Size Guidance
by
Fit advice for ecommerce teams: read where returns cluster by size and reason, then rewrite the guidance on the SKUs that are actually costing you.
Size-related returns are the most expensive kind, because the item comes back, the shipping goes both ways, and the customer usually does not reorder. Most stores answer this with one size chart applied to a catalog made by different manufacturers to different specs, which is why a customer who takes a medium in one of your products takes a large in another and finds that out by returning something. The information needed to fix it is already in your returns data. This skill reads where returns cluster by size and reason, then rewrites the guidance on the SKUs costing you most.
What It Does
- Groups returns by SKU, size and stated reason rather than in aggregate.
- Finds the SKU whose sizing departs from the rest of your catalog.
- Prices the return cost per SKU including both shipping legs.
- Detects the size that is consistently ordered and sent back.
- Writes guidance specific enough to change the size someone picks.
- Rewrites the size advice per SKU, with the return cost it addresses.
Use Cases
- Find the SKU that runs a size small and says nothing about it.
- Cut the returns on the product costing most in two-way shipping.
- Replace a generic size chart with per-product guidance.
- See which size is being ordered and returned most often.
Example Session
Which products need better size guidance, and what should it say?
Three SKUs carry most of the size returns. One runs a full size small.
- SKU | Size return rate | Pattern | Cost per quarter | Guidance to add
- Jacket, navy | 31% | M returned, L reordered | $41,000 | Runs small, size up
- Tee, boxy fit | 22% | L returned, M reordered | $18,000 | Oversized cut, size down
- Trouser, slim | 17% | Length, not waist | $12,000 | Inseam by height
- Cap, one size | 2% | None | $900 | No change
The navy jacket has a clean signal: customers order a medium, return it, and reorder a large, which is a full size and worth $41,000 a quarter in two-way shipping alone. One sentence above the size selector fixes most of it. The trouser problem is different: the waist is right and the inseam is not, so the guidance needs to be length by height rather than a size recommendation.
Why This Beats Prompting It Yourself
One Size Chart vs Per-Product Fit Reality: The Hidden Risk
A generic prompt improves the size chart, which is the visible artifact and the wrong level. Your catalog comes from different manufacturers cutting to different specs, so a single chart is wrong on whichever products depart from it. The customer discovers the departure by returning something, which is the most expensive way for anyone to learn it.
When You Still Need The Reason Behind The Return
Size returns and fit returns are not the same problem. The trouser is returned for length while the waist fits, so a size recommendation would not help and length by height would. Reading the stated reason rather than the size exchanged is what separates a guidance fix from a product fix.
Known Limitations
Return reason quality depends on what your returns flow captures, and free-text reasons are parsed with the usual limits on short entries. The reorder signal needs returns linked to subsequent orders by customer, which some setups do not maintain. It writes the guidance and does not publish it to your storefront.
Install It
- Download the ZIP.
It is free and there is no account to create. - Unzip it into your agent's skills folder.
Claude Code reads ~/.claude/skills/, which is hidden by default: the command in the folder block opens it. Other agents scan their own directory, so drop the same folder there instead. - Ask your agent to use it.
Restart the agent if it was already running, then it picks the skill up with no config.
skills folder
Copy
~/.claude/skills/size-guidance/
SKILL.md
references/fit-patterns.md
guidance/
# one file per SKU, with the copy to add
# macOS: create the folder and open it in Finder
mkdir -p ~/.claude/skills && open ~/.claude/skills
# Windows: paste in the Explorer address bar
%USERPROFILE%\.claude\skills\Frequently asked questions
How is this different from improving the size chart?
A chart describes measurements; guidance tells someone what to do. "Runs small, size up" changes the size in the cart and a measurement table usually does not. This also works per SKU, which is the level at which fit actually varies.
Which agents does the skill run in?
Any agent that supports the open SKILL.md format: Claude Code, Cursor, Codex CLI, GitHub Copilot, Gemini CLI, Manus, Grok and others load it unmodified. The format is portable, the location is not. Each agent scans its own skills directory, so you drop the same folder into whichever one yours uses. There is no config file to edit and no API key to provision.
What does it need connected to work?
Returns with SKU, size and reason, orders linked to customers so reorders can be matched, and shipping cost. Free-text return reasons work and structured ones work better.
How is 31% returns worth $41,000?
Both shipping legs plus handling on every returned unit, and the margin on the customers who return and do not reorder. The skill shows the components so the figure can be checked rather than taken on trust.
Will guidance actually change what people order?
A specific instruction placed next to the size selector does. A measurement table three clicks away does not, and neither does a general note on a separate sizing page. Placement is part of the recommendation for that reason.
What is included with the skill?
The skill itself, the fit pattern reference, the per-SKU guidance format, and the return cost model. It is free and security scanned. We re-publish the ZIP when the model changes, so download it again if your rankings start shifting.
Returns cluster where the size advice is wrong.
Polar joins return reasons to SKU margin across 45+ sources, so you rewrite the guidance on the products where it actually costs you.
<script type="application/ld+json">{"@context":"https://schema.org","@type":"FAQPage","mainEntity":[{"@type":"Question","name":"How is this different from improving the size chart?","acceptedAnswer":{"@type":"Answer","text":"A chart describes measurements; guidance tells someone what to do. \"Runs small, size up\" changes the size in the cart and a measurement table usually does not. This also works per SKU, which is the level at which fit actually varies."}},{"@type":"Question","name":"Which agents does the skill run in?","acceptedAnswer":{"@type":"Answer","text":"Any agent that supports the open SKILL.md format: Claude Code, Cursor, Codex CLI, GitHub Copilot, Gemini CLI, Manus, Grok and others load it unmodified. The format is portable, the location is not. Each agent scans its own skills directory, so you drop the same folder into whichever one yours uses. There is no config file to edit and no API key to provision."}},{"@type":"Question","name":"What does it need connected to work?","acceptedAnswer":{"@type":"Answer","text":"Returns with SKU, size and reason, orders linked to customers so reorders can be matched, and shipping cost. Free-text return reasons work and structured ones work better."}},{"@type":"Question","name":"How is 31% returns worth $41,000?","acceptedAnswer":{"@type":"Answer","text":"Both shipping legs plus handling on every returned unit, and the margin on the customers who return and do not reorder. The skill shows the components so the figure can be checked rather than taken on trust."}},{"@type":"Question","name":"Will guidance actually change what people order?","acceptedAnswer":{"@type":"Answer","text":"A specific instruction placed next to the size selector does. A measurement table three clicks away does not, and neither does a general note on a separate sizing page. Placement is part of the recommendation for that reason."}},{"@type":"Question","name":"What is included with the skill?","acceptedAnswer":{"@type":"Answer","text":"The skill itself, the fit pattern reference, the per-SKU guidance format, and the return cost model. It is free and security scanned. We re-publish the ZIP when the model changes, so download it again if your rankings start shifting."}}]}</script>