It took me a couple of years of SEO’ing to understand how deceiving site links can be when you are reviewing Search Engine data and specifically impressions.
I have to admit that over the years I must have taken several inadequate strategic decisions because of this. In this blog post, I have screenshots with impressions, rankings and filters on how I have come to this deceiving conclusion – from two of the best rankings gambling sites in India.
In the end, I realize my claim about deceiving GSC also has a second sites of the coin – Continue reading to find out!
Why are impressions & GSC data so important?
When I create long & short term strategies I ALWAYS start in Google Search Console to see what’s up with the site and to get a hum of what direction I can take the project.
When I use the GSC data for short-term strategies it is mainly for the following:
- Conversion Rate Optimization – If page A receives a heck of impressions on a so-called transactional keyword, I optimize the content towards a strong CTA such as sign-up/download/buy NOW.
- Cannabilization – If I notice both page A & B receives impressions on the same keyword/s I de-optimize a page B for that/those keywords, and polishes up page A for that keyword.
- Intent – Intent is the equation of CRO and cannabilization which is calculated like (CRO – Cannabilization = Intent). If page A is cannabilizing on page B and does not convert my customers, then its a clear sign of optimizing towards the wrong intent.
What I mean about the short-term is a fix that can be done in a workday and usually yields results in the coming week. The purpose is to end up with a quick bang for the buck. Nice.
However, this requires that I have analyzed the data from GSC correctly.
As for long term strategies GSC impressions & average ranking is key. I weigh how many impressions a keyword has and which position it has. If a keyword with 1,000+ impressions is hovering somewhere on page two (let’s say position 11-17) I make sure to give that keyword & page some extra love.
I expect to see the results of this love in the coming six months. After all, I do SEO in very competitive niches and getting a high-value keyword as “live casino” on page one requires patience and sweat.
I use the data to conclude the following long-term strategies:
- Internal linking – Boys n gals! Time to go through all 500 pages and make sure we are internally ranking for the given keyword. Are you ready to do some mind-numbing work the coming week?
- Link building – Having a valuable keyword on page two in a competitive niche will require you to focus your link building to reach your goals. This drains plenty of resources from the organization.
- Intent – As mentioned above in short-term, intent boils down to all strategies into one result. Same goes for long-term strategies like If you’re aiming for a broad-term such as”live casino”. It will require you to optimize the page for a broad intent. This cancels out long-tail keywords which yields quick revenue and changes/jeopardizes the projects (or even the companys’) financial & growth strategy.
Using accurate data is of utmost importance for setting long-term SEO strategies. It’s like when you are navigating a boat crossing the Atlantic; if you set the bearing only 1 degree wrong you will end up in a whole other place than you intended.
Okay, now I have established how important accurate GSC data is for operating a website & an SEO driven company. Let me reveal what I have done wrong.
Old vs. New Site Links
It all started during the May 2020 update, which affected the SEO community at large. Internally I named this update “The Ego Stroke Update” since everybody in our different SEO departments found different areas which were affected and could be improved. The content team found pages & templates to improve, team tech found sloppy code, and so on. That is what usually happens during a core update – it is affecting all your SEO efforts.
Team Crunch SEO Data, which is me, found out that there was something “wrong” in impressions vs. our financial performance. The impressions hit rock bottom on several valuable keywords, but our revenue was steady.
Site link impression attribution 2011
Of course, there can be several reasons for this. But when I set out to research how impressions are attributed by site links I could find NOTHING, except for one deleted post from on Google Webmaster Forums in 2011. One of Google’s lead developers, Asaph Zemach, says the following:
It’s very clear: Site links does not attribute several impressions in GSC.
But let’s take a trip down memory road to create a better picture of why things are in the way they are – and why I write this topic with some agitation.
Site Links 2007-2016
In 2007 Google announced the feature of site links. At this time, you could demote site links if you found them unsuitable for your website.
As a webmaster, you have never been able to set your own site links. According to Mr.G they were assigned based on:
We only show sitelinks for results when we tink they’ll be useful to the user. If the structure your site doesn’t allow our algorithms to find good sitelinks, or we don’t think thats the sitelinks for your site are relevent for the user’s query, we wont show them. […] Sitelinks have evolved into being based on traditional web ranking, so the way to influence them is the as other web pages.”Site links are just an ingredient of the magic sauce in the ranking algorithm.
During this period I have been looking for other creditable sources saying that site links does not generate more than 1 impression – without success. So I can only assume that Asaph Zemach’s statement of “they do not generate any additional impressions in GSC” was still true.
2016 – May 2020: The Scam is Unravelled
Welcome to the period where I have been messing up – Please line up and throw a pie at me – free of cost!
When you are working or building websites with, let’s say 50 pages or more, you usually apply tech and content templates to guarantee quality, consistency and to speed up the process.
In our organization, I am the man behind the templates and responsible for implementing them. My long & short term strategies for content development are based on GSC data; Impressions, clicks & average position.
When you create medium to large websites you have to structure the pages into silos. The pages in the silos follow a template made for that specific silo. When your website gets affected by an algorithm you start by reviewing the site from a categorical point of view; which of my silos were hit and what was the outcome?
So in the weeks after the May 2020 update, I started reviewing the results and found out that some of our silos had been hit:
This was a tough hit on my ego and I was certain this was going to affect our business KPI’s negatively. I started digging into the silo to further investigate which specific keywords had dropped.
That’s when I found out how Google had scammed me over I-don’t-know-how-long on impressions generated by site links.
May 2020 And Onwards
Investigating further into I found no real common denominator for the keywords which had lost rankings & impressions, apart from one; the rank of the keywords itself.
In the picture below I have exemplified with the keyword “online lottery” which has the most impressions in the silo I reviewed.
How can the same keyword be top-ranking on several pages, at the same time? “Online lottery” showed up on our main casino landing page, software reviews and even our about us page. This was something that had been going on as far as I could look back in the GSC data (16 months).
This doesn’t make sense!
When you have the same keyword on several pages it’s called “cannibalization” and is a sign of your incapability of keeping intent & topics separate. Normally, one of the cannibalizing keywords is ranking at a position between 50-100, meanwhile, the primary keyword is hovering at positions 5-30. These are all ballpark figures, but just to give you an idea of how it works.
Attribution of site links impressions in modern time
In my case, ALL of the pages with the same keyword were on position 1. This made me think of how the data in GSC is generated and came to the following conclusion:
- Impressions – Impressions can be said the amount of people who want’s to go on a home party on a Friday night. When I was in my teens we were a group of friends rooming the streets looking for home parties. Sometimes we got in, sometimes we had to sit at the curb drink our lukewarm beers. It’s a hit n’ miss.
- Clicks – A click is when a someone finds & enters a party which we want go to (erhm, invited to), kicks the door open, pukes in the dog bed and leaves with the wrong right footed sneaker. Pretty much what our users does to our sites.
- Average Position – Meanwhile clicks & impressions are the lads, ‘Average Position’ is the rumour after the weekend. The rating of whom had the best party. Meaning, my group of friends contributes to the party rating, but it is not attributed by us.
In other words, impressions & clicks are generated by users and average position is generated by Google.
A keyword can only have 1 first position in the SERP, but in my case, several pages were assigned at position 1 for that keyword. The only function which exists in Google’s toolbox generating several pages on position 1 is called site links. Somewhere between 2016 and May 2020 Google changed how site links for a page triggered impressions.
“Online lottery” is a broad tail and this is just one example out of several broad tail queries in my case. Creating this blogpost I have researched the world wide web for more than 9,000 hoursss trying to find when that changed happened.
I can’t find a jack.
What Does This Mean?
I can only speak for my own experience, but it has led me to make inaccurate business decisions. When you are dealing with several websites with 500+ pages, all running traffic, you only do a big over-haul of the projects on yearly basis – don’t repair something which is not broken.
Since the pages were generating a good amount of impressions, clicks & ranking I thought “nothing was broken”. I thought we had a huge success coming up for the silo!
But if you strip the pages of the deceiving site link generated jumbo-mumbo it actually looks like this.
On the lower image, prior to the update, I removed the site link data and the results are:
- 76% decrease in clicks
- 93% decrease in impressions
All of a sudden my cash-cow pages were not worth JACK SHIT!
How it Affects Data Overall
Evaluating pages as I have explained above can result in uncertainties. However, the overall data in GSC is not skewed by site links impressions.
As you can see above the overall impression data for the keyword “online lottery” is 21.8k.
/lottery/ is the correct URL for ranking “online lottery” and as shown above it also stands at 21.8k.
How’s The Situation in 2021?
The same as above, apart from we have lost rankings on “online lottery”.
As you can see, the total impression for the keyword is reflected on the page it is supposed to rank on. However, what I believe Google has done since May 2020 and will continue doing, is displaying more relevant site links.
As a webmaster, you should therefore ask yourself why site links about gambling laws in the states Kerala, Maharashtra & West Bengal show up as the most relevant? Does it have to do with the legal situation of online gambling here…?
Blogging for me is a tool to improve myself and that is exactly what this topic has given me. I have learned that I have been wrong in how I have filtered and analyzed data coming from site links.
Since the total data is not affected by site links I can’t claim Google has done “something wrong”. In my opinion, it is just deceiving how they are displaying GSC data and requires a butt-ton of work filtering out all the unnecessary data to get to the core. Not very webmaster nor beginner-friendly.
Being able to filter impressions for site links effectively can leverage how Google & the users perceive relevance for a page – just look at which one-liner site links shows up in the SERP and what kind of traction they are getting.
But if you are looking at a site from a categorical or a birds-eye perspective, I still claim the statement of site links & GSC data to be very deceiving.