Monday's offer lists
Three importers send updated offer lists this week. One is a PDF with lots grouped by country. One is a spreadsheet with prices per kilo in pounds. One links to a web portal with prices per pound in dollars and a note about which lots are spot and which are forward. You are looking for a washed Central American to replace a component of your house blend, at a price that keeps your wholesale espresso where it is.
So you copy the likely candidates into your own sheet, convert units, and request samples of four. Next month the lists change and the sheet is out of date.
Every importer formats differently
The information on offer lists is broadly the same: origin, region, producer or cooperative, variety, process, grade, cupping score, bag size, warehouse, availability and price. But each importer structures it differently, uses different units and currencies, and changes the list every week or two.
- PDFs, spreadsheets and portals all in use at once.
- Prices in different currencies and per kilo or per pound.
- Spot, afloat and forward positions marked differently.
- Cupping scores and descriptors written by each importer's own cuppers.
- Nothing records which lots you already sampled and what you thought.
What manual comparison costs
Comparing offers by hand takes the buyer a lot of time each cycle and is easy to do badly. Good lots are missed because they were on a list nobody got round to reading. The same lot is sampled twice because nobody remembered it from last year. Price comparisons are unreliable when units and terms differ. And buying history, meaning what you bought, from whom and at what price, is scattered across old emails.
The sample room suffers too. Sample requests go out from the buyer's email, samples arrive in small bags labelled with the importer's reference, and by the time they are roasted and cupped the link between the cup and the offer line is a handwritten note. When a coffee cups well, finding the offer again to check the price and position takes longer than it should.
The offer tracker we build
- Offer lists are forwarded to a dedicated address or uploaded. A parser reads each importer's format, with an extraction step using a language model for PDFs, and a person checks anything uncertain.
- Every lot is stored with standard fields: origin, region, producer, variety, process, score, bag size, position, warehouse and price converted to your currency and unit, with the original kept alongside.
- The same lot appearing on successive lists is recognised, so you see its price and availability over time.
- You filter across all importers at once, for example washed, Central American, spot in the UK, within a price range.
- Sample requests are recorded against the lot, and cupping results are linked back when you score them.
- Lots you buy become contracts in your green coffee records, with the offer details attached.
| Question | How you answer now | With the tracker |
|---|---|---|
| Which lots fit this brief? | Read every list | One filter across all importers |
| Is this a good price? | Convert units by hand | Prices in one unit and currency |
| Did we sample this before? | Memory | Sample and cupping history on the lot |
| What did we pay last year? | Search old emails | Buying history by origin and importer |
Buying with the whole market in view
The buyer spends time cupping and talking to importers rather than reformatting spreadsheets. Candidate lots are found quickly and compared fairly. The sample room works from a list linked to real offers, and next year's buying starts with a record of what worked. The buying decision itself stays entirely with you.
Is green buying a copy and paste job?
- You copy lots from importer lists into your own sheet.
- Prices are converted between currencies and units by hand.
- You have missed a lot because it was on a list nobody read.
- You cannot easily find what you paid for a similar coffee last year.
- Sample requests are not linked to cupping results.