Tutorial
Read a price change without being misled by it
Turn a movement on the dashboard into a decision, and recognise the comparisons this data cannot support.
About 8 minutes. Last reviewed against the product on .
A price you pay moved. This tutorial is about deciding whether it matters, and it spends more time on the comparisons the page refuses to make than on the ones it shows you.
1. These are your prices, not the market's
Price changes are built from your own imported purchase records. Nothing on this page comes from the public market data elsewhere on this site. It compares what you paid to what you paid before.
2. Read the financial impact before the percentage
The page leads with impact rather than percent on purpose. A 40% rise on something you buy twice a year is not the problem that a 4% rise on your highest-volume line is.
| Column | What it measures | Why it can be blank |
|---|---|---|
| Current | The most recent unit price paid | Never, if the item has any record |
| Change | Current against the previous unit price | No baseline yet: only one purchase exists |
| Weekly impact | The change applied to your recent weekly quantity | Usage cannot be established over the window |
| Weekly spend | What this line costs you in a normal week | Never, it is summed from records |
Measured from the product model in src/lib/purchasing/intelligence.ts.
3. Treat a blank as information, not a gap
A line with no baseline has been bought once. A line with no weekly impact has no reliable usage behind it. In both cases the page shows nothing rather than an estimate.
4. Check the direction against the count
No baseline is not the same as unchanged. Unchanged means two prices were compared and matched;
no baseline means there was only ever one price. Sort by purchase count when a movement surprises
you: a change computed from two records is a weaker signal than one computed from twenty.
5. Leave with one supplier conversation
The action a price rise leads to is almost always a conversation with the supplier who raised it, before the next order. Take the largest weekly impact you understand the cause of, and leave the rest.
If you want to know whether the wider market moved too, that is a different page with different data: what the market price data is, and is not.