Traffic Dropped After a Core Update? Split SEO and GEO

When a Google core update rolls out, don't look only at ranking changes. Record SEO metrics and GEO signals for the same period together, but interpret their causes separately. A core update is a broad change that recalibrates evaluation criteria across Google's ranking systems, so simplifying it into "an event where one factor changed" easily leads to the wrong response^[1][2].

Responding to core updates used to be relatively simple. You listed the keywords that went up and down, then checked content quality and backlinks. But as AI Overviews and generative search experiences have moved into the search results, reading the same numbers only the old way now means missing important parts.

A core update is a "scoring rubric adjustment"

Google describes core updates as broad changes that recalibrate its overall evaluation so search results are more useful and reliable^[2]. Unlike small updates that fix a single feature, a core update can affect page groups differently, such as informational posts, product pages, and comparison content.

What matters is that Google doesn't disclose detailed weightings or specific influencing factors. There's no announcement like "this time we're weighting brand trust more" or "we changed how AI answer sources are selected." So before drawing conclusions, practitioners need to set observation criteria and look at metrics separately.

Why SEO metrics alone aren't enough

In the past, the impact of core updates was mostly viewed through SEO metrics.

• Which keyword rankings went up
• Whether organic clicks dropped
• Whether CTR changed while impressions held
• Which page group was hit the hardest

These metrics are still important. But as AI search experiences have moved into the search results, there are more situations where the same numbers need to be read differently. For example, clicks can drop even when rankings haven't fallen much. You could see this as simply "a ranking problem," but it's possible that AI Overviews had already resolved the user's question^[3]. Conversely, clicks may drop while your brand is mentioned more often in AI answers, and if you only look at the traditional SEO report, you miss that result.

So it's safer to split analysis of a core update period into two layers.

The key is to look at the same period without bundling everything under the same cause.

Practical checklist: what to look at first

During a core update, you shouldn't draw conclusions too quickly. Rankings and impressions can fluctuate during the rollout, and changes over a few days may not reflect the final state.

1. Fix the baseline period first

Compare 2-4 weeks before and after the update start date on the same basis. If possible, wait at least one more week after the rollout completes and look again. Don't declare "this is the cause" while it's still in progress.

2. Separate brand and non-brand queries

Searches containing your brand name and general informational searches mean different things. If brand queries fluctuate, trust and reputation issues may be mixed in; if non-brand queries fluctuate, content quality or changes in competing content may play a bigger role.

3. Look by page group

Looking only at the overall traffic average misses problems. You need to split page groups, such as blog posts, service pages, comparison content, and pricing pages. From a GEO perspective, look separately at how explanatory content and evidence-based content are used in AI answers.

4. Look at search data and behavior data together

In Search Console, check impressions, clicks, CTR, and average position; in your analytics tool, check actual behavior after the visit (inquiries, signups, resource downloads).

5. Record AI Overview visibility separately

Not every query triggers an AI Overview. For each key query, record whether an AI Overview appears, which sources are attached, whether your brand is mentioned, and whether competitors appear.

6. Classify the types of cited sources

AI answers not citing your site isn't an immediate failure. If media, reviews, communities, and partner sites describe your brand positively, that can be an indirect signal. Conversely, if your page ranks but the AI answer centers on competitors, you need a GEO response.

7. Don't jump to a cause during the rollout

During a core update, ranking adjustments, changes in AI Overview display, competitor content updates, seasonality, campaigns, and technical issues can all overlap at once. Rather than saying "we dropped because of this update," first look at "which metrics moved together during the same period."

Why do core updates keep happening?

If you treat a core update as a one-time event and think "once we get through this one, we're done," you'll be caught off guard all over again with every update. Google regularly carries out ongoing recalibration of search quality, and these updates are likely to keep repeating^[1][2].

So the right approach to core updates isn't "analyzing this one update" but "building an analysis routine you can repeat every time." With a routine in place, you won't have to repeat the trial and error of setting a baseline period and splitting metrics from scratch when the next update arrives.

A repeatable routine looks roughly like this.

1. Record SEO metrics for key queries and page groups weekly, even in normal times
2. For the same queries, also record visibility and citation status in AI search engines weekly
3. When a core update is announced, automatically split and compare the periods before and after the start date
4. Tag page groups and brand/non-brand categories in advance so you don't reclassify them with every update

If you build up data like this in normal times, when an update is announced you can skip asking "what changed?" and go straight to checking "which metrics deviated from the baseline?"

An example makes it clearer

Imagine a B2B SaaS company. After a core update, organic clicks for the non-brand query "AI search optimization tool" dropped. Average position slipped slightly, from 3rd to 4th. Looking only at these numbers, you might see it as a simple ranking drop.

But when they searched the same query directly, a large AI Overview appeared at the top, and the answer cited comparison review sites and community posts, not the company's own site. In this case, the response isn't simply changing the title or adding more keywords to the body. First you need to check whether the site has definitions, comparison criteria, use cases, limitations, and pricing criteria that AI can easily cite, and at the same time check what sentences describe the brand in external reviews and communities.

Let's tweak the same case slightly. If this company had been recording AI answer visibility all along, it would have already known the pattern "comparison review sites are often cited for this query" before the core update. Then, seeing the drop in clicks after the update, instead of panicking, it could interpret much faster that "as expected, the AI Overview impact showed up strongly on non-brand exploratory queries." Whether or not you keep records in normal times makes a big difference in how fast you respond to a crisis.

In the Korean market, you need to look wider

In Korea, looking only at Google isn't enough. Depending on the industry, Naver, YouTube, communities, review sites, and map services together shape the search journey. Looking at GEO in light of a Google core update doesn't mean "optimizing only for Google." It means using the changes on Google as a starting point and also looking at the other surfaces where users actually check information.

So a GEO review for Korean companies needs to separate at least three things.

• The quality of explanations on your own site
• How your brand is described externally
• The sources and sentences AI answers actually select

If you don't separate these three, it's hard to know what to fix even after the core update has passed.

Conclusion: treat it as a signal to change how you respond

It's natural to ask "what changed?" with every core update. But since detailed causes are mostly undisclosed, your practical response needs to be more specific. Don't just look at whether rankings went up or down; look at how SEO metrics and GEO signals moved together during the same period.

Now you need to ask: how did our pages move in search results, is our brand visible in AI answers, which sources are cited instead of us, and have we distinguished whether the drop in clicks is a ranking problem or an AI answer problem? GEO isn't a concept that replaces SEO; it's an operating approach that adds AI answer visibility and source trust on top of SEO^[4]. Using a diagnosis tool like TRAIL Search, you can turn the work of observing these two layers separately on the same timeline into a repeatable routine.

For related reading, see What is GEO (Generative Engine Optimization)? Which pages to start with, and in what order, SEO and GEO visibility strategy: when you rank high in search but don't appear in AI answers, and SEO fundamentals in the AI search era: what changed in on-page, off-page, and technical SEO.