What product data do AI assistants and publishers actually need?

The texture and porosity the product is formulated for, the key ingredients, price and availability, and a stable product URL. The texture range is the field most often missing and the one that decides whether a recommendation reaches the right person.

Which schema type should a haircare product page use?

Schema.org's Product type. Google documents how product markup is used for product snippets and how review markup is used for review snippets. Include name, image, description, brand, and an offer with price, currency, and availability.

What are positiveNotes and negativeNotes?

They are properties Google documents for expressing the pros and cons of an editorial product review in product snippet structured data. Their existence means the pros-and-cons format is a recognised shape, and a review with only positives leaves a documented field empty.

Can structured data claim things the page does not say?

No. Markup has to describe what is actually on the visible page. Asserting a rating, price, or claim in the markup that the page does not contain is a guidelines violation rather than a clever optimisation.

Why does narrowing my texture range help sales?

A vague product gets recommended vaguely, bought by people it does not suit, and collects reviews saying it did nothing. A precisely described product reaches people it was made for. Precision reduces returns and improves the reviews that future buyers read.

What breaks a publisher placement most often?

A dead product URL after a site redesign or a discontinued line nobody flagged. Telling a publisher that a product is gone, before a reader clicks and finds out, is the most useful message a brand can send.

Is it acceptable to ask a publisher to remove a paid label?

No, and it damages the relationship. You would be asking them to carry a legal and reputational risk for your benefit. Ask how partners label paid and gifted coverage before the campaign, not after.

What image format should product images be supplied in?

Square, plain or transparent background, 1000px or larger on the short side, from the original file rather than a compressed web copy. Detail removed by an earlier compression pass cannot be recovered later.

Will fixing product data make a mediocre product sell?

No. Good data makes a product easy to describe accurately, which means the right people find it and the wrong people do not. It is not a substitute for the product being good.

Getting your product data into a shape machines can use

The field almost every haircare brand leaves out

Texture range. Not curly, but the actual range, such as suits 3B through 4A, or built for high porosity hair.

When it is missing, everything downstream has to guess. A publisher building a routine guesses. A retailer's filter guesses. An assistant summarising options guesses, and it will usually guess from your marketing copy, which was written to sound broad rather than to be accurate.

Brands leave it out because narrowing feels like shrinking the market. In practice the opposite happens. A vague product is recommended vaguely, gets bought by people it does not suit, and collects reviews saying it did nothing. A precisely described product is recommended precisely and collects reviews from people it was made for.

If you do one thing from this page: publish the texture range on every product page, in text, in the words shoppers use.

The structured data layer

Schema.org's Product type is the vocabulary for describing a product to a machine. Google documents how this markup is used to generate product snippets and, for review content, review snippets.

The basics worth getting right are the product name, an image, a description, brand, an offer with price and currency and availability, and a stable identifier. If reviews appear on the page, mark them up honestly rather than aggregating numbers the page does not actually show.

Google's product snippet documentation also describes positiveNotes and negativeNotes properties, which express the pros and cons of an editorial product review. That is worth knowing for two reasons. It means the pros-and-cons shape is a recognised format rather than a stylistic choice, and it means a review that only lists positives is leaving a documented field empty.

The hard rule underneath all of it: markup describes the page. If your structured data asserts a rating, a price, or a claim the visible page does not contain, that is a violation of the guidelines rather than an optimisation.

What a publisher partner needs from your feed

Separate from the search question, this is what a partner like Kurli needs in order to place a product without guessing.

Stable product URLs. A link that survives a site redesign. Redirect chains are tolerable; dead links are not.
Texture and porosity mapping per product, including the textures it does not suit.
Category, in normal words: cleanser, conditioner, leave-in, cream, gel, oil, tool.
Price and the markets you ship to, so readers are not sent somewhere that cannot serve them.
A square product image at 1000px or larger, from the original rather than a compressed web copy.
Discontinuation notices. The single most useful message a brand can send a publisher is this product is gone, before a reader finds out by clicking.

Full specifications are on the media kit page.

Where disclosure fits into the data

If a piece of content about your product is incentivised, that has to be visible in the content, and it should not be laundered out by the markup.

Google's structured data guidance tells publishers not to include fake or undisclosed incentivised reviews on a page or in the markup. The FTC's endorsement guides require disclosure of a connection between the seller and the endorser that could affect the weight a reader gives an endorsement and that the reader would not already expect, which covers gifted product and paid placement.

For a brand this means two practical things. Ask your partners how they label paid and gifted coverage, before the campaign rather than after. And do not ask a publisher to soften or remove a label, because you are asking them to take on a risk for your benefit, and the good ones will say no and remember that you asked.

A short audit you can run on one product page

Take your best seller and check:

Does the visible page state the texture range in plain words?
Does it state who the product is not for?
Are the key ingredients in text, not only in a packaging image?
Is there valid Product structured data, and does every field in it match the visible page?
Is the price and availability correct for every market the page is served in?
Is the product URL the one you would still be using in two years?
If reviews are shown, is any incentivised review labelled as such?

A page that passes all seven is easy for a search engine, an assistant, and a human partner to describe accurately. That is the whole objective. Nothing here makes a mediocre product succeed, and nothing here is a substitute for the product being good.