Par levels for a pub: how much stock to hold, and when to reorder

Every cellar answers one question, well or badly: how much stock is enough? Too much and you've turned cash into boxes that sit there depreciating. Too little and Saturday night finds the one tap that matters running cloudy at 9pm. A par level is just that question answered per product, in writing, from your own numbers instead of your gut. Here's how I set mine.

What a par level actually is

A par level is the amount of a product you restock to at every order. Not a vibe, a number: Madri, par 4 kegs. If the count on order day says 2, you order 2. That's the whole mechanism. The value isn't in the arithmetic, it's that the decision gets made once, calmly, from sales data, instead of fifty-two times a year at 8am with a delivery deadline and a hangover of guesswork.

The working formula is simple: what you'll use before the next delivery lands, plus a buffer for the week that surprises you.

Par = rate of sale over the delivery cycle + buffer

On a weekly delivery, take your busiest recent week for that product and add one unit. A lager doing 2 to 3 kegs in a normal week and 3 on a big one gets a par of 4. A vodka doing a bottle and a half a week gets a par of 3, because spirits don't go off and a £15 bottle is cheap insurance. A slow liqueur doing a bottle a quarter gets a par of 1, and a hard look at whether it earns its shelf at all.

Rate of sale comes from the count, not the till

The till tells you what you sold. The count tells you what you used, which is what you have to replace. On draught those are not the same number: line cleans, fobbing and waste sit in the gap, and on a badly run line the gap is pints per week. Set pars off till data and you'll systematically under-order your fastest lines, because the fast lines are where the waste lives too. If you're not counting regularly yet, here's how often a pub should count, and the short answer for setting pars is weekly for four weeks, then you know your rates of sale better than your brewery rep does.

Cask is a timetable, not a number

For keg, bottles and spirits, par levels are about cash. For cask they're about the clock. The Cask Marque cellar card, the standard the assessors audit against, says check there's at least 14 days of best-before left when the delivery arrives, hold a minimum of three days' stock, and once a cask is on sale, empty it within three days. So a cask par isn't "how many firkins fit in the corner", it's "how many can I empty within their window". If your ale sales can't kill a firkin in three days, your par is pins, not more firkins. I've covered the whole routine in cellar management basics and the container maths in the keg and cask calibration guide.

What getting it wrong costs, both ways

Overstocking feels safe and quietly isn't. A cellar holding £1,000 more than its pars need is £1,000 of your cash converted into stock that pays you nothing, goes out of date at the edges, and makes every count take longer. Run that state permanently and it's an interest-free loan you've made to your own shelves. The worst of it calcifies into dead stock, the cases nobody orders because nobody remembers why they're there.

Understocking costs more per incident and it's louder. Run the best-selling lager dry on a Friday and you don't just lose those pints, you teach your regulars to check the competition. And the emergency top-up run means cash-and-carry prices on stock you'd normally buy better, so the panic pints are your lowest-margin pints of the month. Those two failure modes are the whole case for pars: the buffer is sized so the panic never happens, and the par is sized so the buffer isn't a warehouse.

Pars move with the calendar

A par sheet set in January and never touched again is wrong by April. Sport, bank holidays, Christmas, the local beer festival: rates of sale swing, and the par should swing ahead of them. The other reason to carry a sensible buffer is that deliveries fail on their own schedule. When a national brand went short in summer 2026, Damm UK confirmed nationwide Old Speckled Hen delays, and pubs holding three days' stock rode it out while just-in-time cellars went dark on that line. Delivery delays and your stocktake covers what to do when the dray doesn't come.

A worked par sheet

An illustrative par sheet for a wet-led pub on a weekly delivery. Your rates of sale will differ; the method won't.

Product	Busiest recent week	Buffer	Par	Count says	Order
Lager (50L kegs)	3	1	4	1.5	3
Guinness (50L kegs)	2	1	3	1	2
Cask ale (firkins)	2	0 (clock rules)	2	1	1
House vodka (70cl)	1.5	1.5	3	2	1
House white (75cl)	6	2	8	3	5
Post-mix syrup (boxes)	1	1	2	1	1

Note the cask line: no buffer, because the buffer would go stale. And note that the order column is just par minus count, which is why the order takes ten minutes once the pars exist. This is exactly what StockTap's order guide does with your live count: it knows your rates of sale from your stocktakes and flags the lines whose coverage is running short, so reorder day starts from your numbers instead of a blank page.

Set them once, then let the counts argue

Don't aim for perfect pars on day one. Set them from your last month's usage, run them for four weeks, and let the counts tell you where you were wrong: the line that hit zero gets a bigger buffer, the one that never dropped below half gets its par cut and the cash comes back off the shelf. A par sheet is never finished. It's just less wrong every month, which is more than you can say for gut feel.