A production plan you will actually keep up to date

Most plans are built for a week that never arrives

The production plan that dies is the one built for a perfect week. Every account orders what they said they would, nothing changes on Tuesday, and the sheet you made on Sunday is still true on Friday. That week does not come.

I planned prep for service for about a decade, and then I planned production for a microgreens business I ran during covid, growing and selling to trade. I have kept plans that worked and abandoned plenty that did not. The ones I abandoned all failed the same way: they assumed the week would hold still. When I planned production from scattered orders plus memory, I got it wrong in both directions. Some weeks I cut and packed more than anyone wanted and watched it go over. Other weeks I was short and had to tell someone no.

Standing orders were the worst offenders, because they rarely stood. A recurring order changed often enough that "recurring" was a fiction I kept up by hand. So a plan built on the assumption that last week repeats was wrong before I had finished writing it.

A plan survives when it expects to be edited. That is the whole idea below.

Three things a plan has to hold

Strip a production plan back and it only needs to do three jobs.

Hold the fixed commitments, so the orders someone has actually placed never live only in your head or in a message thread.

Hold the guess, separately, so you can plan for the accounts who usually order but have not confirmed yet without pretending that guess is a promise.

And be small enough that updating it on a Wednesday night takes a minute, not a sitting. A plan with forty cells does not get updated. A plan with a dozen does.

The template

Here is the sheet. It is one row per thing you make, and six columns. Paper, whiteboard or spreadsheet, it does not matter which.

Line item	Confirmed	Likely	Make	Done	By
e.g. pea shoots	4	2	6	☐	Thu am
e.g. sourdough	20	8	28	☐	Fri 6am

Line item. One row for each thing you produce and sell as its own unit. Not "the café order", which hides three different products inside it. The row is the trap, split it into the actual things you make.

Confirmed. The quantity someone has actually asked for and you have said yes to. This column is a record, not a hope. If it is in here, you owe it to someone.

Likely. Your guess for the accounts who usually order this and have not been in touch yet. Keep it in its own column so you never confuse it with the confirmed number. When the order lands, it moves left into Confirmed. When Friday comes and it never landed, you have lost nothing but a guess.

Make. What you are actually going to produce. Often it is Confirmed plus Likely, but not always, and this is the column where your judgement goes. A grower plants more than the order because some of it will not make grade. A baker rounds to a full tray. Write the real number you will make, and you will learn, over a few weeks, how good your guesses are by comparing this column to what went out.

Done. A tick. That is the entire job of this column, and it is the one that stops you standing at the bench on Friday morning wondering whether you did the thing.

By. The make-by or deliver-by point. More on this below, because it is the column people leave off.

Sunday, and the one habit that keeps it alive

Fill it once at the start of the week. Put in every confirmed order you have, put your honest guess in Likely, and set your Make numbers. Fifteen minutes, most weeks.

Then the habit that actually matters. When an order changes, and it will, you change one cell. A café doubles their Friday order on Tuesday afternoon, you add it to Confirmed and adjust Make. That is the whole update. You are not rebuilding the plan, you are nudging it.

This is why the sheet has to be small. If editing it is a job, you will not do it midweek, and a plan you stopped editing on Monday is lying to you by Thursday. The test I use: if updating the plan is more effort than the message that prompted the update, the plan is too big.

The column everyone leaves off

The "By" column looks optional. It is the one I would fight to keep.

Most small producers do not miss orders because they forgot to make something. They miss because they made it at the wrong time. Something with a make-by point, bread, cut greens, anything that is only good for a day or two, has to be produced against the delivery, not against the week. Two accounts ordering the same product for different days are two different production jobs, even though they are one line in your order book.

Put the day and the hour it has to be ready in that column. Sort your Sunday plan by it if you can. Then Thursday's work and Friday's work stop being one undifferentiated pile you dig through at 5am.

Where the Confirmed column should come from

The weak point in all of this is the Confirmed column, because filling it means gathering every order that arrived as a message, a call, an email, a note to yourself, and copying them across by hand. That copying is where mine used to go wrong. An order sat in a thread, never made it to the sheet, and was still sitting there when I was packing everything else.

This is the join makeweek is built around. Orders come in as line items, and the line items across your confirmed orders add up on their own into what you need to make. The Confirmed column stops being a thing you transcribe on a Sunday and becomes something that is already counted. You still make the calls on Likely and Make, because that is your craft and no software knows your yield. But you are making them on top of a real total instead of a remembered one.

Start with the paper version this week. Six columns, one row per thing you make, a tick when it is done. Keep it small enough that you will still be editing it on Wednesday night, and it will already beat the plan you abandoned.