Tip 4: Show Your E-E-A-T

E-E-A-T is a lens, so give it something to look at

Most advice on this topic starts from the wrong place. It treats experience, expertise, authoritativeness and trust as a score Google calculates, then tells you to add trust badges and an author box until the number goes up.

There’s no number. It’s a framework in Google’s Search Quality Rater Guidelines, the handbook human raters use to judge whether search results are any good. Raters don’t change rankings directly. Their judgements help Google check whether its systems are surfacing the kind of pages people would trust. So the useful question for a store owner is simple: when someone lands on your page, what do they see that proves you know this product?

For an online store, the honest answer is usually “less than we actually know”. The evidence exists. It’s sitting in your returns spreadsheet, your customer service inbox and the head of whoever chose the fabric. It just never makes it onto the page.

This tip is about moving that evidence into your content. If you want the separate question of keeping AI-written copy factually true to your brand, that’s covered in keeping AI content accurate to your brand. Here we’re dealing with proof.

What each letter asks of a store

The four parts overlap, and the guidelines put trust at the centre: a page that seems untrustworthy doesn’t get rescued by expertise. For ecommerce, they translate roughly like this.

• Experience: has someone actually used, worn, washed, fitted or broken this product? First-hand detail is the hardest thing for a competitor or a content farm to copy.
• Expertise: does the writer understand the category well enough to explain why, as well as what? Knowing that 17.5 micron merino feels softer than 21 micron is expertise. Saying “premium soft wool” isn’t.
• Authoritativeness: do other people treat you as a source? Reviews, press, links and customers who name you when they talk about the category. You can’t write this into a page, but you can make pages worth citing.
• Trust: is the page accurate, honest about limits and clear about who’s selling? Delivery costs, returns terms, contact details and claims you can back up all sit here.

Notice that only one of those is mostly outside your control. The other three are editorial choices.

The evidence your store already has

Before writing anything new, list the sources of first-hand knowledge you already hold. Most stores have far more than they publish. A typical inventory looks like this:

• Returns data: the reason codes tell you exactly where expectations and reality diverge. “Too small” clustered on one style is a sizing insight you can publish.
• Fit and feel feedback: review text, exchange requests and post-purchase survey answers, especially the ones that mention body shape, activity or climate.
• Customer service questions: the questions asked three times a week are the questions your product pages fail to answer.
• Product testing notes: wash tests, wear trials, samples you rejected, the supplier you dropped and why.
• Team knowledge: the buyer who has handled 40 fabric swatches this season, the founder who designed the first version, the packer who knows which boxes arrive crushed.
• Sales patterns: which sizes sell out first, which colours get returned, what people buy together.

None of this needs a research budget. It needs someone to read it with a writer’s eye and ask which bits a shopper would want to know before buying.

A worked example: merino base layers

Take a fictional store selling merino base layers. Its buying guide for “best merino base layer for hiking” currently reads like every other result: merino regulates temperature, resists odour and is naturally soft. All true. All available on 50 other pages.

Now run the evidence inventory against it.

1. Returns data shows the 150gsm crew is returned for being “too thin” far more often than the 200gsm. That becomes a section on fabric weight, with a plain recommendation: 150gsm for high-output summer hiking, 200gsm if you stop often or hike above the tree line in spring.
2. Fit feedback shows customers with broad shoulders size up in the women’s crew. That goes into the sizing section, attributed to customer feedback, with the measurement where it starts to matter.
3. Testing notes from the team’s own wear trials record that the 100% merino top pilled under a rucksack hip belt after a few weeks of use, while the merino-nylon blend didn’t. That’s a trade-off worth stating, even though it makes one product look worse.
4. Customer service keeps getting asked whether the tops can go in the tumble dryer. A care section answers it, with what actually happens if you do.
5. A named author: the guide is signed by the store’s product lead, with a two-line bio saying what she’s responsible for and how the tests were run.

The rewritten guide is barely longer. It’s far more useful, and almost nothing in it could be produced by summarising the pages already ranking. That’s what experience looks like on a store: specific, occasionally unflattering and traceable to a real person or dataset.

Named authors who have earned the byline

An author box helps only when the author is real and relevant. A generic “Team Sprite” or “Admin” byline tells the reader nothing, and an invented persona with a stock photo is worse, because it’s a small lie sitting at the top of the page.

What works is modest and true. Name whoever knows the subject best, say what they do at the company, and link to a short author page listing what they’ve written. If the product lead didn’t write the words but supplied the testing notes and checked the draft, say “reviewed by” and name her. Readers understand that. Sprite’s own author page for Richard Newton is an example of the plain format: who he is, what he works on, what he’s published.

For products where a bad recommendation could hurt someone, the bar rises. Children’s sleep products, supplements, mobility aids and anything electrical fall closer to what the guidelines call “your money or your life” topics. There, a reviewer with a relevant qualification or clear professional experience is worth arranging, and claims should be conservative.

Trust lives in the boring details

Raters are told to look for who’s responsible for a site and how to contact them, especially where money changes hands. Shoppers do the same thing, just faster. A surprising amount of trust is decided by pages nobody thinks of as content.

• Delivery costs and times visible before checkout, and consistent across the site.
• A returns policy written in plain language, with the actual window and who pays postage.
• A physical address and a working contact route, on a page linked from the footer.
• Reviews shown unfiltered, including the three-star ones. A wall of five-star reviews reads as curated.
• Claims that match the product. If the page says “machine washable at 40°C” and the care label says 30°C, trust is gone the moment the jumper shrinks.
• Dates on guides that go stale, so a reader knows when “this season’s range” was written.

Structured data can make some of this machine-readable: Product markup for price and availability, Organization details, review markup where it reflects genuine reviews. It describes evidence that’s already visible. It can’t create evidence that isn’t there.

An evidence audit you can run this week

Pick your five highest-traffic informational pages in Search Console and your five best-selling product pages. For each one, go through these checks and note the gaps.

1. Does the page contain at least one detail that could only come from handling the product? Circle it. If you can’t find one, that’s the first fix.
2. Is there a named person responsible for the content, with a reason to trust them?
3. Does it state a limitation or trade-off? Pages that only praise read as sales copy.
4. Are the top three customer service questions about this product answered where a shopper would look?
5. Do the factual claims match the spec sheet, the care label and the returns policy?
6. Would a first-time visitor know who runs the store and how to reach them within one click?

Then fix in that order. Adding one real testing note to a buying guide usually does more than redesigning the About page, because it changes what the page says rather than what surrounds it.

Where AI fits, and where it can’t

Google has said plainly that it cares about the quality of content rather than how it was produced, which we cover in more detail in does Google ban AI content. The catch is that a language model has never worn your base layer. It can only restate what it’s given.

So the split is practical. People supply the evidence: returns patterns, testing notes, fit feedback, the judgement of whoever handles the product. Automation can then carry that evidence consistently into every relevant guide, product description and FAQ, which is the part small teams struggle to keep up with. Feed a tool nothing but product titles and you get the same generic page as everyone else. Stories are one of the best containers for this material, which is why Tip 3: Tell Stories pairs so closely with this one, and the depth that evidence makes possible is the subject of Tip 6: Go Deep.

At Sprite, that’s how we treat brand inputs: the testing notes, sizing quirks and customer language a store supplies become source material the writing draws on, with a named reviewer where the subject calls for one. The expertise stays yours. We just make sure it turns up on the pages where shoppers are deciding.

---

This is Tip 4 in Sprite’s 11-part series on ecommerce content that compounds. Next: Tip 5: Cluster Up.

Frequently asked questions