Food traceability software for a Canadian food maker has to do what Part 5 of the Safe Food for Canadians Regulations asks. It records every lot one step back and one step forward, with names, addresses and dates. It puts a lot code on the label, and it hands the CFIA one plain-text file within 24 hours. A good one also keeps the link between supplier lots and your batches without anyone typing it. Here are ten tests you can run in a free trial, each tied to the rule it serves.
The short version
- Run your last mock recall in the trial, with your own data, and time it.
- The trace has to start from any lot: yours, or a supplier's.
- The batch's ingredient lots have to be recorded as you work, not typed in afterwards.
- The answer to the CFIA has to download as one CSV file, with the addresses in it.
- Ask what it costs for the number of people who actually record batches.
- TaroStack passes the ten tests on every plan, from $9.99 a month. Run them against it too.
The ten tests
| # | Test | Why (SFCR) | How to run it in a trial |
|---|---|---|---|
| 1 | Records the supplier's lot at receiving, ideally by scanning the case label | 90(1)(c) and (d), and narrow recalls | Receive one real delivery |
| 2 | Records which ingredient lots went into a batch, without anyone typing them | The link one step back to one step forward | Make one batch from two deliveries of the same flour |
| 3 | Gives each batch a lot code and prints it on a label | 90(1)(a) and 92 | Print one label and read it |
| 4 | Records which lots went to which customer, with their address | 90(1)(b) | Ship one order of two lots |
| 5 | Traces from any lot, yours or a supplier's, both ways | The whole of Part 5 | Start from the flour lot in test 2 |
| 6 | Shows what's still on your shelves from a lot, and where | Holding product, and the CFIA's distribution list | Compare it with a shelf count |
| 7 | Downloads the trace as one CSV file with names, addresses and dates | 91(1)(b): "a single file and in plain text" | Download it and open it in a spreadsheet |
| 8 | Times a recall simulation and keeps the record | 84(2): every 12 months, kept two years | Run your last mock recall again |
| 9 | Keeps the two recall contacts and complaints on record | 84(1) and 83 | Look for where they go |
| 10 | Keeps everything for two years, reachable from Canada | 90(3) | Ask what happens to records when someone deletes an item |
In my view, tests 2 and 7 deserve the most time. A batch's ingredient lots recorded as a side effect of finishing it, and the answer to section 91 as one file rather than a report per lot or a PDF, are what a recall depends on.
This is general information, not legal advice. Your CFIA inspector has the last word on what applies to you.
The kinds of software, against the tests
| Kind | Check hardest | Suits |
|---|---|---|
| A spreadsheet | 2, 5 and 6, which depend on typing; 8 and 9, which depend on memory | A few products, one person |
| General inventory software | 2 and 3: whether there's a production side and lot labels at all | Resellers more than makers |
| Manufacturing planning (MRP) | 7 and 8, and the price per user | Several products, planning-heavy |
| Food ERP | The price and the setup project | Several plants, big-retail programs |
| Food safety platforms | 2, 5 and 6, if the platform doesn't hold your stock | Writing the plan and its logs |
| Maker inventory (TaroStack's kind) | Whether it fits a very large operation | Small and mid-sized food makers |
These are kinds, not verdicts on any one product. Run the ten tests on the one you're looking at. Our food manufacturing software comparison has named tools and their prices at list.
How to run the trial in an afternoon
- Pick a real lot from last month: one supplier lot that went into two or three of your batches.
- Enter only what's needed: the supplier, the ingredient, the two or three products, the customers who got them.
- Receive the deliveries, finish the batches, and ship the orders, the way your people would.
- Start a trace from the supplier's lot and time it.
- Download the result and open it. Check that it's one file, that the addresses are there, and that nothing is missing.
- Count how many people would need a login to do this every day, and price that plan.
If a system can't do step 4 without someone telling it which lots went into which batch, it will be no better than your spreadsheet on the day it matters.
Where this stops working
Every system on that list depends on the same thing: the lot being recorded when the food moves. Software that makes recording a batch slower than writing on a clipboard will lose to the clipboard, and its traces will have gaps. The test of traceability software isn't the trace screen. It's whether the person at the mixer at 5 a.m. records the batch in it.
The other trap is price by the seat. A plan that's cheap for one login and expensive for five pushes a business into one shared login, which removes the record of who did what.
How TaroStack does it
TaroStack is maker inventory built around lots. These are the ten tests, in its terms.
Receiving records the supplier's lot, typed or scanned from a GS1-128 case label (1). Finishing a batch takes ingredients from the lots that expire first and records which ones it used, so the batch's lots are written as you work (2). Each batch gets a lot code and a printable lot label with a GS1-128 barcode (3). Shipping records which lots went to which customer (4). Any lot opens a trace, both ways, with what's still on your shelves and where (5 and 6).
The trace report carries your business's, your customers' and your suppliers' names and addresses, downloads as one CSV file for one lot or several (7), and lot labels carry your business name and address.
Traceability drills are timed and saved with what they found (8).
The two recall contacts live in settings and print on every drill and recall record, TaroStack reminds you before 12 months pass without a drill, and complaints are recorded against their lot (9). Nothing is deleted, so two years of records are there when you need them (10).
Lots, recalls and drills are on every plan. Starter is $9.99 a month for one person; Growth, $39, is for three people, and Business, $89, for up to ten. It's free for 14 days, so run the ten tests with your own lot. If you'd like a hand setting up the trial, ask, and we'll do it with you.
