by SizeGuide (LX AI)
SizeGuide turns product measurements and fit notes into clearer size-chart copy — so shoppers guess less and merchants see fewer wrong-size returns.
Decision-support only — confirm before you publish or act.
I built SizeGuide after noticing that many online stores lose sales because shoppers cannot tell if a garment will fit them. The tool takes raw measurements — chest, waist, inseam, sleeve length — and any fit notes you have from product testing or vendor sheets, then turns them into plain‑language size‑chart text that a shopper can read quickly. The output is not a guarantee; it is a decision‑support aid that you should review before publishing.
In this guide I walk through the exact steps I use each week to keep our size charts accurate and helpful. I focus on concrete actions, not theory, and I show how to avoid common pitfalls that lead to returns.
Start by pulling the spec sheet for each style. If you work with a factory, request a PDF that lists the garment’s flat measurements for each size. If you develop in house, use a flexible tape measure on a sample garment and record the numbers in a spreadsheet. I keep a master sheet with columns for size, chest, waist, hip, inseam, sleeve, and shoulder.
Beyond numbers, note how the garment feels on a model or wear tester. Words like “runs tight across the bust”, “generous in the thigh”, or “short in the torso” are valuable. I add a free‑text column next to each size and fill it after each fitting session. These notes become the basis for the copy that SizeGuide will later generate.
Before moving on, I run a quick sanity check: every size must have a value for each measurement column. Missing data leads to vague copy and increases the chance of a shopper guessing wrong. I highlight any blank cells in red and either measure the missing point or mark the size as “data pending”.
I upload the completed spreadsheet to SizeGuide’s import page. The tool accepts CSV or XLSX files and expects the column headers I described. After the upload, I click “Generate copy”. SizeGuide reads each row, applies a set of phrasing rules, and produces a draft size‑chart block for each size.
The output looks like this example for a medium men’s shirt:
I read each line aloud. If a phrase feels awkward or repeats information, I edit it directly in the SizeGuide editor. The tool lets me override any sentence while preserving the underlying data.
Our brand uses a friendly, straightforward voice. I adjust the generated text to match that tone — replacing “fits true to size” with “fits as expected” and adding a brief line about fabric stretch when relevant. I keep the changes minimal so the core measurement facts stay intact.
Once satisfied, I click “Export”. SizeGuide gives me a plain‑text block and an HTML snippet. I copy the HTML into our product‑page template and the plain text into our internal style guide for future reference.
I create two versions of a product page: one with the old size chart and one with the new SizeGuide copy. Using our analytics platform, I split traffic 50/50 for a week. I monitor the metric “size‑related return rate” (returns where the reason code mentions fit or size).
At the same time, I run a short pop‑up survey after purchase: “Was the size chart helpful in choosing your size?” with options Yes, Somewhat, No. I also add an open‑text field for comments. I gather at least 200 responses per variant to get a stable signal.
If the new chart reduces the size‑related return rate by at least 15 % and the helpfulness score rises above 80 %, I consider the copy successful. If not, I return to the SizeGuide editor, adjust phrasing based on shopper comments, and repeat the test.
I write a short internal note that records the test dates, variants, metrics, and any copy tweaks made. This note becomes part of our size‑chart SOP and helps the team avoid repeating the same mistakes.
Our storefront uses a Liquid‑based theme. I placed the SizeGuide HTML snippet inside the product‑description block, right after the price and before the “Add to cart” button. This placement ensures shoppers see the chart before they make a size selection.
Because the size‑chart text rarely changes, I enable theme caching for that snippet. The cache refreshes automatically when I re‑export from SizeGuide, which I do whenever a style gets a new measurement update.
I set up a simple error tracker that watches for missing HTML tags in the rendered page. If the snippet fails to load, I receive an email and can revert to the previous chart while I fix the export.
Every month I run a quick audit: I pull the latest measurement sheet, run SizeGuide, compare the new output to the live version, and publish any differences. This keeps the chart aligned with any subtle changes in fabric or cut that might occur over a production run.
SizeGuide does not promise that a shopper will never need to return an item because of size. It is a decision‑support tool that makes the size chart clearer, but the final fit still depends on the individual’s body shape and personal preference.
The tool does not replace a professional fit model or a wear‑test panel. It works best when you already have reliable measurement data and honest fit notes from those sources.
SizeGuide is not a magic fix for high return rates caused by quality issues, inaccurate photography, or misleading descriptions. If those problems exist, you must address them separately.
Finally, the output is not a legal guarantee. Always treat the generated copy as a suggestion and verify it before you publish.
I keep a written checklist next to every release that touches this topic. The list is boring on purpose: name the owner, name the evidence artifact, and name the rollback. If a step cannot produce a file, a ticket, or a log line, it is not a step — it is a wish.
| Signal | What to record | Stop-ship? |
|---|---|---|
| Unexpected write or send | Tool name, args, identity | Yes until gated |
| Missing confirmation on destructive action | Call path and bypass | Yes |
| Untrusted text treated as instruction | Source surface and resulting action | Yes for agent reach |
| Scanner clean, judgment open | Deferred items + owner | Only if core path safe |
sizeguide helps with the measurable first pass. Architecture choices, legal sign-off, and customer-facing claims still need a human owner. Treat every checklist row as decision support, not a certificate.
When a prospect asks whether the product "covers everything," I answer with the boundary: automated checks catch known patterns; novel chains and policy judgment still need review. That honesty keeps support load down and keeps the blog aligned with what the product actually does.
If you need a second opinion, pair this write-up with the references below and with your own staging reproduction. Do not skip the reproduction step just because a model wrote a confident summary.
SizeGuide turns product measurements and fit notes into clearer size-chart copy — so shoppers guess less and merchants see fewer wrong-size returns.
Decision-support only — confirm before you publish or act.
Want to try the product? SizeGuide
Generated 2026-09-17 · primary sources only · no fabricated traffic metrics