Hey there 👋

A month ago, in the information gain issue, I pointed out that Google had never used that phrase in its guidance. What it did have was a plain question: does your page add substantial value compared to the other pages in search results?

On October 2, Google went a step further by adding two new sections to its "Creating helpful content" guide. One explains what counts as the main content of a page and why that's what quality is judged on, while the other lists the four things Google's human reviewers are trained to look for, which include effort, originality, talent or skill, and accuracy.

One line in there will sting a lot of content teams: "Attribution or giving credit to other sources doesn't replace the need for original effort."

That's Google describing the carefully sourced summary post, which a lot of marketing blogs are built on, and saying it isn't enough.

This week, I'll go through what Google actually changed, what it doesn't mean, and how to run the four tests on your best page.

Let's go. 🚀

Alongside this newsletter, I help financial brands get found on Google, ChatGPT, Perplexity, and Google AI Mode, either directly or through my agency, Rise Up Media.

If you'd like help with yours, tell me what you're working on. It takes two minutes.

Work with me →

ChatGPT is citing original sources over the sites that summarize them. Ahrefs tracked US ChatGPT citations from March to September and found Reddit's share of answers fell from 19.1% to 2.3%, and Wikipedia's from 14.1% to 2.9%. PubMed climbed from 64th to 4th among the most-cited domains, while Investopedia, Britannica and the New York Post dropped out of the top 200. The data comes from free, logged-out ChatGPT before GPT-6 launched on October 7, so the newest model may behave differently. It still fits this week's theme neatly: the summarizers are losing ground to the sources they summarize. [Source: Search Engine Journal]

Google now tells you to fact-check every piece of AI-generated content by hand. Alongside the helpful content update, Google revised its guidance on using generative AI on your site. The new wording says "it is critical to manually factcheck and review all AI-generated content" for accuracy and trustworthiness before publishing, because AI output can contain hallucinations. Google says the change brings the document in line with what it already tells developers at events. [Source: Search Engine Roundtable]

Google's September spam update finished after 13 days and 16 hours. It started on September 24 and was complete as of October 8, making it the longest of Google's four spam updates this year. Google hasn't said what kind of spam it targeted. If your traffic dropped in that window, the place to start is Google's spam policies, not a content rewrite. [Source: Search Engine Journal]

Google has now written down the four things its reviewers judge on every page. Each step below takes one of them and turns it into a check you can run on your own content this week.

First, it's worth being clear on what changed, because this update will get misreported.

The first new section defines main content as "any part of the webpage that directly helps the page achieve its purpose." That includes the obvious, like the article text, images, and video. It also includes things content teams often forget to count, such as calculators and other interactive tools, user reviews and forum discussions when they're the point of the page, information inside tabs and expandable sections, and the page's own title and headings.

The second section lists the four attributes raters are trained to evaluate in that main content: effort, originality, talent or skill, and accuracy. The guide was last updated on October 5.

What this update doesn't do is give you a list of ranking factors.

Google's own page is clear: "Search raters have no control over how pages rank. Rater data is not used directly in our ranking algorithms."

Raters are how Google checks whether its systems are working. So the right way to read the four attributes is as a public description of what Google's ranking systems are trying to reward. That's still worth acting on, but anyone selling you "the four new ranking factors" is overselling it.

Here are the four tests.

Google defines effort as "the extent to which human work went into creating the content or the systems powering it."

Its examples of high effort are writing original analysis, translating a poem by hand, and building a custom interactive map. Its examples of little to no effort are generating pages automatically from feeds, and using AI to produce large amounts of text "without manual oversight or curation."

Then comes the line that matters most for content teams: "Attribution or giving credit to other sources doesn't replace the need for original effort."

That rules out a format most marketing blogs depend on: the post that gathers ten statistics from ten studies, credits every one of them properly, and adds nothing of its own. It's well sourced, and by this definition it's still low effort.

The check: open your best page and list every element that couldn't have been produced in an afternoon from other people's pages. Your own data, a test you ran, a tool you built, a screenshot from your own account, a worked example from a real client. If the list is empty, the page is a summary, however good the citations are.

Google's definition is "the extent to which the content offers unique, original information or perspectives that aren't already available on other websites."

That's the information gain argument from September, now in Google's own guidance. Last month I quoted Google asking whether your content provides "substantial value when compared to other pages in search results." This version is more direct: if the information already exists on other sites, it isn't original.

The Ahrefs data in this week's news shows the same pressure from the AI side. Investopedia, one of the best-known financial explainer sites on the web, has dropped out of ChatGPT's top 200 cited domains. Ahrefs' read is that ChatGPT "seems to be prioritizing original sources over sites that summarize them."

The check: search your page's main question on Google and in ChatGPT, open the top five results and every source cited, and write down what they all say. Then mark the parts of your page that aren't on that list. For most pages that's a small share. That small share is the bit worth building out.

Google's wording is "whether the content shows the talent, skill, or expertise necessary to provide a satisfying experience for visitors." Its examples are clear writing, well-produced video, and page tools that actually work.

Google also adds something useful: "not all content requires specialized expertise." Its example is a person sharing a new way to clear snow from their driveway. You don't need credentials for that, just the experience and the ability to explain it well.

For financial brands, though, this one usually comes down to execution. A fee calculator that breaks on a phone fails this test, no matter how good the surrounding article is. So does a comparison table that's out of date, or a guide written so cautiously that it never actually answers the question.

The check: open the page on your phone and use it the way a customer would. Run every calculator with real numbers. Read the opening paragraph out loud and ask whether it answers the question someone searched for. Anything clumsy, broken or vague goes on the fix list first, because it's usually the cheapest of the four to improve.

For informational pages, Google says content "should be factually accurate." For topics that could significantly affect people's lives or well-being, which Google calls YMYL ("Your Money or Your Life"), the bar goes up. The content "must be highly accurate and consistent with established expert consensus."

Finance is YMYL by definition. So every fee, rate, tax rule, and product claim on a financial brand's site is held to the higher of the two bars.

Two practical consequences. First, "consistent with established expert consensus" means a contrarian take needs more evidence than a mainstream one. You're allowed to disagree with the consensus, but say so explicitly and show why. Second, Google's revised AI guidance in this week's news now says to "manually factcheck and review all AI-generated content." If AI drafts any of your financial content, a person has to check every number before it goes live.

The check: highlight every number on the page and confirm each one links to its primary source and has a date. Then ask who would put their name to the page as a reviewer. If nobody would, or nobody has, that's the gap.

Run these on the page that matters most to your business. A "no" on any of them tells you where to spend your next week.

This prompt runs the effort and originality checks for you, against the pages you're actually competing with.

The Scenario: You've got a page that matters to your business, and you want to know how it measures up against the pages Google and AI engines already show for the same question, before you decide what to rewrite.

The Prompt:

Run it on your most important page first. Expect step 5 to take the longest, and to be the most useful.

These are the most popular AI tools we use at Rise Up Media. If you're not using them already, they're worth a look.

Full disclosure: some links above are affiliate links. If you sign up, we'll earn a small commission at no extra cost to you.

Find the parts of your page that took real work, check what's on it that isn't anywhere else, use it like a customer would, and get every number checked by someone who'd sign off on it.

None of this is actually new advice. What's new is that Google has written it down in plain language, including the line that citing other people's work doesn't count as your own effort.

For financial brands, accuracy is the one to start with. Google's bar is highest here, and a single wrong fee or rate can undo everything else on the page.

Until next time, keep exploring the horizon. 🌅

Alex Lielacher

P.S. If you want your brand to show up in Google AI Mode, ChatGPT, and Perplexity, reach out to me.