Resources

The Hidden Cost of Writing Descriptions Manually

Written by Martin Lotz III | Aug 16, 2026, 8:54:30 PM

Ask a merchant how long it takes to write a product description and most will answer with the time it takes to type one: maybe ten or fifteen minutes for a decent paragraph. That answer is almost always wrong, not because the typing estimate is off, but because typing is the smallest part of the job. I learned this the expensive way while trying to fix product content across roughly 60,000 listings at Lotz Outdoors. The actual cost of a manually produced description includes research, prompting, formatting, copying, importing, checking, and repairing, and every one of those steps takes real time that rarely gets counted.

The steps nobody counts

Here is the full sequence my old process actually required, per batch of products, not per description:

  1. Export product data from Shopify into a large CSV file.
  2. Open and manage the spreadsheet, tracking which products have been handled and which have not.
  3. Research the product, brand, and specifications separately, often across several browser tabs, because the existing supplier text rarely had enough to work from.
  4. Copy the product description and supporting data into a general AI tool.
  5. Type the same instructions again, for every product or batch, because the tool had no memory of the last one thousand times you had already explained your tone and structure preferences.
  6. Copy the generated content back into the spreadsheet.
  7. Prepare and format the file for re-import, checking that fields lined up correctly.
  8. Import the file into Shopify and wait to see what happened.
  9. Check whether the import actually worked, which was never guaranteed.
  10. Repair failed imports, overwritten fields, or mismatched listings, which happened often enough to be a real, recurring cost rather than a rare exception.

Ten steps, and only one of them is writing. The other nine are the hidden cost that never shows up when someone estimates that a description takes fifteen minutes.

Why research is usually the slowest part

Writing is fast once you know what to say. Finding out what to say is the slow part. A thin supplier description gives you almost nothing to work with, so before you can write anything useful, you have to track down materials, dimensions, intended use, and whatever else a shopper would actually want to know. That research happens outside the writing tool entirely, in browser tabs, spec sheets, and sometimes emails to a supplier who may or may not answer quickly.

This is the part of the process most merchants underestimate the most, because it does not feel like writing and so it does not get counted in a time estimate. It is, in practice, often the largest single cost.

Repeated prompting is a hidden tax

A general AI tool has no memory of your brand voice, your structure preferences, or your banned claims from the last session. Every batch means retyping the same instructions: keep the tone direct, always include a materials line, never claim waterproof unless verified, format as HTML with these specific tags. Multiply that retyping across hundreds or thousands of products and the tax adds up to real hours that produce no new information, only repeated instruction-giving.

The cost of fixing what breaks

The most expensive step in the old workflow was never the writing. It was the repair work after a failed import: a field that got overwritten, a description that ended up on the wrong product, a batch that partially failed and required checking every single row to find the one that did not take. This kind of repair work is unpredictable, which makes it hard to plan around, and it often happens hours or days after the original work, when the context has already faded.

A rough way to see your own hidden cost

You do not need precise numbers to see the shape of this cost. Try tracking a single product through the full manual process, honestly, including every step above:

  • Time spent researching the product and its specifications.
  • Time spent writing or typing prompts and instructions.
  • Time spent formatting and copying content between systems.
  • Time spent importing and then checking whether it worked.
  • Time spent fixing anything that broke, averaged across how often that happens.

Add those five numbers together, then multiply by the number of products you actually need to touch. Most merchants find the real total is several times larger than their first instinct, because the first instinct only counted the writing.

What actually disappears with a different workflow

This is the specific cost the Enrichio workflow, build a template, load products, generate, then review and publish, is designed to reduce. A template removes repeated prompting, since your tone, structure, and rules are set once and applied automatically. Loading products directly from Shopify removes the export-and-spreadsheet step entirely. Research happens as part of generation, rather than in separate browser tabs. Review and publish inside the same system reduces the risk of the failed imports and overwritten fields that caused the most expensive repair work in the old process.

None of this eliminates the need for human judgment. You still review drafts, verify facts, and approve what gets published. What changes is how much of the surrounding labor, the exporting, the repeated instructions, the formatting, the import checking, disappears from the process entirely.

Putting a number on your own catalog

If you want a starting estimate for your own store, try this simple exercise:

  1. Pick ten products that need real content work.
  2. Time yourself honestly through research, writing, formatting, and import for those ten.
  3. Divide by ten to get a per-product figure that includes the hidden steps, not just typing time.
  4. Multiply that per-product figure by the number of weak listings in your catalog.

That number is usually the moment a merchant realizes the real constraint was never writing speed. It was everything happening around the writing. Enrichio supports research and generation across that whole workflow, but the time savings will vary by catalog, category, and how thin your existing content is, so treat any specific hours-saved figure as an estimate you should verify against your own experience, not a fixed guarantee.