Repricing Without Losing the Sale: How to Test a Price Increase

When a supplier raises your cost, the arithmetic for the hold-margin price is easy. The harder question is whether customers will still buy at that price. You can't know that in advance, but you can work out how many sales the new price is allowed to lose, and then test the increase on a slice of the catalogue before you roll it out everywhere.

The setup

One SKU runs through the whole article: a ratchet strap set. It costs $18.00, sells for $30.00 and moves about 120 units a month. That is a 40% margin and $12.00 of gross profit per unit, or $1,440 a month.

The supplier's new price file puts the cost at $19.80, a 10% rise. At the old retail of $30.00 the strap now earns $10.20 a unit, a 34.0% margin, and $1,224 a month at the same volume. To get back to 40% you would need $19.80 ÷ 0.60 = $33.00. The formulas behind that step are in how to recalculate your margin after a cost increase.

So if you move to $33.00, how many of those 120 sales can you lose before the increase does more harm than good?

The break-even volume loss

Last month's $1,440 no longer exists at the new cost, so compare the two options you have now: keep $30.00 and earn $10.20 a unit, or move to $33.00 and earn $13.20 a unit.

The formula:

max volume loss = 1 − (unit profit at old price ÷ unit profit at new price)

For the strap: 1 − 10.20 ÷ 13.20 = 22.7%. At $33.00 you can sell 22.7% fewer units, about 93 instead of 120, and still earn the same $1,224 you'd earn by doing nothing. Lose fewer and the increase pays; lose more and holding the price would have been better.

The same calculation for a range of prices, all at the new $19.80 cost:

A 10% rise on a 34% margin leaves a lot of room: you would need to lose more than a fifth of your volume before it hurts. And the thinner the margin, the bigger that room gets, because each extra dollar of price is a larger share of what you earn per unit. Stores with thin margins tend to fear price increases the most, and they are the ones with the most slack to absorb lost units.

What elasticity means for a small store

Price elasticity is the percentage change in units divided by the percentage change in price. If the strap goes from $30.00 to $33.00 (+10%) and monthly sales fall from 120 to 108 (−10%), elasticity is −1.0. Gross profit at 108 units is 108 × $13.20 = $1,425.60, against $1,224 for holding the price. Raising the price earns $201.60 a month more.

The break-even elasticity for this move is 22.7% ÷ 10% = −2.27. Customers would have to be more than twice as price-sensitive as that −1.0 case before raising to $33.00 lost money.

You can't borrow this number. Published estimates cover categories and whole markets and don't transfer to one store with its own traffic, customers and competitors. What you can do is look at which of your SKUs are likely to be less sensitive:

• Specific parts, where a customer who needs one part number for one machine shops around less than someone buying a generic item.
• Urgent purchases: when a broken part stops work, a few dollars of saving rarely matters.
• Small add-ons bought alongside a bigger item, which rarely get price-checked on their own.
• Listings that are hard to compare, because the same SKU at the same pack size isn't easy to find elsewhere.

The opposite cases, like commodity items, exact-match listings sold by many stores, and items people search for by price, are where you should test before you raise.

How to test an increase without an A/B tool

Shopify shows one price per variant to everyone, so a clean split test on a single product needs extra tooling. A simpler method works for most small stores: raise the price on one group of similar SKUs, leave a comparable group alone, and compare the two over the same weeks.

1. Pick two groups. Choose SKUs from the same category with similar volume. Put half in a test group and half in a control group. Ten and ten is enough to start.
2. Record a baseline. Pull units and gross profit for both groups from the sales by product report for the last four full weeks.
3. Change only the test group. Raise prices to the level you are considering. Don't run discount codes or promotions on either group during the test.
4. Run four full weeks. Full weeks cancel out weekday patterns. Avoid holiday periods and the start of a season.
5. Compare against the control. The control group tells you how much of the change came from the market and how much came from your price.

Say that over the four weeks before, the test group sold 480 units and the control group 400. Over the four weeks after, the test group sold 444 units (−7.5%) and the control group 384 (−4.0%). Part of the test group's drop happened to everything, so remove it:

change due to price = (444 ÷ 480) ÷ (384 ÷ 400) − 1
                    = 0.925 ÷ 0.960 − 1
                    = −3.6%

A 3.6% loss is far inside the 22.7% break-even for a 10% rise, so the increase can go to the rest of the category. If you had only looked at the test group, you would have blamed the price for a 7.5% drop it mostly didn't cause.

Decide the stop rule before you start

Small numbers are noisy. A SKU selling 120 a month sells about 28 a week, and a slow week can look like a collapse. Write down before the test what result makes you roll back, and stick to it:

• Roll back if the change due to price is more than half the break-even loss. For a 10% rise on the strap, that is about 11%. Using half leaves room for noise.
• Extend if the groups are too small to read. Below roughly 100 units per group across the test, run another four weeks rather than deciding on a handful of orders.
• Roll out if the loss is well under that line, and keep checking the group once a month after.

Slow movers that sell two or three a month can't be tested one by one. Group them by category and test the group, or price them straight to your margin floor, because the profit at stake on each is small.

Keep the loop running

A test only helps if you notice the cost change in the first place. The repricing lag in the KPIs for a reselling store measures how long you sell at the old price after a rise, and each strap sold at the old price during that lag gives up $3.00. CostSentry compares each new supplier price list with the last one and shows the margin at your current retail and the price you would need to hold your target. That gives you the starting point for the break-even calculation above. To see the same numbers for one price file right now, the supplier margin checker works without an account.