Write for the Machine, Sell to the Human
Every guru tells you to 'write for your customer, not the algorithm.' It’s a nice sentiment, and it’s completely wrong. You have to do both, at the same time, in a single block of text. Treating them as separate tasks is why so many product pages are either unreadable keyword salads or beautifully written sales copy that nobody ever finds. The real skill isn't prioritizing one over the other; it's about building a cage of algorithmic constraints and then writing compelling prose inside it.
The Algorithm Isn't Your Audience, It's the Bouncer
["Think of the search algorithm as the bouncer at an exclusive club. The bouncer doesn't care about your personality, your witty conversation, or how well you dance. They have a simple checklist: Are you on the list? Do you meet the dress code? Are you of age? If you tick those boxes, you're in. If you don't, you're left on the curb, no matter how charming you are.", "On Amazon, Etsy, or Google Shopping, the algorithm is that bouncer. Its job is to filter billions of products down to a few dozen relevant results. It uses a crude but effective checklist: keywords in the title, completed attribute fields, sales history, image quality. It’s not reading your copy for emotional resonance. It’s pattern-matching. Your first job is to give the bouncer exactly what's on their checklist so you can get inside where the actual customers are.", "This is why obsessing over a perfect opening sentence before you've mapped your keywords is a waste of time. Your beautiful prose is for the customer, but the customer will never see it if the algorithm doesn't let your listing through the door. The machine is a gatekeeper, not a connoisseur. Treat it as such. Feed it the structured data and keywords it needs to see, clearly and unambiguously."]
The 'Write for Humans First' Fallacy
["The most common, and most destructive, piece of advice in this space is to write your perfect, human-centric copy first and then 'sprinkle in some SEO' later. This is fundamentally backward. It’s like building a house and then trying to add the foundation. You end up with a compromised mess that serves neither purpose well. Your elegant sentences get mangled by shoehorned keywords, and the SEO feels unnatural because it’s an afterthought.", 'The professional approach is to start with the constraints. Begin with the non-negotiable elements the machine requires. What is your primary keyword? What are the five most important secondary keywords? What are the character limits for the title, bullets, and description on your target platform? What specific attributes—like material, color, or size—must be included for filtering?', 'Build a skeleton using these hard points. This keyword goes in the title. These three go in the first three bullet points. This one must be in the first sentence of the description. Once you have that rigid structure, your creative task is to write compelling copy that connects those dots. It’s a more challenging puzzle, but the result is a cohesive listing where the SEO and the sales copy are one and the same, not two warring factions.']
Structured Data: The Machine's Native Language
["Sellers obsess over the visible text—titles, bullets, descriptions—but often neglect the most important part for algorithmic discovery: structured data. These are the backend fields, item specifics, and attribute tags that customers rarely see directly but use constantly for filtering. On eBay, they're called Item Specifics. On Etsy, Attributes. On Amazon, they're buried in the flat file or the backend UI. This is where you speak directly to the machine, and it's not a conversation you should ignore.", "This isn't just about helping users filter. The platforms themselves use this data as a primary signal for relevancy. Etsy's 'Odyssey' search update in Q4 2025, for example, began heavily penalizing listings that had a mismatch between their user-selected attributes (e.g., tagged as 'Linen') and the keywords used in the description. Our internal tracking at rankd showed that sellers with these mismatches saw an average 8-12% drop in search impressions in the first month after the rollout. The algorithm is checking your work.", "Think of it this way: your prose is for persuasion, but your structured data is for categorization. If you sell a '10-foot braided USB-C cable,' the prose describes why your braiding is superior and the connectors are gold-plated. The structured data, however, just needs to state: Cable Length: 10 ft, Connector Type: USB-C, Material: Braided Nylon. Filling out every single one of these fields is the single highest-leverage activity for getting your product properly indexed and shown in filtered searches, which is where the highest-intent buyers live."]
An Actionable, Machine-First Workflow
- Step 1: Keyword Harvesting. Before writing a single word, use a tool (or just the platform's search bar) to identify one primary keyword, 3-5 secondary keywords, and a handful of long-tail and semantic variations. Do not skip this.
- Step 2: Field Mapping. Create a simple map of where each keyword will live. The primary keyword must be in the title. The top three secondary keywords must be in the bullet points. Long-tail keywords are for the description and backend search terms.
- Step 3: Write to the Skeleton. With your map in hand, now you write. Your goal is to weave these required keywords into readable, persuasive sentences within their assigned locations. The constraint makes you more creative, not less.
- Step 4: Populate All Structured Data. Go to the backend of your listing and fill out every single field available, even if it seems optional. Match this data perfectly to your product's actual specifications.
- Step 5: Write the 'Invisible' Copy. Fill out the backend search term fields on platforms like Amazon. This is pure machine food. You don't need grammar or prose; just space-separated keywords and synonyms you couldn't fit elsewhere.
- Step 6: Readability Check. Read the entire customer-facing listing out loud. If it sounds clunky or robotic, revise the connecting language, but do not remove your mapped keywords. The keywords are the pillars; the rest is just drywall.
Amazon's A10 and the Power of Invisible Text
["Amazon's algorithm, currently A10, is notoriously opaque, but we know it places immense weight on text the customer never reads. Backend Search Terms and Subject Matter fields are essentially a direct line to the algorithm, free from the need to sound human. This is your chance to be a robot. You can include common misspellings, Spanish-language equivalents, and hyper-specific long-tail phrases that would ruin your customer-facing copy.", "Don't waste this space. The 'Search Terms' field allows for 249 bytes of space-separated keywords. Use all of it. Don't repeat words from your title or bullets; A10 already indexes those. This is for net-new terminology. If you sell a garlic press, your title has 'garlic press.' Your backend terms should have 'mincer crusher peeler rocker stainless steel professional'.", "Furthermore, the placement of keywords in your *visible* copy still matters immensely for initial indexing. Our scrb data, analyzing over 50,000 top-performing listings in the 'Home & Kitchen' category, shows a 22% higher click-through rate from search for products that include their top three most-searched keywords in the first 150 characters of the product description. This suggests that even if customers don't read the whole description, the algorithm does, and it rewards early, clear signals of relevance."]
Character Counts Are a Feature, Not a Bug
["Sellers constantly complain about character limits—Etsy's 140 for titles, Amazon's ~200 recommendation, Shopify's 70-character SEO title limit. They see it as a restriction on their creativity. It's not. It’s a guardrail designed to force clarity and prevent the kind of keyword-stuffing nonsense that made search results unusable in 2010.", "A hard limit forces you to make choices. You can't be a 'Premium High-Quality Professional Ergonomic Best-Selling Stainless Steel Garlic Press Mincer and Crusher for Home Chefs'. You have to be a 'Professional Stainless Steel Garlic Press'. The constraint prunes the fluff. It demands you identify the single most important terms a user would search for and put them front and center.", "The best copywriters embrace these limits. They use them as a creative challenge to pack maximum information into minimum space. They use pipes (|) or dashes (-) to create scannable title sections. They understand that on a mobile search results page, only the first 60-70 characters are likely to be visible anyway. Writing a tight, impactful title isn't just good for the algorithm; it's good for the hurried, distracted human who's scrolling on their phone."]
When to Update: The Myth of 'Set It and Forget It'
["Product copy is not a static asset. Search trends change, competitor strategies evolve, and platforms update their algorithms. A listing optimized for Etsy's search in 2024 might be underperforming by mid-2026. You have to treat your listings like living documents, subject to periodic review and revision.", "A good rule of thumb is to perform a copy audit quarterly. Look at your analytics. Are impressions dropping for your main keywords? Has a new competitor entered the space with a different angle? Use your platform's search analytics (like Etsy Search Analytics or Amazon's Search Query Performance dashboard) to see the literal terms customers are using to find you. You might discover they're searching for 'gift for dad' in November, which warrants a temporary title tweak.", "Don't perform wholesale rewrites unless a listing is dead. Instead, iterate. Try changing the order of clauses in your title. Swap out the keyword in your third bullet point for a new contender. Change the main image and see if it affects click-through rate. Small, controlled changes allow you to isolate what works without resetting the listing's indexing and sales history with the algorithm."]
FAQ
Can I just use an AI writer for all my product copy?
You can, but you shouldn't use it blindly. AI is excellent for generating a first draft based on a keyword skeleton or for rewriting a boring paragraph. It is terrible at understanding your brand's specific voice and identifying the most critical, high-intent keywords without human guidance. Use it as a tool to overcome writer's block, not as a replacement for strategy.
How important are bullet points versus the main description?
On Amazon, bullet points are vastly more important. They appear high on the page ('above the fold' on desktop) and are the most-read piece of copy after the title. On Shopify or your own site, the format is more flexible, but the principle of scannable, benefit-oriented points remains critical for user experience.
If I update my copy, will I lose my search ranking?
It's possible to see a temporary dip, which is why you should avoid complete rewrites of successful listings. The algorithm needs time to re-index the changes. However, if your listing is underperforming, the potential upside of a copy refresh that better targets user intent almost always outweighs the risk of a temporary shuffle in the rankings.
Does keyword density still matter at all?
Not in the old sense of 'repeat the keyword 10 times.' Modern algorithms are smart enough to understand synonyms and context. What matters is keyword *presence* in key locations: title, bullets, backend fields, and the first paragraph. Repeating a term excessively is a negative signal that makes you look like a spammer to both the machine and the human.