The Question That Decides Whether You Get the Shelf Space

The producer makes marinades and finishing sauces out of a 22,000 square foot facility west of Hamilton. Fifty-odd staff, a product line built over eleven years, and a founder who can tell you the viscosity target of every SKU from memory. Nine months of relationship-building with a national grocery buyer had finally turned into a real conversation about listings.

Then the supplier approval package came through. Sixty-one questions, most of them routine. Insurance certificates, allergen controls, a copy of the recall procedure, GFSI certification status. Standard stuff for anyone who has sold into retail before.

Question 34 asked for the results of their most recent traceability exercise, including start time, completion time, and the mass balance reconciliation for the lot tested.

They had a recall procedure. It was thorough, written by a consultant, reviewed annually, and completely accurate about what the company intended to do. What they didn't have was a record of having actually done it end to end with the clock running. When they tried to generate one, it took eleven days and two people working evenings to reconstruct a single finished lot back through four ingredient suppliers and forward to eight customers. The founder told me the arithmetic was the worst part. The paperwork existed. The numbers didn't reconcile, because nobody had ever asked them to.

Those eleven days were only the trace itself. Completing the rest of the questionnaire, getting the reconciliation into a format the buyer's QA team would accept, and waiting out their review cycle took another month. Six weeks total from the package arriving to the conversation restarting.

They got the listing. A smaller initial one than what had been discussed in the summer, in fewer stores, with a review at six months. The founder is honest about why. He spent six weeks proving something about his operation instead of selling into it, and by the time he was back in front of the buyer the momentum had gone somewhere else.

I bring this up because most conversations about traceability in this industry start with recalls and food safety, which is correct but incomplete. The commercial consequences arrive far more often than the safety ones.

#One up, one back is a floor, and floors are low

Under the Safe Food for Canadians Regulations, licence holders need to be able to identify the immediate supplier a food came from and the immediate customer it went to, keep those records for two years, and provide them to CFIA within 24 hours of a request. Lot identification has to appear on the label or accompanying documents. That's the requirement, and a well-kept binder system can meet it.

What the regulation doesn't ask is whether you can do it in a single pass, whether the quantities add up, or whether the trace holds when one ingredient lot was split across nine batches over five weeks. CFIA's requirement is about the existence and accessibility of records. A retailer's requirement is about your capability under time pressure.

BRCGS puts a number on it. The standard says traceability should be achievable within four hours, including reconciliation of quantities. That's a specific bar from a specific scheme, and it's worth knowing whether your buyers audit against it, because four hours sounds generous until you try it on a product with fourteen ingredients where two of them are sub-recipes made in-house last month.

"We passed our SFCR review with no findings. Then a customer asked us to trace one case and we needed most of a week. Both of those things were true at the same time and I still don't love saying it out loud."

#Rework is where the lineage gets messy

Here's the part that separates plants with clean traceability from plants with technically-compliant traceability, and it's rarely the thing people worry about.

Rework. Blended batches. The 40 kilos of product that came off the line slightly under spec and went back in at 3% into the next four runs. The tote of base that got topped up rather than run down.

Manual systems handle a clean one-to-one lineage well enough, because each batch has a sheet and each sheet has a list of what went into it. The moment a lot appears in a second lot's history, the record structure has to hold a relationship it was never designed for. What usually happens is that somebody writes a note in a margin. Reworked 40kg from 4438. No quantity breakdown, no indication of which subsequent runs received it, no link back to the ingredient lots inside that original batch.

That note is the weak point in your entire trace. Not the missing record, the informal one.

In-house sub-recipes create the same problem one level deeper. A spice blend or a base sauce made in your own facility last month is an ingredient with its own lot history, its own supplier lots underneath it, and its own rework going in and out of it. When that base gets used across three finished products over five weeks, tracing a supplier concern back to a specific delivery means unwinding the base's lineage first, then following it forward into every finished lot it touched.

The math gets worse quickly. One ingredient lot split across nine production batches, three of which contributed rework into four more, means a single supplier notification touches thirteen finished lots and whatever customers received them. On paper, that's a wall covered in batch sheets and a person with a highlighter trying to hold the whole structure in their head.

When a trace has to run in both directions through rework, you need the system to know that finished lot 4471 contains material from lots 4460 and 4452, that 4452 itself contains rework from 4438, and that supplier lot BX-9912 therefore touched product shipped to a customer six weeks after it was received. No auditor will find that chain in a binder. Neither will you, at least not in an afternoon.

#Mass balance is the part nobody prepares for

Ask an experienced auditor what they actually check, and reconciliation comes up fast. You received 1,800 kilos of a supplier lot. Your batch records show 1,640 used across production. You have 90 in inventory. Where's the rest?

Sometimes the answer is legitimate, because there's line loss, sampling, giveaway on fills, waste. Sometimes the answer is that the receiving weight was recorded off the packing slip instead of the scale, and the two were never the same number. Either way, you need to be able to say it with numbers rather than with a shrug and an explanation about how things work on the floor.

Plants running real-time capture sit around 98% inventory precision because the weight comes off the scale into the record with nobody retyping it. That precision isn't interesting in itself. It's interesting because it's what makes reconciliation a report instead of an investigation.

"The auditor asked me to account for 160 kilos and I knew where it went, I just couldn't show him. Knowing isn't the same as showing. That was an expensive lesson in about four different ways."

#If you sell into the US, the ground has shifted

FSMA 204, the FDA's Food Traceability Rule, does something the Canadian requirements don't. It names specific Key Data Elements that have to be captured at specific Critical Tracking Events, and it requires that you produce an electronic sortable spreadsheet within 24 hours when FDA asks. Not a binder, not a PDF of scanned sheets. Sortable.

The compliance date has moved out to 2028, and the industry pushed hard for that extension, so it's worth checking the current status before you plan around any particular month. What hasn't changed is the direction. If your product appears on the Food Traceability List, or if you supply someone whose product does, the expectation is now machine-readable data at defined points in your process. Paper doesn't convert to that on demand.

I've watched a few operations treat the delay as a reprieve. The ones who used it as runway are in much better shape.

#Why this ends up on the sales side of the ledger

Retail buyers, foodservice distributors, and export brokers are all doing the same calculation about you, and it isn't really about your food safety record. It's about how much of their time you're going to cost them if something goes wrong.

A supplier who can answer a trace question in an hour is a low-maintenance supplier. One who needs a week is a liability with good margins, and buyers price that in, or they pick your competitor who answered the questionnaire without hesitating. Nobody says this in the meeting. It shows up in how fast the deal moves.

"Our biggest customer's QA lead told me straight out that we were easier to work with than a company twice our size. Nothing to do with our product. It was because we answered his questions the same day."

#So what happens when your questionnaire arrives?

Not if. Every growing food business in this country eventually gets one, and the pattern is consistent: you get asked to prove a capability at the exact moment when proving it is most inconvenient and most expensive.

The sauce producer is doing fine now. Bigger listing than the one they started with, and the six month review went well. But they'd tell you themselves that they spent nine months building a commercial relationship and then very nearly cost themselves the deal over a records problem they'd been aware of for years and had never had a reason to fix.

The reason usually arrives with a deadline attached.

Book a demo at book a demo. We'll walk you through a full forward and backward trace on a live system, including a batch with rework in its lineage, so you can see exactly what the reconciliation looks like and compare it against what your current process would take.