Lean in cannabis · packaging and fulfillment · vault to delivery
Pack what's ordered.
Packaging and fulfillment is the last stop before a store, and the last chance to catch an error. Every label, weight and manifest has to be right. A mistake here comes back as a relabel, a manifest that got kicked back, or a store that's out. Co-packing adds a second licensee: the client's tested product reaches you by a transfer entered in the state tracking system, and it gets packed to the client's spec. Confirm who owns the product and how its tags are handled with your CCC compliance officer. These are the 8 wastes from the vault to delivery, with how to measure each one and the first thing I'd try. Written for packaging leads, fulfillment and inventory staff, and whoever runs the line. I run continuous improvement at a Massachusetts cannabis company, and in that work packaging went from a cost center to a line that co-packs for other brands.
DOWNTIME · vault to delivery
The 8 wastes of packaging and fulfillment
For each one: what it looks like, how to measure it, where to look, and one thing to try first. Measure before you change anything, so you can show the difference afterward. Required compliance checks are not on this list. The rework when something fails one is.
Defects
What it looks like
- The wrong batch or potency on a label.
- Underweight jars, caught at the check-weigher or not caught at all.
- "Manifest got kicked back." Usually a tag that doesn't match.
Measure it
Label error rate, fill-weight accuracy, and manifests rejected or corrected.
Where to look
The check-weigher, the relabel bin, rejected manifests in the tracking system.
First thing to try: mistake-proof the label. Before a run, check the first labeled unit against the batch's record and its package tag, then run the rest. Then list every place a label field is typed by hand, and replace one with a scan or a pick from a list. My free mistake-proofing skill walks through the options.
Overproduction
What it looks like
- Running every weight of a strain when only one of them sells.
- Jarring past the next replenishment, so packed units sit in the vault.
- Packing to keep the line busy, not to fill an order.
Measure it
Packed units with no order or replenishment behind them, by SKU and weight.
Where to look
The vault shelves, next to the replenishment list and open wholesale orders.
First thing to try: pack to replenishment. For each SKU and weight that sells, set a small finished-goods level, and pack only to refill it when an order pulls it down. Slow weights and co-pack runs get packed against an order.
Waiting
What it looks like
- The line idle, waiting on labels or lab data.
- A driver waiting on a late pick.
- The receiving store waiting on the manifest.
Measure it
Order cycle time, from order to delivered, split by step. And minutes the line sits idle, by reason.
Where to look
The line whenever it's stopped, staging, the loading door.
First thing to try: a line-stop log. Every time the line goes idle, write the time and the reason on a sheet at the line: labels, lab data, materials, changeover. After two weeks, sort it. The top reason is your first kaizen. If it's changeovers, my free SMED skill walks through cutting them.
Non-utilized talent
What it looks like
- Packaging leads who know what moves, and nobody asks them before the run is planned.
- Hand-keying manifests while a picker knows the faster way.
- Packers who know why the labels jam, never asked.
Measure it
Ideas raised by packers, pickers and leads, and how many were acted on, each month.
Where to look
The packing line and the fulfillment desk, at the end of a shift.
First thing to try: end each shift with one question on the board: what made you stop or redo something today? Act on one answer each week, and say which one.
Transportation
What it looks like
- Product going vault to pack, back to the vault, to staging, then to the truck.
- Extra trips to cover short picks.
- Staging at the far end of the building from the loading door.
Measure it
Moves per order from packing to the truck, and extra trips per month to cover short picks.
Where to look
The vault log, the delivery schedule, and a floor plan.
First thing to try: draw one batch's path on a floor plan, from the vault through packing and back, then one order's path to the truck: a spaghetti diagram. Circle every move that doesn't change the product. Keep every move the security rules require, and question the rest.
Inventory
What it looks like
- Finished goods sitting in the vault while stores are out.
- Pre-printed labels for dead SKUs.
- Co-pack materials for a client brand with no run on the schedule.
Measure it
Finished-goods days on hand, and store outs that were sitting in the vault at the time.
Where to look
The vault, the list of store outs, the label shelf.
First thing to try: a daily match. Each morning, put the list of store outs next to what's in the finished-goods vault. Anything on both lists ships first. Then walk the label shelf with the current SKU list and red-tag anything that isn't on it.
Motion
What it looks like
- Walking to the printer for every run.
- Hunting for the right tag roll.
- Reaching across the table for jars, lids and bags.
Measure it
Steps to finish one run (count them once), and units packed per labor-hour by station.
Where to look
The packing stations, the path to the printer, the pick path through the vault.
First thing to try: point of use. Put a label printer at the line and the tag rolls in marked spots beside it, so printing on demand is a step, not a walk. Then write the station layout down as standard work, so every station looks the same.
Extra processing
What it looks like
- Relabeling finished goods after a rule or design change.
- Scanning every package tag twice when one scan would do.
- Repacking to a buyer's preferred size after the product is already jarred.
Measure it
Units relabeled or repacked each week, by cause, and tag scans per package.
Where to look
The relabel bench, and every point where a tag gets scanned.
First thing to try: one scan per package. Walk one package and write down every time its tag gets scanned, and what each scan is for. The scan that records it in seed-to-sale tracking stays. The goal is one good scan that does the jobs the extra scans were doing, never a skipped one.
Before you cut anything
What isn't waste here
- 1Printing labels on demand. The compliance label carries the batch's own test data, so it gets printed when that batch is packed, not in a long run ahead of time. A long run ahead is how one batch's numbers end up on another batch's jar. Make printing on demand fast and close to the line.
- 2The tag scan. The state package tag is the product's unique ID in seed-to-sale tracking. Scanning it is not the waste. Scanning it twice when once would do is. The fix is one good scan, never no scan.
- 3The checks before product leaves. Certified child-resistant packaging, the manifest, and the receiving licensee's re-weigh and acceptance all stay. Don't cut them. Make them pass the first time.
If you only measure two things
Where I'd start
- 1Label error rate. Packages with any label error, out of all packages labeled. Count the errors you catch and the ones a receiving store catches. Start with a tally at the final check: packages checked, errors found, and what was wrong. It's the quality measure of the line, and the one I'd put on the board first.
- 2Order fill rate. Lines shipped complete, divided by lines ordered. A line is one product on an order, and it counts only if the full quantity shipped. Start from last month's orders and manifests, then track it every week. When it drops, check the vault first: was the product there and not picked, or never packed?
Free · one page · print it
The packaging walk sheet
One printed page: the eight wastes as they show up between the vault and the store, where to look, the question to ask, and room to write what you saw. Follow one order, write it down, and bring the sheet back to your team before anyone proposes a fix.
The rest of the operation