Service level: deciding how often you're allowed to run out
Service level: deciding how often you're allowed to run out
Never running out is not a goal, it is a budget decision — and an expensive one. Service level is the honest version: choosing, product by product, how much stockout risk you will accept, and paying for exactly that much protection and no more.

Jaswant Singh
Co-Founder, CTO & COO, Kauzio
Ask an owner how often they are willing to run out of a product and the instinctive answer is "never". But never has a price, and it is steep: the last few percentage points of availability cost more in safety stock than all the rest combined. Somewhere between "never out" and "always out" is a number you are choosing whether you admit it or not. Service level is just choosing it on purpose.
One number per product, not one number per shop
A 95% service level means you accept running out on roughly one replenishment cycle in twenty. Whether that is reckless or extravagant depends entirely on the product:
Your A items — the products your regulars come for — deserve 98% or better. A gap on that shelf costs more than the sale: it teaches your best customers to check somewhere else first.
Mid-range products can live at 90–95%. An occasional gap is an apology, not a crisis.
The long tail can run far lower. Holding deep safety stock on products that sell twice a month is how stockrooms fill with frozen cash — the disease we described in The cash frozen on your shelves.
The two numbers that set the buffer
For any product, the safety stock that delivers your chosen service level comes from two things: how much its demand bounces around, and how variable your supplier's lead time is. Steady demand and a reliable supplier need almost no buffer at any service level. Volatile demand and a supplier who ships "in one to three weeks" need a big one — or a cheaper fix: fixing the supplier.
That last point is the most underrated inventory lever there is. Cutting lead-time variability does more for availability, at zero stock cost, than any amount of buffer.
Deciding it beats defaulting it
Most shops run an implicit service level set by gut feel and shelf space — high on whatever the owner personally notices, low on whatever they don't. Making it explicit, even crudely — three tiers, three targets — converts stockouts from recurring surprises into a budgeted, chosen cost. Kauzio watches sell-through and flags the products drifting below their tier while there is still time to reorder; the tiers themselves are a judgement only you can make. Decide the number. Do not let the shelf decide it for you.
Read next
Forecasting covers: the restaurant version
Forecasting covers: the restaurant version
August 7, 2026 · 2 min read
From 'what happened' to 'what to do': prescriptive analytics, honestly
From 'what happened' to 'what to do': prescriptive analytics, honestly
August 7, 2026 · 2 min read
How to choose decision intelligence software
How to choose decision intelligence software
August 7, 2026 · 3 min read
