le SGE and your traffic: 3 SGE recovery cases //

Many articles have been written on Google Search Generative Experience, but the important questions remain unanswered.

• How will Google SGE impact organic traffic if and when it goes live?
• How much will our traffic decrease?
• What can we do?

This article provides a framework which can help answer these questions. This framework allows us to:

• Estimated traffic drop on 23 websites
• A new optimization technique has been discovered that helps to rank pages in the SGE snapshot cartousel.
• We have implemented three SGE recovery projects to mitigate (at least in part) the traffic drop expected from Google SGE.

We will share our findings and data, but we’ll also share our framework, which you can use to better understand the risks Google SGE poses for your website and to find solutions.

TL;DR

This article is extensive, as we have tried to provide you with all the information you will need to evaluate the impact of Google SGE, and in case there is a significant traffic drop, to design your own SGE Recovery Project while sharing our own data.

The following are some key points, along with links for more information in this article.

• 
  You can estimate the traffic that you will lose or gain by using Google SGE
  
  .
  • Anyone can use an Excel spreadsheet to implement our open SGE Impact Model. This model will allow you to estimate the impact of SGE on your traffic, based on a variety of outcomes. Here is an example where a website will lose between 44-75% organic traffic as a result of Google SGE. Discover how the model functions.

• 
  A study of 23 sites found that the organic traffic dropped by 18-64% as a result SGE
  
  .
  • Our study concentrated on websites within the technology industry. The traffic was primarily from informational keywords. Our sample is very diverse, with some sites gaining traffic by as much 219% while others lose as much 95%. View the results of this study.

• 
  You can optimize your pages so that they appear in SGE carousels.
  
  • We found a successful technique, tested it in a controlled experiment, and implemented it across several websites. We have an experimental framework to help you quickly find the right technique for your niche. See the results of our experiment and learn how to design SGE tests.

• 
  Combining the SGE impact models and experimental framework can help you mitigate SGE risks.
  
  • You can prevent a site from experiencing a significant traffic drop by optimizing the keywords for SGE. This is a relatively easy way to save your website. Learn how to create an SGE recovery project.

• 
  Shared results of the first three SGE recovery project in the world.
  
  • In our most successful case, we saw a client who was expected to lose 79% of their organic traffic now grow 97% in the best scenario. In order to achieve this result we optimized 11 pages for Google SGE, and created four “ideal” pages that are friendly to SGE. Learn more about the three case studies.

Before

After

• 
  We are committed in helping the community grow and survive from Google SGE.
  
  • Join me for a free Webinar on Thursday, September 28th – Overcoming The Google SGE Challenge : Assessment and Recovery Strategy. We will go into greater detail and answer questions from the audience.

The SGE Impact Model

To understand the risks of SGE we need to first understand where our keywords rank in the new experience and how these ranks translate into traffic.

We can determine if organic traffic will drop or grow when Google switches to SGE by comparing the expected traffic to our current traffic.

To achieve this, we built a simple website model that estimates the traffic changes expected across four scenarios. This is an open model that we propose to the community.

The principles listed below can be used by anyone to create an SGE impact spreadsheet in Excel or Google Docs to better understand the effects of SGE on their website, and to get useful insights to help with recovery.

Simplifying assumptions

Google SGE has many unknown parts. It is a complex and large system. We use some simplifying assumptions to model it. Although these assumptions are controversial and not always true, we think they are reasonable because of the thousands of keywords that we observed ranking on SGE.

• Assumption #1: You will receive (some) organic visitors if your page is ranked in the carousel of snapshots.
• Assumption 1: Traffic generated by the carousel of images will be the same or lower than that which is currently ranked on Google.
• Assumption 1: The blue links are the most popular links. If you’re not in the SGE carousel but in the 10 traditional blue links, then you’ll get no traffic. (This is a conservative estimate; in reality some users may scroll past SGE to click on the blue links.
• Assumption #4: You will still rank at the same level as today if the SGE experience does not exist.

The CTR and ranking are the two parameters we use to determine “how bad” SGE is for our website.

Click-through Rate (CTR) for SGE

The CTR for different positions on the SGE snapshot is not known. We use two extreme scenarios, one optimistic, and one pessimistic to explore the possibilities.

The table below shows:

• Current Google CTRs according to Advanced Web Rankings. This data is for non-branded informational keywords, and it’s current for June 2023. The model can be adapted to different types of keywords using a CTR that is not the baseline.
• CTR (Optimistic). The CTR optimistic for the top 10 places is similar to Google’s today. This is the “best case”. We expect that the CTR for the top three positions will be the same, as they are displayed side-by-side, and not stacked one on top of the other, like Google does today.
• CTR in Google SGE (Pessimistic). We estimate that SGE results in half of the CTR at the top 10. We assume that the CTR for the top three positions of SGE is identical.

The actual rank in SGE

You may have noticed that rankings are very volatile in SGE. Pages can disappear, reappear, or change rank. In some cases the SGE experience will not show up at all. Your page may have a different rank in multiple observations for the same keyword.

We create two “ideal” ranks in our model by taking a set rank observations:

• This is the highest rank. If you checked three different times and got 3, 5, and once didn’t get a rank, then the optimistic ranking is 3.
• Pessimistic ranking:The ranking is pessimistic if even one observation shows that the page did not rank. We take the average if the page ranks in all. The pessimistic ranking is ‘not ranked’ in the example.

Four scenarios

We have defined four scenarios of SGE’s impact, from the best to the worst, based on these parameters.

Best scenario

• Optimistic ranking / Optimistic CTRL: We take the optimistic ranking (best rank ever seen) for all keywords and the CTR (like Google does today).

Mixed scenario

• Pessimistic ranking / Optimistic CTR: We take the optimistic rank and combine it with the CTR pessimistic (half the current CTR) for all keywords.

Mixed scenario

• Pessimistic ranking / Optimistic CTRL: We take the pessimistic ranking combined with the optimistic CTRL (like Google does today).

Worst scenario

• Pessimistic ranking / Pessimistic Click Through Rate: The worst case scenario is to combine the pessimistic ranking with the pessimistic click through rate (half of what you have today).

Selecting keywords

This is an important part of our SGE Model. We want this model to be actionable, i.e. we want to improve and optimize traffic based on its results. We want to identify a few keywords that can “explain” the majority of organic traffic.

Here are some steps to help you find the 20% of words that have 80% impact on your website.

• Identify “SEO relevant” traffic on your site. Exclude sections of your site from the analysis if you do not optimize them or don’t “care” about them for SEO.
• Identify top pages driving at least 70% organic traffic. Even if you have hundreds of pages, only a few will drive the majority of traffic.
• Select one representative keyword for each page. Even if you have many keywords ranking on a particular page, it is likely that there is a “flagship” keyword for the page. Most likely, if that keyword is ranking, the rest will be as well.

You can use this method to arrive at 10-50 keywords that are responsible for the majority of your SEO traffic (for smaller websites). For larger websites, you may reach hundreds of keywords. You should use these keywords to estimate the impact of SGE.

Then, you should identify strategic keywords. These are keywords which, although they may not have appeared in the analysis above, are still important to the business. We want them to rank well in SGE.

Results of the SGE Impact Model for 23 Websites

It must make you curious to know what our model predicts in terms of traffic growth (or decline)! You must be curious to see what our model shows in terms of traffic drops (or growth!) The results are:

The following is a list of the datasets.

Agile SEO is a technology-focused agency.

We tested our impact model with 23 clients, including startups and established technology companies.

Each of them has between 100 and 10,000 indexed pages as well as between 1,000 and 300,000 organic monthly visits.

Keywords

The pie chart below shows the expected changes to our client’s most important keywords once SGE is live.

We identified 1,242 keywords with high impact across 23 websites.

It is a shocking fact that 64% strategic keywords or high traffic keywords have lost their ranking in SGE. They do not appear on the SGE snapshot cartousel.

The good news is that 27% of keywords pose no risk. This means that traffic is expected to be the same as today, or even higher. Or that strategic keywords rank roughly the same as today.

A second important finding is that only 8% of all keywords are SGE-free. Multiple studies have shown a similar 8-10%, which indicates that Google activates the SGE for the majority of keywords.

Traffic aggregate in SGE

Below we can see the expected traffic for all 23 websites.

We analyze 628,000 organic monthly visits (driven primarily by the 1,242 keywords with high impact shown above).

The four columns and four bar graphs show our four scenarios.

• The websites’ organic sessions drop by 18% when using the ideal scenario. We estimate a high CTR here (roughly equivalent to Google’s current CTR) and take the best ranking for each keyword.
• The worst case scenario would see the websites fall by 64% and wipe out 405,000 organic traffic sessions per month – ouch. We estimate that the CTR for Google SGE will be about half what it is today.
• There are still significant drop-offs of 38-54% in the mixed scenarios .

Even in the best-case scenario, SGE will result in significant traffic reductions.

If we assume the worst on one of our parameters, such as CTR or rank, then those traffic drops will increase by 2X-3X.

Traffic per website (SGE)

The aggregate data conceals important differences between websites.

Below is a chart that shows the data for all 23 websites. It includes the number of keywords with high impact, the current organic traffic and the expected traffic changes in Google SGE.

Some of the 23 websites in our sample actually show traffic growth.

It is because their keywords are more important and rank higher in SGE. This means that they get more traffic than the current situation.

Two of these examples are highlighted in green. Most notably, the website #4 has seen a traffic increase by 219%.

We also saw this in our SGE recovery cases (see below).

Some websites have been very hard-hit – look at the two red examples. Website #16 for example is down by over 90% in all scenarios.

There is a cause for serious concern across all of the websites we sampled. We now move on to the next phase of the discussion. What can we do in response to SGE traffic drop?

A framework for SGE experimentation

Let’s begin at the end. This is the keyword that we optimized to put our client in position 1 of the SGE snapshots carousel.

This page was ranked in Google between positions 1-3 for the keyword “incident response”, but it didn’t rank on SGE.

The page is now consistently ranked in the top 10 in SGE. This has been the case for more than two months.

We need to identify pages that do not rank in SGE and get them ranked as high as they can in the snapshot carousel.

Four rules for an effective SGE experiment

You can use the following conditions to carry out experiments and discover how you can achieve the result above.

• Concentrate on one page or keyword at a given time: The interventions should be restricted to a single page/keyword, while keeping site-wide factors constant.
• Concentrate on factors that you can control. For instance, if the design or content of the page is not under your control, these factors cannot be included in your experiment.
• Make the change within a week. This step is crucial because SGE rankings can quickly shift. It is less likely that a rank change is the cause of a result if you measure it within a week.
• Create a control group. Find a set of similar pages or keywords, and compare their rankings while you are implementing your intervention. You’ve got a winner if their rankings remain stable on average and your experimental page increases in ranking.

This framework might be considered too restrictive, as it doesn’t consider factors that are site-wide or that can’t be changed in a week.

It’s true. But if you find a way to satisfy these rules, and it works, then you will be able to easily recover from SGE traffic drop.

Our SGE experiment results

We conducted an experiment, following the above rules, to try and optimize pages so that they appear in the SGE carousel.

The SGE carousel does not show the keywords that are in the top 10 of Google.

We used two different intervention methods (A & B), which involved both on-page and offline factors. Each intervention included seven keywords.

Parallel to this, we also tested a group of keywords similar in nature on the same website, but without any intervention.

We checked the ranking of the 14 keywords in each Google account before the experiment.

• Account 1: No keywords are ranked.
• Account 2 only 2 ranked

We checked the rankings of seven keywords after intervention A in three Google accounts.

• Account 1: No keywords are ranked.
• Account 2 – 2 ranked
• Account 3: No keywords are ranked for citations

We checked the rankings of 7 keywords after intervention B in three Google accounts.

• Account 1 – 6 in the top 3 and 1 without SGE
• Account 2 – 6 in the top 3 and 1 without SGE
• Account 3 –

Control Group:

We checked the SGEsnapshot Carousel for 89 keywords that were not previously ranking.

• 21% of these words entered the carousel of snapshots during the week of experiment.
• 79% of the people were not included in the snapshot carousel.

The odds that a keyword will enter the snapshot carousel “by chance” are 21%. In our experiment, we were able to place 93% (15 out of 16 positions) of the tested positions into the snapshot carousel when SGE was active. This shows that the intervention caused a rank change.

We concluded that the intervention B achieved SGE ranking for keywords in the top 10 list, but not in SGE.

Note

We will not share our technique, as it was designed for a specific application in the SaaS/technology market. Google’s algorithm is known to work differently for keywords and niches.

There will be different techniques in every niche. We have shared our general framework for experimenting, so that anyone can find the right technique to fit their particular use case.

SGE Recovery Project Process

This process can be used to design a complete SGE recovery for a site once you have found a technique that is repeatable and can help you achieve SGE rankings for high-impact keywords:

Step 1. Perform SGE impact analysis as shown in (above).

Step 3: Create a list with keywords that are:

• Google SGE is a strategic ranking loss. This is called “strategic ranking loss.”
• According to the worst scenarios, driving high traffic and expecting to have less traffic on Google SGE. This is what we call “maximum impact.”

Step 3 : Prioritize the keywords in this order:

• Take a look at the keywords that your business cannot afford to lose.
• Second, choose the keywords that will have the greatest impact on your traffic (you can eliminate keywords with lower search intent).

Step 4: Optimize existing pages using your experimental technique.

Step 6: Other ways to achieve SGE rank

• Create the “ideal” page that you think would rank higher, based on your ranking factors. In most cases, the new pages will be ranked in SGE.
• Find another page with the same parameters, but not as well focused on the keyword. Rewrite the page to emphasize the keyword that you want to rank for in SGE.

Step 6. Run the SGE impact analysis again and check if:

• Reduce the number keywords that are losing ranking.
• At least in some of the scenarios, mitigate the traffic drop.

We have used this process successfully for several clients since we first discovered it. Here are some of the results that we achieved.

Three SGE Recovery Case Studies

We were focused on saving our clients’ traffic from expected drops as soon as we realized the threat of SGE. We are proud to announce that we have successfully completed three SGE recovery projects.

Note We do not share the names or domains our clients in order to protect their privacy. We can share this information in confidence with the media or potential clients. Please contact us if you’re interested.

Case 1: Cloud Native Developer Tool, 175,000 organic monthly sessions

Background

• Website of a corporate company with product pages, a blog and e-learning center.
• 755 pages indexed.
• The organic traffic increased from 21,000 in September of 2021 to an average of 174,000 in Q3 2023.
• SGE Recovery Project focuses on 135,000 visitors driven by 38 keywords

SGE Impact Analysis

• 
  Before you begin, please consider the following:
  
  The website was to be:
  • You will lose ranking for 7 strategic keywords.
  • According to the SGE scenario, organic traffic can be lost between 60,000 (44%) and 102,000 (75%).

• 
  What we did
  
  • Optimized 9 existing pages for SGE snapshot carousel, using our experimental technique.
    • Eight out of nine people entered the SGE carousel snapshot (88% success rate).
  • We wrote 8 “ideal” pages that we thought would rank higher in SGE and then applied relevant off-page criteria.
    • Three out of eight entered the SGE carousel snapshot (37% success rate).

• 
  After
  
  The website should:
  • Saved 4 keywords.
  • In the best case scenario, the drop is only 5% (compared to the 44% that was expected before the project).
  • The worst case scenario will drop 57% (compared to 75% prior to the project).

Evaluation

In the best case scenario, we have reduced the risk to this client. Two other scenarios still pose a significant risk. You can continue to optimize keywords and reduce the risk in all scenarios.

Case 2: Enterprise Cloud Platform, 35,000 organic monthly sessions

Background

• Website of a corporate company with product pages, blog, and learning center.
• 1 250 pages indexed
• The organic traffic increased from 10,000 visitors in January 2021, to 45,000 users in 2023 Q2. Due to the site’s relaunch, traffic dropped to 35,000 users in 2023 Q3.
• SGE project focuses on 19,000 visits driven 68 keywords.

SGE Impact Analysis

• 
  Before you begin, please consider the following:
  
  The website was to be:
  • Lose ranking for 23 strategic keywords
  • You will experience a traffic drop of between 15,000 (79%) and 16,000 (85%).

• 
  What we did
  
  • Optimized 11 pages for SGE snapshot carousel, using our experimental technique.
    • 8 of 11 people entered the SGE carousel (72%) success rate.
  • Write or rewrite 4 pages in order to create the “ideal page”, and apply relevant off-page factor.
    • Three out of four entered the SGE carousel snapshot (75% success rate).

• 
  After
  
  The website should:
  • We saved 17 strategic words (there were multiple strategic keywords ranking on the same page).
  • In the best scenarios, you can expect to grow by 97% or between 2 and 28 percent.
  • Drop only 31% when comparing the worst scenarios (compared to 85% prior to the project).

Evaluation

We were able to turn SGE Impact around, from a massive drop in traffic to a possible increase of almost 100 % in organic traffic (in the best scenario).

We also saved 17 strategic keywords which were losing their visibility in SGE.

Case 3: Cybersecurity Solution, 15,000 Organic Sessions per Month

Background

• Website of a corporate company with product pages, blog, and learning center.
• There are 888 pages indexed.
• The organic traffic increased from 3,000 visitors in January 2022, to a median of 15,000 users in Q3 2023.
• The SGE project was focused on 5,000 visitors driven by 34 keywords. (The rest of the traffic has become less important for the business because it is changing its strategic focus).

SGE Impact Analysis

• 
  Before you begin, please consider the following:
  
  The website was to be:
  • You could lose between 1,100 (22%) to 2,100 (42%) of your valuable organic traffic.

• 
  What we did
  
  • Optimized 3 pages for SGE snapshot carousel, using our experimental technique.
    • Two out of three entered the SGE carousel snapshot (66% success rate).
  • Write or rewrite one page to create the “ideal” page for ranking on SGE. Apply relevant off-page factor.
    • The page has entered the SGE carousel of snapshots (100% success rate).

• 
  After
  
  The website should:
  • Grow 84% with the best case scenario. Or, move from a growth of 11% up to a slight drop of 9% when using mixed cases.
  • Drop 34% when compared to the worst scenarios (compared to 42%) before the project.

Evaluation

This is an example where a relatively minor drop in SGE traffic was greatly mitigated by optimising only four pages.

This highlights the importance of focusing your efforts on the right keywords in order to recover SGE.

Final thoughts and Next Steps

We haven’t seen any concrete advice on how SGE affects websites or what they can do in order to recover.

This article will hopefully help you to understand Google SGE from three different perspectives:

• Impact Understand what the likely impact of Google SGE will be on specific websites in concrete terms, i.e. what keywords are at risk and how much traffic is at stake – by using our SGE impact report.
• Action: Understand what we can do via our experimental framework to mitigate specific rank drops and traffic drops identified by the Impact Model. We have successfully implemented SGE recovery projects using this framework.
• Hope: Our experiences show that, even if SGE is a threat to the organic traffic built up in your SEO project, you can adapt your practices to suit this new search experience. is not the end of SEO as some claim. is just the beginning of an exciting new era.

SGE Recovery Webinar

• On Thursday, September 28th at 1 pm. On Thursday, September 28 at 1 p.m. ET, I’ll be holding a free webcast – overcoming the Google SGE Challenge : Assessment and Recovery strategies. We will go into greater detail about the SGE impact model, our experimental framework, our SGE recovery cases studies, and answer questions from the audience. Save your spot.

This article is intended to inspire a community-wide effort to collect more data and develop solutions for SGE across a range of niches.

SEO: Here’s to another 20 years!

The post Google SGE and your traffic: 3 SGE Recovery Case Studies first appeared on Search Engine land.