We have a kraft gift box that goes into three of our bestselling bundles. The primary supplier lands it at $2.10 a unit on a 12-working-day lead time. There is a second supplier on file who holds the same board locally and turns an order around in four working days. They charge $2.56.
In early July the primary went quiet on a delivery. I had about five days of cover. I looked at the 22% premium on the alternative, decided it was too much, and waited. The box turned up eleven selling days later than promised, which left us short for six of them.
The premium on that order would have been roughly $1,380. Those six days cost more than that, and I had everything I needed to work that out before I decided.
I run Personalised Favours, a Sydney-based manufacturer on Cin7 Core and Shopify. We hold a bit over 15,000 sellable variants and I actively manage around 700 of them, so the supplier question comes up often enough that a rule beats an instinct.
The comparison nobody runs
The switch decision is one line of arithmetic. You're weighing a certain cost against a probable one, and both sides come out in dollars.
Switch when: (days short × daily demand × unit contribution) > (order quantity × price premium per unit)
The left side is what the gap costs you. The right side is what closing it costs you. Every input is something you already have.
Run it on my box. Daily demand was 34 units. The shortfall between what the primary could deliver and when I needed stock was six selling days. Contribution on the bundles that box sits in is about $9 a unit, after the box itself, the product inside it, and the labor to personalize it.
- Cost of the gap: 6 × 34 × $9 = $1,836
- Cost of the premium on a 3,000-unit order: 3,000 × $0.46 = $1,380
The switch cleared by $456, and that is before anything else is counted. Two of those bundles were wedding orders with fixed dates, so the sale doesn't simply arrive later.
A note on that $1,836, because it is deliberately the smallest honest number. I've argued elsewhere on this blog that lost margin is only about 15 to 20 percent of what a stockout really costs you, once the rush freight, the wasted ad spend, the support load, and the customers who quietly don't come back are all in the total. Contribution alone understates the left side of that inequality by a multiple. I keep it in the test anyway, because a conservative left side means the test never pushes me into an expensive order I didn't need. When even the floor says switch, there is nothing left to deliberate about.
Flip the inputs and the answer flips with them. Had I been sitting on 40 days of cover with a primary who was merely slow, the left side collapses toward zero and that $1,380 is money set on fire. You are only ever paying the premium for days it actually buys you.
Two inputs that get fumbled
Days late and days short are different numbers. A supplier running eleven days behind only hurts across the days you were genuinely out of stock. Five days of cover absorbs five of them, which is precisely what buffer stock is for. Size the left side on your exposure, and if you don't know your exposure, that is the first thing to go and measure.
A quoted lead time on a supplier you haven't used is a sales promise. Four days is what they say when they want the order. If you've placed fewer than half a dozen orders with them, you have no measured figure to put in the formula, and the gap between quoted and actual lead times is its own expensive problem.
Dormant backups aren't backups
The part that actually caught me in July was that a second supplier has to be kept warm.
Ours hadn't taken an order from us in fourteen months. So I had no idea whether $2.56 was still the price, or how long reactivating a dead account would take now that our old contact had left. All of it had to be sorted out while I was already short, which is the worst possible moment to be finding anything out.
A backup you can use on the day needs four things on file, and none of them survive neglect:
- Their product code for your SKU, because your code means nothing to them and the back-and-forth just to identify the item costs a day on its own
- A lead time you have measured from your own received orders
- A price confirmed inside the last six months
- At least one order placed recently enough that you're still a live account
I now push a small order through every secondary supplier once or twice a year, deliberately, at a quantity I'd have bought anyway. It costs a bit on unit price. It buys a supplier who picks up the phone.
The seasonal version of this
There's a version of the calculation that shifts in the back half of the year, and September is the right time to look at it.
Daily demand sits on the left side of the inequality. As a SKU moves into its peak, that side grows while the price premium on the right stays flat. A switch that was clearly wrong in June can be clearly right in November for the same SKU at unchanged prices, purely because daily demand tripled.
Which means a standing preference for the cheaper supplier isn't a policy, it's a habit. The threshold has to be recalculated against current demand, and for anything seasonal your July answer tells you very little about your November one.
What to do this week
Take your ten highest-contribution SKUs and ask whether each one has a second source. Where the answer is no, decide whether it should. A single-source SKU carrying real volume is a risk you're holding whether or not you've ever priced it.
For the ones that do have a second source, write down four numbers: the two unit prices and the two measured lead times. Then work out the daily demand at which switching pays, and record that figure against the SKU. Having it written down turns the decision into a thirty-second check on the day, rather than the afternoon of second-guessing it cost me.
Stocura carries a primary and a secondary supplier on each SKU, with the primary driving the reorder trigger and the alternative available when you go to raise the order. Each supplier's lead time is learned from your own purchase order history and takes over from the manual estimate once enough orders have been received to trust it. Each SKU also carries the supplier's own product code, so the PO goes out in their language rather than yours.
Start with the SKU that scared you most this year. Mine was the box.
Two suppliers per SKU, both lead times learned from your own orders.
Stocura sits on top of your Cin7 Core data and keeps the alternative in front of you at the moment you're placing the order. Free until August 1, 2026 during soft launch.
Start free trial