What Goodspring stands for
Measurement you can check, from someone with nothing to sell you after.
Goodspring is an independent measurement practice for Shopify stores, and the house of one product. The practice finds where the numbers are wrong and where the money leaks, in dollars, and hands you the map; it does not sell the fix. The product, Keel, fixes one thing, product data, at a flat price you can read before you buy.
Vendor-neutral: no tools sold, no referral fees, no implementation arm waiting. One price, agreed before the work starts and stated in public. Nothing renews by default. A human in the loop, and a crawler that identifies itself and publishes its own rules.
The creed, on shirts, in our own store: a real one, with real orders, and where Keel is tested first.
The practice
Every leak I can find in your store, ranked: evidence, dollar cost, root cause, what to do about each. The number is set to your store on a 15-minute call, agreed in writing before I start, and it doesn't move after. Same scope at every size; several storefronts or markets are scoped on the same call. At the readout, one thing is offered and nowhere else: a one-time re-measurement, about sixty days after you ship the fixes, that shows whether they worked. The diagnostic, in full.
Product data that agrees with itself.
The admin, the product page, its structured data and the Merchant Center feed each tell a different story about the same product. Keel reads all four, shows you the difference, fixes the data underneath in bulk, and proves it with a re-scan.
What it represents
Measurement is free.
The audit reads your whole catalog and shows every finding with its evidence before anyone asks you for money. You have a real number first.
The fix is a product, not a proposal.
A flat price, printed where you buy it. Never per order, never a share of revenue, no retainer, no implementation team at the end of a report.
Your catalog is never a guess.
Keel never invents an identifier and never writes without a preview, a snapshot and an undo. What it cannot read it reports as unknown, not as a pass.
Feed apps submit what you have. SEO suites grade it. Keel is built for the part in between, which the App Store has barely had: read every surface a shopper never sees, name what disagrees and the Shopify field that fixes it, fix it in bulk without a developer, and keep it correct as new products arrive and as Google and Shopify change the rules underneath you.
Tell me what is wrong in my terms. Fix it without a developer. Prove it worked. Warn me when it breaks again.
The four jobs Keel is built to do, in the merchant's own words.
The diff between surfaces
One product, four surfaces. Search engines, ad feeds and shopping agents read the ones you never look at.
Illustrative, a made-up product. This is the shape of what Keel shows: which surface disagrees, what it says, and the field that fixes it.
What it checks
What the channels match and classify by.
A GTIN missing where the category expects one, or failing its check digit; a brand missing; a part number missing when there is no GTIN; no taxonomy category, or its attributes left empty.
What machines are told.
Product markup missing, duplicated by the theme and an app, or disagreeing with the admin or the visible page; Merchant Center items disapproved, each reason mapped to the Shopify field that fixes it.
Why the subscription exists.
A new product created incomplete; an edit that breaks a check it used to pass; Google or Shopify changing a rule, with the count of your products it affects and the source linked.
How it works
- Install. The catalog read starts at once, with no setup screen. The first finding, from your own catalog, is built to show within a minute.
- The three biggest issues. Each in plain language, with a count over the products it applies to, an example, and its evidence. What could not be read is reported as unknown, beside the count and never inside it.
- One free sample fix, then before and after. Up to ten products, with a diff preview before anything is written and identifiers from you or your supplier's sheet, validated, never guessed. Keel re-scans what it changed, the page as well as the admin, and shows what resolved. Then, and only then, the plans.
- Every write has a preview, a snapshot and an undo. Bulk jobs included.
- Counts with denominators, never a score. "20 of 190 products lack a GTIN" is a finding; a number out of a hundred is not. Every finding carries its evidence.
- Neutral where it cannot help. Keel names the tool that can, with no affiliate link. On uninstall the data Keel wrote stays, because it is yours.
Plans and the free grader
Free shows the full extent. Paying zooms in and fixes it.
The audit is free: every finding, with its evidence, for the whole catalog, plus a small allowance of fixes. Fixing is paid, two ways: a Cleanup pass, one supervised bulk fix charged once; or a monthly plan (Keep, Fix, Monitor) for the change stream. Prices are flat and printed in the App Store listing when it is live. Cancelling is one click down to the smallest plan.
A store URL in. A linkable report out. No install, no email.
The grader reads a sample of a store's product pages from the outside, the way a search engine or a shopping agent would, and prints the sample size and what it could not read. It cannot see admin fields or Merchant Center status, and says so; that gap is the reason to install. It reads through the same crawler as the practice; a store that blocks it gets an honest "could not read", never a workaround.
Where it stands
September 2026: the engine is built and tested. The app and the grader are next.
The check engine, the rules corpus and a two-hundred-product test store with known defects pass their tests, deterministically. The grader ships first; the listing follows app review. No date is printed because none is certain; sign up below and you hear the day it is live.
Keel fixes the data, and most current themes generate their Product markup from it. Where a theme does not, Keel says so and offers its own markup, off by default. The first installs are worked by hand, and support answers within one business day.
Hear when Keel is live.
One email when the grader and the listing are live, and nothing else. Your address sits in my inbox, not in a list vendor; reply "stop" and it is gone. Say what you would want fixed first and you can be among the first installs.
Here for the diagnostic? Worth 15 minutes.
Screen shared, no deck. I walk what I found on your store, you tell me which parts you already knew, and we work out what's worth fixing first. If we're not a fit, you keep the list.