Butterfly Effect

#249 How Core Updates make subtle user intent changes

A warm welcome to 24 new Growth Memo readers who joined us since last week! Join the ranks of Amazon, Microsoft, Google and 12,500 other Growth Memo readers:

---

---

👉🏻How Scaled Localization Drives High Authority PR Coverage

Scaled localization is digital PR’s version of the “near me” keyword modifier.

You know how "dentist near me" is way more valuable than just "dentist Austin," right?

Let me explain simple localization and its importance in digital PR.

An Austin reporter cares more about a story that shows how national trends are impacting the Austin area than she does about one that doesn’t.

She always needs to tie her stories back to the people in her community. So if you’re pitching something that doesn’t, there’s a very low chance she’ll cover it.

Scaled localization is when you create a story that has dozens of local angles so that it can be credibly pitched to hundreds of local reporters.

Take this study on the top neighborhood complaints in America. Here, we analyzed a huge data set of 58,735 posts and 354,972 comments on the Nextdoor app in 50 of America’s biggest cities.

Then, we looked at over 200 keywords representing common community issues.

This approach allowed us to produce multiple rankings tailored to cities across the US, grabbing the attention of reporters in various places.

Our media relations team could craft precise outreach campaigns with rich data for multiple cities and topics—again and again and again. 😄

The result? We earned 66 placements, with 33 of them boasting a Domain Authority of 80 or higher. Notable placements were the Miami Herald, The Baltimore Business Journal, The Fort Worth Star-Telegram, The San Antonio Current, and The Houston Chronicle. 🚀

Exclusive for subscribers of the Growth Memo: We're opening up 3 slots for bespoke ideation decks full of digital PR ideas ready to execute.

These are the same kinds of ideas that have averaged 80 links per campaign, including The New York Times, The Atlantic, NPR, The Independent, and more.

Get your DPR ideas!

---

Sites can lose significant organic traffic from Core updates without obvious patterns when analyzing rank changes. One hidden cause is subtle intent shifts.

Two outdated models prevent us from seeing what's really going on.

One, we often think of user intent as transactional, navigational, informational and commercial. But today, user intent is much more refined and specific.

Two, not all spots in the search results are the same. Google splits the top results into dominant, common and minor intent. Unless we’re talking about longtail keywords with a very clear intent, keywords can have several meanings. Some people searching for “ai ecommerce”, for example, want to know how AI is used in ecommerce while others want to know how ecommerce changes due to AI. Subtle but different. As a result, Google shows a mix of results trying to answer each of those intents as good as possible.

Google reranks search results during Core Updates when it detects that user intent shifts, which can be especially hardcore when dominant intent and the top 3 results are affected. In the grand scheme of things, small rank changes often seem minuscule and chaotic. But what we're seeing is the Butterfly Effect: a small change making a massive impact.

Butterfly Effects from intent shifts

Without knowing what to look for, it’s hard to detect and understand subtle rank changes due to user intent shifts (Butterfly Effects).

Our understanding of keyword meaning is often too static. Topics change all the time. On one end of the spectrum, you have news-related keywords that fall into QDF (query deserves freshness) filters. Look at the news for a day and you see the peak of how fast the meaning of a keyword can change. On the other end, you have evergreen keywords that barely change in meaning.

Take the example “ecommerce ai”. Between early February and mid-April, the dominant user intent changed from "information about AI in ecommerce" to "how ai transforms the ecommerce industry”.

In the screenshot below, I color-coded the different types of intent to highlight how subtle the differences can be.

It makes sense. AI changes rapidly, and so does its impact. People learn about it and come up with more questions. When (dominant) user intent changes, ranks follow suit.

The weighting of intents shown in the SERPs for keywords can also change. A dominant user intent can become a common one and a common minor one. Intent shifts are the reason why Google says that not always are site owners doing something wrong when an update rolls out and there is “nothing to fix” and that you cannot "recover" from an algorithm update. Well.

An even bigger challenge arises when intent shifts and two URLs suddenly cannibalize each other. For example, when an article covers an introduction to a topic ("what is...) and another covers a more comprehensive guide to the topic, they can both rank at the same time or suddenly compete.

How do you find Butterfly Effects? When organic traffic drops after a Core Update, pay attention to keywords that dropped out of the top 3 positions. This is where it hurts the most. Even a single position drop has outsized impact.

To quantify intent shifts, you can analyze the titles of results ranking before and after the update. Titles are not the end-all-be-all, but they give important clues about user intent. At scale, you can use your LLM of choice to categorize titles with prompts like:

• ”Cluster the following titles into one of the following groups: {intent 1}, {intent 2}, {intent 3}, etc.”

• “What intention could users have when clicking on the following search result?“

For very popular queries, we can take it a step further and analyze the trend of searches in Google Suggest to see if search volume is growing or shrinking. Rapidly rising search demand for a related keyword could alter the dominant, common or minor search intent for the root keyword. Note that we don’t know all keywords with growing search volume or where Google’s threshold is for determining that user intent for a keyword shifts. There is a time factor present as well since some keywords rapidly change their meaning (think “independence day” and “wuhan”).

How to act on intent shifts? Once you’ve identified that intent has shifted for a keyword, you have three options:

• Fix potential cannibalization (delete or consolidate)

• Rewrite affected articles to match new user intent

• Create new content based on changed user intent

Core updates are multi-faceted

Refining user intent is not all that Core Updates do. They’ve become kitchen sinks for all sorts of systems:

• The Helpful Content classier has been integrated with Core Updates (link)

• Other quality systems like Panda and Penguin have been integrated with Core Updates years ago

• 15% of searches Google gets are net new, which means Google needs to test the initial results mix and iterate based on user signals.

All of these influences make Core Updates multi-faceted, complex and unpredictable. However, considering the different forms of user intent and subtle changes gives us a path to move from confusion to problem-solution.