.

We would love to test this theory with other keywords, but finding suitable terms that both meet the criteria and for which we have data for, is challenging.

However, it’s clear to us that based on this data, it’s not realistic to say the deletion of Google authorship photos has no impact on the CTR of other elements on the SERP.

Comments

Hashim Warren
Jul 10, 2014

I called Google out on the lie right in the original thread. Thanks for putting data behind it

Feroz
Jul 11, 2014

Nice post backed by statistics 

Mark Traphagen
Jul 11, 2014

I am here neither to try to support or debunk that ad revenues were a factor in the author photo removal, but rather to caustion that a one off example that has not been reproduced hardly constitutes proof. A careful data scientist should note that. You have not controlled to eliminate other factors that could cause a sudden CTR jump, and therefore have established nothing more than an interesting coincidence.

Larry Kim
Jul 11, 2014

Hi Mark. Thanks for commenting here, and also on LinkedIn, Inbound and Google+.I agree with you when you say that it is very important to control and eliminate other factors. That is exactly what I and our internal data scientist have done here - we set out to control all of the variables except for the removal of author photos. To do this, we isolated a single, high-volume keyword where there was only one ad listing, thereby eliminating any change in ad CTR due to changes in ad position. The rest of the SERP results didn't change at all.Furthermore, by using ad Click Through Rate (as opposed to studying traffic to a page) I've normalized by the number of impressions. I didn't explain all the math in the post, but there is a mathematical concept taught in engineering school called confidence interval which allows one to determine mathematically if what is being observed is merely an interesting coincidence, or if something significant actually changed. Given the high volume (thousands) of ad impressions per day and the magnitude of the change, we compute there is a 99% probability that this isn't just an interesting coincidence.Refuting an experiment by asking for more study hardly constitutes an argument. Not only did I say that we would very much like to replicate the experiment with other keywords in the blog post, but I wouldn’t make such a strong claim in the first place without complete faith in the scientific accuracy of our data or the strength and validity of my point. While we seem to agree on the importance of accuracy and data integrity, I am at a loss as to why you felt the need to refute the experiment not just on our blog, but pretty much everywhere this information was posted. Is it possible that you might be feeling a bit defensive about the loss of Google+ author photos and what that means for the future of Google+? 

Mark Traphagen
Jul 12, 2014

I've responded to this elsewhere, Larry (this thing now exists in way too many places)! In that response, I noted that I now understand that your individual test was better controlled than it sounded at first. However, asking for reproduction by others is not at all out of line or extraordinary, but rather standard scientific practice before anything can be declared "proof." While you have controlled for many things, you can not control for the possibility of coincidence. Although uncommon, CTR's for ads do sometimes take leaps (or falls) that are hard to explain. Having others who could reproduce the conditions of this experiment and also show a CTR rise would be significant in increasing my confidence that you have esablished causal (and not just correlative) relationship.

markhare
Jul 12, 2014

Thanks for putting the time and effort into testing out your/others theory and suspicions. I don't know as it necessarily indicates intent on Google's part but its more so to me interesting in what it says about user behaviour. This indicates that the public median opinion of SERP results is at least partially clouded with doubt about the paid results they're getting. So people are very aware that ads are being used to get them to click through to something somewhat unrelated to what they searched but the addition of personal credibility helps legitimize the content.To me that would be more of a cue directing Google to make the decision around what I thought was the reasoning from the start: by allowing authorship in the form it was/is rolled out in the opportunities to abuse that trust are potentially poisonous to the user experience. At least in taking the headshots away it removes the credibility that's implied by having it there. Plus there could have been complaints from users who got duped by assuming a post with a headshot was somehow more legitimate. Making major decisions that affect the user based on financial reasoning sounds more like something Apple would do than Google.        

markhare
Jul 12, 2014

Oh also I've noticed there is a devout crew of G+ enthusiasts who's business is very dependent on G+ so you're likely to hear back across multiple platforms from a few people whenever contesting something about how anything related to G+ works.I went through the same thing when pointing out that indirect social signals such as referral conversions off of social media do factor into rank. I got 'Matt Cutts says....' back and I've seen quite a few others go through the same thing. There's never room for discussion which is so silly when it comes to SEO. All that seems to happen is an argument that goes nowhere followed by passive aggressive concession on both sides and I get more followers. 

Ayush Gupta
Jul 11, 2014

Thanks for posting this data. Google Fishy! I should experiment with Authorship too.

Zion
Jul 11, 2014

Why google remove authorship snippet in SERP results?

Shane Todd
Jul 11, 2014

To be fair mate and this is going to be pretty honest but you're no stud muffin... could it be that with the images that Sex sells and if you don't have it you're more than likely not going to get the click...... #justsaying 

Rand Fishkin
Jul 11, 2014

Thanks for putting this together Larry. Data rules the day. If you could expand to show this across larger groups of keywords, that would be greatly appreciated, and I think helpful in holding Google to task for putting forward at best, incomplete, and at worst, dishonest, messaging on the topic.

SimonTheSorcerer
Jul 11, 2014

I really liked my google authorship photos. It really helped us small guys trying to make living from our websites. It helped my CTR.Why not seperate desktop, tablet, mobile search? I understand it the author photos took up too much space on mobile, but on tablet or desktop they looked okay.I think utilizingall the capabilities of the device used for searching would only lead to a better search experience.    

Jasper Oldersom
Jul 11, 2014

Thanks for this Larry, It's hard to find evidence to back this up, but you found a great keyword to test it.I hope more people will put this to the test, so there won't be any doubt about it. Cheers!

Daniel
Jul 11, 2014

Great post and interesting data! It's obvious that Google did this to boost their revenue. But I never liked the photos is the search results in the first place.

Larry Kim
Jul 11, 2014

i miss them. it was incredibly powerful for personal branding.

Augustina
May 27, 2015

Hurrah, that's what I was seeking for, what a material!

present here at this website, thanks admin of this website.

Cristian Sepulveda
Jul 11, 2014

It makes sense t o think that Google prioritizes paid ads above the organic, after all that is your business. However, according to a recent MOZ post by Mark Traphagen, authorship photos offered a level of trust that was lost by this change in the SERPS. If photos offered greater confidence for the user, it was obvious that they sooner or later would eliminate.

Larry Kim
Jul 11, 2014

right. the same folks that argued that photos made the organic search results more "trustworthy" seem to be having a hard time accepting that the removal of those photos would have the opposite effect.

Steve Scott
Jul 11, 2014

So glad to have seen this study because this was my exact thinking behind the removal of Google Authorship photo's.  Look, I'm a big fan of Google search.  But let's be clear, they are in business to make money.  I've seen many eye-tracking studies showing the "pull" of those authorship photo's to search results lower in the SERPS than for those closer to the top of the listings.  It seemed rather obvious to me that removing those photo's would likely boost the CTR of Ads and your post goes a step further in proving that.  Thanks! 

James Carswell
Jul 11, 2014

Nice find Larry. Did ad position or top of page rate vary in the time you were looking? Obviously if you were showing higher up the page or in the banner position more often that would have a positive impact on CTR.Also, you state "I set about looking for evidence to support my theory". Did you find data that didn't support your theory too?  As Rand mentions it would be nice to see more data.

Larry Kim
Jul 11, 2014

sure. great question. I can tell you what didn't work. What didn't work was the first approach that we took, which was to aggregate click through rates for all the keywords in the adwords account. The problem with this approach was that it was flawed. CTR was changing (up and down), but there were many reasons for this, including changing Ad Positions, etc. It was too difficult to separate the signal from the noise. For this reason, we changed the approach to just try to find one "smoking gun" - to find one high volume keyword where we could isolate the variables and just observe changes in CTR before and after. 

Colby
Jul 11, 2014

I would liked to have seen a control screenshot of the SERP during the period of when CTR for the ad was lower, because perhaps the SERP listings were different on those days, and there was a more compelling organic listing that searchers were clicking on -- the SERPs are changing all the time. Overall I think you are probably right though.

Larry Kim
Jul 11, 2014

in some ways, it matters less (i'm not saying that it doesn't matter here) but it matters less if the organic search results are changing position in the serp. What matters more is that the ad position is constant at the top of the page, and that there were previously search listings that had photos in them, which were later deleted. The reason is because i'm not necessarily trying to prove the exact magnitude of the change, but rather to point out that the deletion of author photos from the organic results has an impact on the CTR of other elements on the page (like the ads).

Darian Schouten
Jul 11, 2014

Great article -and thanks for the mention, Larry!

Mary Kay Lofurno
Jul 13, 2014

Hi Larry - Thanks for the post, I think your on to something there, nice to see data.  I read the Moz piece, articles on search engine land and other blogs.  When I saw the announcement from Google. Not so much why Google did this but its impact on orgs and resource allocation priority as it relates to implementing techincal seo.  A quick full disclosure: I have been in marketing for over 26 years, started optimizing web sites back in 1998, had my own firm for a number of years and have now been an in-house SEO/SEM for over 11 years.   I have worked on all sites of all different conversion types.   I see from this is Google shooting itself in the foot some here and I will explain. I have spent many years wrangling resource time from IT departments to make changes on web sites as the industry has evolved from the early days of search until today.   IMO, after many years of working with CEOs, CFOs, CTOs, Marketing VPs, IT developers, IT Directors etc, many will say,  "why should we allocate resource time to make these changes in CMS on the backend [think large ecommerce/content sites] if Google is going to turn around and kind of discontinue it, to a certain extent.  See what they recently did with Author, who knows if its really going to be worth it."I have been in the trenches now writing business cases with  justifications for Microformats/Rich Snippets/Schema, etc for the last 4 years plus.  By far, Author was the easiest one to get concurence from the senior management team to implement it because they could get that one right away.  And as you know, its been well documented in many articles that the adoption of these standards [microformats, rich snippets, RDFa, schema],  has been scant when you look at it in aggregate. I am only bringing this up because I have not seen anyone else mention it.  As an in-house, am I going to stop advocating for those standards I mentioned above - no certainly not.  And, does Google have enough data to infer authorship/citations/trust.  IMO, they do.  And maybe that was part of it as well, we will never really know for sure. Anyway, its getting late, take care, Mary Kay   

Louise
Jul 16, 2014

I've no real doubt that this is the case on informational searches where many blog articles exist and authorship photos were previously prevalent. I'm less convinced by more commercial search terms where generally there weren't authorship photos. This is surely the case with most AdWords ads.Louise

Ledz Gonzales
Jul 16, 2014

Based on the data and screenshot you've shown, i think there is actually big difference in the results. I was conducting a research on how this changes will work through my campaign. Thanks for the great info Larry!!

Leave a Comment