Before and After: Product Description Rewrites
Four worked examples showing what changes when a product description gets rewritten for search, and why the second version outranks the first.
Most advice about product descriptions stops at the principle. Be specific. Write for humans. Answer the customer’s questions. All true, all frustrating, because the gap between agreeing with that and knowing what to type is where everyone gets stuck.
So here are four rewrites with the reasoning attached. These are constructed examples rather than client work, and I’m not going to attach traffic numbers to them, because any number I gave you would be made up. What I can show is what changes and why the second version gives a search engine more to work with. It’s the same comparison FullSweep SEO puts in front of you as a before/after diff on every change it proposes.
Why is supplier copy so bad for search?
Start with the most common problem in any catalog that grew fast.
Before:
Stainless Steel Mixing Bowl Set. High quality stainless steel construction. Durable and easy to clean. Perfect for all your kitchen needs. Set of 3.
After:
A nesting set of three stainless steel mixing bowls, 1.5, 3 and 5 quarts. The 5-quart handles a double batch of cookie dough without flinging flour over the counter; the 1.5 is the one you’ll actually reach for to whisk two eggs.
Mirror-polished 18/8 stainless, dishwasher safe, and light enough to hold in one hand while you pour. The flat rolled rims mean they don’t stick together when nested, which is the small annoyance that makes cheaper sets irritating. No lids included.
What changed: capacities, alloy, weight behavior, a use case per size, and one honest omission. The first version could describe four hundred products. Every retailer selling this set from the same supplier feed has close to identical text, which means none of them have a reason to outrank the others.
The second version can answer a real search. Somebody typing “mixing bowl big enough for double cookie batch” now has a page that speaks to that.
What about copy that’s well written but says nothing?
The opposite failure. This one usually comes from a brand that hired a copywriter.
Before:
Crafted for the modern woman, the Aria blouse effortlessly transitions from desk to dinner. Timeless elegance meets contemporary ease.
After:
A relaxed-fit blouse in lightweight viscose, cut long enough to wear untucked with high-waisted trousers. Falls to mid-hip on a 5’6” frame.
The white is on the sheer side, so most people wear a nude or white bra underneath. Three-quarter sleeves with a single button cuff. Machine wash cold and hang dry; it comes out of the dryer looking like it needs an iron, and it does.
Runs true to size. If you’re between sizes and want the drape rather than the fit, size up.
Nothing about the voice got destroyed. It’s still readable, still has a point of view. What it gained is sheerness, drape, length on a specific height, sizing guidance and care instructions, which happen to be the five things your customer service inbox is full of.
Sheerness is the interesting one. Saying a white top is a bit see-through feels like an own goal. It isn’t. It’s the exact question people search for before buying, and it’s the thing that generates returns when nobody answers it.
Can a description have too many specs?
Yes, and this is the version that engineers write.
Before:
- 2000mAh Li-ion
- IPX7
- BT 5.3
- 40mm drivers
- USB-C
- 250g
After:
Around 30 hours of playback per charge at moderate volume, and a fast top-up gets you a few hours from a ten-minute plug-in over USB-C. Bluetooth 5.3 holds a stable connection across a normal apartment.
IPX7 rated, which means sweat and rain are fine and full submersion is survivable. Not swimming headphones. At 250g they’re light enough to forget after twenty minutes, though the clamping force is firm and glasses wearers should expect some pressure on the temples for the first week.
Same facts. The difference is that the numbers now have consequences attached. “IPX7” is a rating nobody memorizes. “Rain is fine, swimming isn’t” is an answer.
This matters more than it used to, because when an assistant builds a recommendation it reasons over statements. A page that says what a spec means gets used. A page listing acronyms gets skipped in favor of the review site that explained them.
What about collection pages?
Easy to skip, and they’re often the highest-value page you own.
Before:
Shop our collection of yoga mats.
After:
Nineteen yoga mats across four thicknesses. If you’re on hardwood or practice a slower style with a lot of kneeling, look at the 6mm and 8mm options. Hot yoga usually calls for the open-cell natural rubber mats, which grip better when wet but need more drying time than the PVC ones.
Everything here is 68 inches or longer. If you’re over 6 feet, the 72-inch mats are grouped separately below.
A collection page targets a broader query than any single product beneath it, and one paragraph of genuine buying guidance turns a category listing into something worth ranking. Most stores leave that paragraph empty forever.
Why do the rewrites rank better?
Four reasons, and none of them are tricks.
They’re unique, so they stop competing with everyone else selling the same supplier text. They contain the terms people actually search, because those terms are just the specifics. They match intent, since a page that answers a buying question satisfies the person who searched it, and satisfied visitors don’t bounce back to the results page. And they’re quotable, which is what determines whether an AI assistant can use you as a source instead of a competitor.
How do you do this without breaking what already works?
The genuine risk in a rewrite project isn’t writing badly. It’s overwriting a page that was quietly earning traffic on phrasing you didn’t know it had.
Two habits prevent most of that damage. Check Search Console before you rewrite anything, and keep the queries a page already gets impressions on. Then read the new version against the old one, side by side, rather than replacing text and trusting it.
At ten products you can do that in a browser tab. At a thousand you can’t, which is where the review step usually gets skipped and where things go wrong. FullSweep SEO handles this by showing every proposed change as a before/after diff you approve, edit or skip individually, with one-click revert if something published that shouldn’t have. You can see the flow at how FullSweep SEO works.
The pattern in all four examples above is the same one, and you can apply it this afternoon without any software at all. Take a description. Ask what a customer emails you about that product. Put the answer in the copy. Repeat until the emails stop.
Frequently asked questions
Do I have to write a unique description for every variant?
No, and it usually hurts. Variants of the same product should share one description on one product page, with the differences handled by variant options. Splitting near-identical copy across separate pages creates the duplicate content problem you were trying to avoid.
How long should a product description be?
Long enough to answer the questions people ask before buying, which for most products lands somewhere between 150 and 400 words. Length isn't the ranking factor. Padding a thin description to hit a word count adds nothing a search engine can use.
Is it safe to rewrite descriptions on pages that already rank?
Mostly, if you keep the terms those pages already rank for and only add specificity. The risk is rewriting away the exact phrasing that earned the ranking. Check Search Console for what a page currently gets impressions on before you touch it.