Skip to content
GearDock

Catalog Operations

From a Thin Medical Spreadsheet to Product Content

Most catalogs start as a spreadsheet with a manufacturer, a model number, and a price. Here is how that becomes a full product page without inventing facts.

The GearDock TeamUpdated 4 min read

Almost every medical equipment catalog begins life the same way: a spreadsheet with a few hundred rows, a manufacturer name, a model number, a price, and a description field that is either empty or copied from a supplier email three years ago.

The gap between that spreadsheet and a product page a buyer will actually trust is large, and it is almost entirely manual work. This article walks through how to close that gap in a way that keeps the resulting content traceable to something real.

Start by Separating Identity from Description

The single most useful thing you can do before writing any content is to decide which columns identify the product and which columns describe it. Identity columns answer 'which exact thing is this?' Description columns answer 'what should a buyer know about it?'

Identity is where accuracy is non-negotiable. If a row is matched to the wrong device, every sentence generated afterwards is wrong in a way that reads perfectly fluent. Description can be improved later; identity has to be right first.

Typical columns in a medical equipment spreadsheet
ColumnRoleWhy it matters
ManufacturerIdentityNarrows the candidate set more than any other single field.
Model / part numberIdentityThe strongest signal available. Worth cleaning before anything else.
Product typeIdentity (weak)Useful for disambiguating between similar model numbers.
ConditionDescriptionCommercial, not factual. Keep it separate from device specifications.
PriceDescriptionNever influences which product a row is.
Existing descriptionDescriptionTreat as a hint, not as a source. Its provenance is usually unknown.

A Cleaned Model Number Is Worth More Than a Longer Description

Stripping trailing whitespace, normalizing separators, and removing internal SKU prefixes from the model column will resolve more rows correctly than any amount of extra prose in the description field.

Attach the Documents You Already Have

Most sellers are sitting on more evidence than they realize: OEM brochures, service manuals, specification sheets, and instructions for use, often in a shared drive nobody has opened in a year. These documents are the difference between content that states a specification and content that can show where the specification came from.

You do not need complete document coverage before starting. A catalog where 40% of products have an attached spec sheet is far more useful than one where none do, and the products with documents can be finished first while the rest are chased.

  • Prioritize documents that cover many products at once, because a family brochure often covers a dozen model variants.
  • Prefer manufacturer-published documents over reseller summaries, which frequently contain transcription errors.
  • Record the document date. A specification from a superseded revision is a real source of published errors.
  • Do not upload service reports, work orders, or anything containing patient information.

Let the Gaps Stay Visible

The instinct with a thin catalog is to fill every field. Resist it. A product page that says nothing about weight is honest; a product page that states a plausible-sounding weight nobody verified is a liability that will eventually reach a customer.

A good workflow distinguishes three states for every attribute: supported by an attached source, present but unverified, and missing. Only the first should be published without a second look. The other two are a work queue, and treating them as a queue rather than as failures is what makes a large catalog tractable.

  1. Import and Validate

    Load the spreadsheet, map the columns, and look at the rows that failed validation before looking at the ones that passed. Bad rows cluster around a single upstream cause more often than not.

  2. Match Before Writing

    Resolve each row to a specific product. Confirm the ambiguous ones by hand. Nothing downstream is trustworthy until this step is settled.

  3. Attach and Extract

    Link the documents you have, and let extraction pull structured facts out of them so they can be cited rather than retyped.

  4. Review the Gap Report

    Work the list of missing and low-confidence attributes as a queue. Decide which gaps block publication and which are acceptable to ship without.

  5. Generate, Then Review

    Produce the draft content, then have a person read it against the evidence. Generation is a drafting step, not an approval.

What 'Publish-Ready' Actually Means

Publish-ready is not the same as written. Content is publish-ready when a person has read it, the factual statements trace back to something, the condition and commercial language is accurate for the specific unit being sold, and someone with the authority to approve it has done so.

That last point is the one most often skipped, and it is the one that matters when a customer disputes a specification eighteen months later. Being able to answer 'who approved this, when, and against what source' is worth the small amount of process it costs.

Where GearDock Fits

GearDock runs this sequence as one workflow: spreadsheet import, product matching, document evidence, gap detection, content generation, human review, approval, and controlled export. The gaps stay visible throughout rather than being quietly filled.

Private Beta

Try This on Your Own Catalog

Reading about a workflow only goes so far. Bring a slice of your products and the documents you hold for them, and we will run it with you.