Two articles about migrating a shop system. One was written by someone who has done forty migrations. The other by someone who read the first one and six others like it. Both are accurate, both are well structured, and a rater can tell them apart in under a minute. The difference is not expertise. It is experience, and since December 2022 it has its own letter.
Key Takeaways - Experience is first-hand or life experience with the topic, added to the framework in 2022 to distinguish doing from knowing about
- Raters are told to look for signs of effort and original content, and to mark pages that paraphrase other sources as lacking experience
- Evidence is specific and situated: your screenshots, your numbers, what went wrong, where the advice stops working
- Experience cannot be faked convincingly, and attempting to is a trust failure. It can only be shown or added by involving the person who has it
What does Google mean by experience?
First-hand or life experience with the topic of the page. That is the definition in the Search Quality Rater Guidelines, and the example that follows it is the clearest statement of the whole idea:
Google Search Quality Rater Guidelines, September 2025, section 3.4 Experience: Consider the extent to which the content creator has the necessary first-hand or life experience for the topic. Many types of pages are trustworthy and achieve their purpose well when created by people with a wealth of personal experience. For example, which would you trust: a product review from someone who has personally used the product or a “review” by someone who has not?
Google explained the addition in its December 2022 announcement. There are situations where what you value most is content from someone with first-hand experience of the topic. The tax example in that post makes the boundary clear. For how to fill out a tax return you want an expert accountant. For what it feels like to use a particular tax software, you want someone who has done it.
So experience and expertise are different axes:
- Expertise is knowing the subject: the accountant, the electrician, the doctor.
- Experience is having lived the situation: the person who used the product, ran the migration, followed the recipe, recovered from the surgery.
Some topics need both. Many topics, especially reviews, comparisons, how-tos and anything with “what it is like” in the question, need experience more than expertise. We covered the whole framework in E-E-A-T explained; this article is about the letter most sites find hardest to show.
How do raters spot missing experience?
By looking for effort and originality, and marking their absence. The rater guidelines contain a table of low-quality examples, and the pattern in it is unmistakable. One example in there is a list of the best vacuum cleaners. Product images and reviews are paraphrased from other sources, with minimal signs of original effort. The guidelines label it as lacking adequate effort and first-hand experience.
That sentence is the test. A rater asks: could this have been written without doing the thing? If everything on the page could have been assembled from other pages, the answer is yes. The experience score is low, whatever the byline says.
Google’s guidance for product reviews turns the same test into instructions for writers. Google asks for three things there. Evidence of your own experience with the product, such as visuals. Quantitative measurements instead of impressions. And a first-hand reason when you rank something as the best.
The signals raters and readers use, in rough order of weight:
- Original media. Photos, screenshots, recordings you made. A review with the author’s own photo of the product on their desk is a different document from one with the manufacturer’s press image.
- Specific details that only doing reveals. The step that failed, the setting that had to be changed, the time it actually took, the thing the documentation got wrong.
- Numbers from your own measurement. “The export took 41 minutes for 12,000 products” versus “the export can take a while”.
- Limits and failures. Where the method broke, what you would do differently, who it is not for. Summaries do not contain failures because their authors never hit any.
- Consistency with the author’s identity. An author page that shows this person actually does this work, linked from the byline.
Experience is not the same as opinion. “I think this tool is great” is an opinion. “I ran this tool on 140 articles and it flagged 31 outdated prices, 4 of which were false positives” is experience. Raters are looking for the second kind, and readers can tell the difference instantly.
What does first-hand evidence look like in practice?
It looks like the things you could not have written without doing the work. Here is the same paragraph before and after, from a guide on content audits.
Before, no experience signal:
A content audit helps you identify underperforming pages. Start by exporting your pages from Search Console, then evaluate each one for quality and relevance. Many pages will contain outdated information that should be updated.
After, with experience:
The last audit we ran covered 140 pages. Search Console’s export capped at 1,000 rows, which was fine here but would not have been for the client’s 3,000-page site. Of the 140, 31 had at least one outdated price or statistic. The surprise was where they sat. Not in the old evergreen guides, but in the 2024 comparisons. Those were written with current data, and that data expired fastest.
The second version is longer, but not because of padding. Every added sentence is something the reader could not get elsewhere. The row cap, the 31 of 140, the counter-intuitive finding about comparison articles. That is what a rater means by effort and original content.
Five patterns that carry experience, for any topic:
- The counter-intuitive finding. What you expected versus what happened. Only people who did it have one.
- The exact number. Time, count, cost, size. Round numbers read as estimates; precise ones read as measurements.
- The workaround. The documented way did not work; here is what did.
- The scope statement. “This works up to about 500 pages; beyond that the spreadsheet approach falls apart.”
- The artefact. A screenshot with your data in it, a photo with your hands in it, a file you actually produced.
Richard Roth Founder CiteRanger, SEO Strategist The cheapest experience signal is the one most writers cut in editing: the part where something went wrong. I keep a note of every problem during a project, and those notes become the paragraphs readers quote back to me. Nobody has ever emailed me about a paragraph that said “follow these five steps”. They email about the paragraph that said “step three silently fails if the CSV has a BOM, here is how to check”.
Can experience be added to existing content?
Yes, but only by involving someone who has the experience. This is the practical difference between experience and the other on-page signals. An editor can add sources, fix dates and complete the author bio without leaving their desk. An editor cannot add first-hand evidence to an article about a tool they have never used.
Three realistic paths for a site with a backlog of summary-style content:
1. Interview the person who did it
Thirty minutes with the practitioner, recorded, produces the specifics: the numbers, the failures, the workarounds. The editor weaves them into the existing article and credits the practitioner in the byline or as a reviewer. The article changes from “what the internet says about X” to “what we found doing X”.
2. Do a small version of the thing
For how-tos and tool guides, actually run the process once, screenshotting as you go. Half a day of doing produces more experience signal than a week of rewriting. The screenshots alone move the page out of the “paraphrased from other sources” category.
3. Reassign or retire
If nobody on the team has done the thing and nobody can, two honest options remain. Find a guest author who has. Or accept that the page will never carry experience and decide in your content audit whether it earns its place anyway.
What does not work: adding “in my experience” to sentences that contain no experience, inventing anecdotes, or generating “personal” details with an AI. Raters are trained to look for effort and originality. The guidelines name deceptive information about the content creator as a reason for the lowest rating. Fabricated experience is not a weak signal, it is a trust failure. We covered why trust overrides everything else in trustworthiness.
Before publishing, highlight every sentence in the draft that describes something you personally observed, measured or did. If the highlighted portion is under a fifth of the article, the piece is a summary wearing a byline. Either add the missing observations or reframe the piece honestly as an overview.
Where does experience matter most?
Wherever the search intent includes “what is it actually like”. Reviews and comparisons are the obvious case, and the case Google’s review guidance addresses directly. But experience also decides the outcome for:
- How-to guides, where the gap between the documented process and the real one is the whole value
- Tool and vendor evaluations, where the reader wants to know about the failure modes the sales page omits
- Case studies, which are experience by definition and read as fiction without specifics
- Cost and effort estimates, where “it depends” is the summary answer and “it took us 41 minutes for 12,000 products” is the experienced one
- Advice on anything with consequences, where the reader is trying to avoid the mistakes the author already made
For pure definitions and reference content, expertise carries more weight than experience. Even there, one first-hand example lifts the page above the summaries. Something like: in the audit above, this is what the decay curve looked like.
How does experience play into AI search?
AI answer engines are, by construction, summarisers. The pages worth citing are the ones that add something to the existing pool of text. Pages that only restate it are redundant. First-hand evidence is the clearest form of adding something. A number nobody else has. A screenshot of a real result. A failure mode that is in no documentation. Those are the fragments systems quote, because they cannot be derived from anywhere else.
Google’s own AI features apply the same quality foundations as Search, which includes the experience signals raters look for. The selection mechanics are covered in how AI search engines choose which sources to cite. Originality and verifiability were central there, and both are what experience provides.
One caution that ties back to the rest of this site: first-hand numbers age like any other numbers. “The export took 41 minutes” was true on the day. Two versions of the tool later it may not be. Experience content needs the same currency routine as every other page, which is what how often to update content lays out.
Conclusion
Experience is the difference between having done something and having read about it. Google added a letter to its framework specifically to reward the first. It shows in original media, exact numbers, workarounds, failures and scope statements: the things that cannot be assembled from other pages. It cannot be faked without a trust failure, and it cannot be added by an editor alone.
Involve the person who did the thing, keep the part where it went wrong, and put a date next to every number. That is what experience looks like on a page, and it is what raters, readers and AI systems are all looking for.