Leveling Up Organic Traffic Analysis with CustomGPTs | Seer Interactive

into our current repetitive processes. As we were thinking about this, we came across

this blog post that we published a couple of years ago, 18 Ways to Diagnose a Decline

in Organic Traffic. Reporting is something that every SEO or marketer will have to do,

whether that's on a weekly, biweekly, or monthly cadence. And while this post was published

a couple of years ago, a lot of the content within it is still relevant today. That includes

looking into seasonality or search interest, competition, keyword visibility, and rankings

whenever you're trying to analyze performance shifts in organic traffic. What we wanted

to do was to create a custom GPT for this process to help it even more efficient. Or

if you're a newer analyst that's just now getting started on reporting, hopefully this

GPT can act as an assistant into walking through that process and providing some assistance

with analysis or optimization recommendations as well. We used the content of this page

as well as some best practices that we have at Sear in order to make this custom GPT,

the SEO Organic Traffic Decline Analyst. What I want to do is walk through the workflow

of this custom GPT and just highlight some of the capabilities as well. To start, the

GPT will ask for an upload of your Google Analytics or Google Search Console report.

This will help it provide a light analysis on the data as well as provide those recommendations

down the line. I provided an export of Sear's Google Search Console report comparing March

2024 to February 2024 data. It starts off by analyzing this data to provide the total

number of clicks, the total number of impressions, as well as the increase or decrease in these

metrics based on the comparison period. One thing that I do want to call out is that whenever

exporting from Google Search Console to watch out for sampling, the actual Sear site had

around 28,000 total clicks for March. However, within the export, we only had just under

10,000. If anyone has any ideas to get around this sampling, definitely let me know. However,

this is still going to provide a good enough directional analysis. However, if you're using

Google Analytics or do have an unsampled report, it should provide an accurate analysis

for the clicks as well as any changes for your performance as well. So just something

to be aware of. After it provides this initial analysis, it has a couple of core functions

to look into. This includes Google algorithm updates, whether any of these happened over

your comparison periods, Google organic ranking changes, search interest, and search feature

changes. Whenever we get to those, that's going to be primarily using SEMrush or Ahrefs. But for

now, we're going to stick with the current flow and look into Google algorithm updates next. So

I should ask to check for some updates during this time period. As you recall, my data was

based in March 2024. So the GPT is able to remember this information and say, yeah, there

was a significant algorithm update in March 2024, the Quora update or March 2024 Quora update. So

it uses information from a couple of different sources such as Google, Search Engine Journal,

Mozcast, in order to get all of this information to provide you with this information as well as

some context and recommendations for next steps as well. So the March 2024 update was primarily

focused on low quality content or AI written content. And so after it provides all of this

information, I say, yeah, we do have some AI assisted content on the SEER site. Let's check

out how rankings changed over this period as well. So for rankings, it'll ask for a export of the

SEMrush or Ahrefs rankings. So I provided the respective exports for each of these comparison

periods. It will initially run into a slight error, but this is just it reformatting into a

different file type. So after this, it'll be able to work through the analysis more efficiently.

The core focuses within the ranking changes are going to be overall change in rankings. So how

many total keywords was your site ranking for previously and in your current period, as well

as striking distance keywords. So at SEER, we classify striking distance keywords as any

keywords in position 11 to position 30. These are the pages two and three of Google, where maybe

just a little bit of love on those pages can have a large impact. We're also looking at page one

keywords, so positions one through 10, as well as the search volume changes. We're going to look

into search volume changes as a proxy for those search interest changes and the seasonality to see

how the volume of users searching for these things has changed for the comparison periods. So it'll

provide some analysis on the keywords. So for instance, it's seen that SEER had a drop in

position 10 to position 80 for this branded term. It's also looking at page one keywords, again,

that branded term and some of the other terms that we're ranking for on page one with other

posts on the site. For example, screaming frog improved from position 11 to three. If you want,

you can drill into this even more to see like what pages these are ranking for. But in its core

element, it's acting more as a consultant in order to provide holistic analysis of the site. It's

also looking into traffic implications and this is also based on the report that you provided

earlier as well. So you can, it'll ask you some questions or if you want to move on to the next

section, feel free to, you know, dig in on some of this too and use some of that consultative nature

in order to really dig in on your performance. So let's look into what the overall change in page

one, overall change in page one keywords from February to March here. So in order to look at

this, it goes back and looks at those comparison periods that we looked at and provides the amount

of keywords that are new on page ones, the amount of keywords that were lost, and the amount that

retained. It's also able to provide some notable observations from this to provide, you know,

some recommendations for optimizations moving forward. So one of the goals within the GPT is to

see, you know, like what the primary cause for the performance shifts is. So one thing I asked is, is

it safe to assume that ranking improvements can primarily be attributed to the increase in clicks

from February to March or something else here? It's now it's going to say, you know, like it's

reasonable to initially look into this, but let's drill into some of those other capabilities that I

mentioned previously, such as search volume changes. So as it provides some of this information, I say,

okay, yeah, that's a good idea. Let's look into search volume trends. So it uses that previous

export that we provided from SEMrush into looking into the changes in search interest and seasonality

from February to March. So as it looks into this, it sums all of the individual unique search volume

from February to March in order to provide some analysis on if this was a, you know, significant

increase, a significant decrease, or if this was relatively modest. So one thing that I like that

this highlights here is that the, while we did experience an increase in search volume for page

one keywords of 21,000 searches, it says that this overall increase was a relatively modest increase.

And while, you know, like we did experience that increase, the overall increase in search volume is

relatively small compared to the total search volume. So this indicates that seasonality and

search interest probably isn't the primary contributor for our increase in clicks from

February to March. So we confirmed that this is for page one keywords.

And the next thing that we're going to look at is for the impact of SERP features.

We're using the impact of SERP features within the SEMrush export as a proxy for competition to

see if there are any other features that are gaining traction on page one. But if you would

like to include exports of a competitor SEMrush file so it can look into difference and changes

into those keywords as well, feel free to do that. So one thing that SERP features can affect is your

click-through rates of your data. So the GPT will help walk through and identify some of the

SERP feature changes and assess this impact. And as I scroll through, it'll provide some analysis.

So one thing that I was really interested in is what the overall frequency changes were

for specific SERP features. Ran into a little trouble to start out here. However, it finished

up by being able to clean up the data and provide some of these overall frequency changes for our

page one keywords. So looking into this, we're able to see that reviews increased for 1,800

keywords over the course of month over month. We had an increase in ads at the top of the search

results as well as jobs, video, and ads at the bottom of the page one search results. Some of

the things that decreased were site links, the video carousel, feature snippets, and image. So

while you're able to look into this, this helps you know like say maybe what you should prioritize

in your campaigns and strategies moving forward. Maybe prioritizing images on your blog post and

such while this still is important for user experience. Maybe trying to win image features

results isn't going to be a primary driver in the competition. So with all these observations,

it's able to provide some of the recommendations and next steps for these. So if you're experiencing

increases in ads at the top of the search results for your keywords, this could affect

click-through rate as your results, even if you're in position one, are being pushed down maybe below

the fold. So as it's able to look into all these things, I ask just to compile all this information.

What are the primary drivers for my performance shifts from February to March and what should I

prioritize as next steps? After it compiles all this information, it's able to look into some of

those primary drivers for performance shifts. So the main thing that it mentions is the ranking

changes that we experienced on gains on page one as well as the keywords that were attained on page

ones. The March 2024 algorithm update could have a major impact on not only March's performance,

but as well performance moving forward. So this is something to look into. And also search volume

trends. It's a modest overall increase in search volume, but it could contribute to some of the

higher traffic. So it provides this analysis as well as some of the recommended actions for next steps

such as content review and optimization and looking into certain feature

optimization to increase that page one visibility as well. So this was the end of

this workflow. You can continue to look into and drill into some of these changes for your performance,

but feel free to work with this GPT. We're going to share out the link and let me know if you have

any ideas to level this up or have any feedback or have any questions. We'd be happy to chat

through this. Hope this helps and yeah, have a great rest of your day. Thanks.