How to Write Product Descriptions

Product descriptions are commercial writing's highest-volume assignment and its most defaulted: the manufacturer paragraph pasted (the boilerplate economy), or the spec list transposed, or the adjective fog ("premium quality craftsmanship") that describes nothing — while the description's actual job sits unattempted: letting the buyer feel the product answering their situation. Here's the writing craft (the template strategy's editorial half): the buyer-situation method, the FAB engine, and the sensory-and-proof layers.

The method: write to the situation, not the object

The description's organising question isn't "what is this?" but "who is deciding what, right now?" — derived per product from buyer-language research, reviews and support questions: the hiking-boot buyer deciding between waterproof-enough and overkill; the gift buyer needing confidence without expertise. The opening line meets that situation (the confirm move at product scale: "Built for day-hikers who cross streams, not summits"), and the structure follows the decision: the fit paragraph (who this is for, honestly — including the who-it-isn't that makes the recommendation trustworthy, per the honesty-converts rule), the FAB-translated feature core (each spec pushed to what-it-means: "700-fill down" → "warm at -10° without the bulk"), the sensory-and-scale layer (weight in hands, size against familiar objects, the texture words online buying can't supply — the imagination bridge that photographs alone don't build), and the objection line (the one hesitation this product's buyers actually have, answered — sizing, care, compatibility — mined from the Q&A stream).

The craft rules at catalogue scale

1. Triage per the revenue tiers: this full method for the head, attribute-driven honest templates for the middle — the craft applied where it converts, per the standing economics.
2. Specificity is the anti-boilerplate: every sentence tested against "could this describe the competitor's version?" — the numbers, materials and use-cases that fail that test are the description working.
3. Keywords as vocabulary, never as inserts: the buyer-language research supplies the terms naturally (buyers search "waterproof" because they say "waterproof"); the description written to the situation contains its keywords by construction, per the silo's standing rule.
4. Scannable by default: the short-paragraphs-and-bullets formatting layer — description prose above, spec bullets below, both present because both reader modes shop.
5. Voice held: descriptions are the tone guide's hardest test — a hundred products in a voice, without the template smell; the register-map's product entry earns its keep here.

Frequently asked questions

How long should a product description be?

The decision's length: considered purchases (the situation, translation, objection layers earn 150–300 words); commodity items (a tight fit-line plus specs — padding a USB cable's description helps no one). The completeness-not-count rule at product scale.

Can AI write our product descriptions?

The standing division: rendering real attribute data into readable middle-tier prose, yes with editing; the head tier's situation-and-objection craft runs on buyer knowledge generators don't have — generated head-product copy converges on exactly the adjective fog this guide exists against, per the help-and-hurt map.

Do well-written descriptions actually rank product pages?

They supply the template contest's uniqueness and long-tail vocabulary (buyer-situation language matching buyer-situation queries — measurably), while the ranking heavyweights stay the page's usual ones: the category above it, the review corpus, the domain's authority (our shelf). The description's first job remains the conversion — which is the metric to grade it on.