You sold 11,847 jars of coffee. But you failed to prove the promotion sold even one extra jar.

Picture yourself sitting in front of your largest coffee supplier for the quarterly review. You present the deck with these whopping success figures: 11,847 jars sold. 117% of target. Redemptions up by 238%.

On the slide, last month’s promotion is a win.

But then the key account manager asks, “How many of these coffee jars wouldn’t have sold without the promotion?”

That number isn’t in the deck and you can’t answer it either because you never measured it.

Yet you’ve called the campaign a success and are now asking your key account manager to fund the next one.

When the key account manager declines to fund it, don’t call them difficult. You pushed that budget out of the door when you counted every sale but couldn’t show how many moved because of the campaign.

Ocado chose the shopper before it chose the coupon.

When PepsiCo wanted to bring light and lapsed Walkers buyers back, Ocado used the previous year’s baskets to separate shoppers by how recently and how often they’d bought the brand.

Customers in the bottom 25% of annual Walkers spend received £1 off. The same offer was given to shoppers who’d bought Walkers during the previous year, but not within the last 90 days. Regular crisps buyers received 80 pence off. (Ocado Ads, April 2025)

Those groups weren’t labels added to the report afterwards. They decided who received which offer before the campaign went live.

Ocado also kept a control group that saw the campaign but didn’t receive the personalised offer.

When the results came in, Ocado Ads reported 1,023% total ROAS and 204% incremental ROAS. Sales from reactivated customers rose 44% period over period, while new-to-brand sales for the featured products rose 177%. (Ocado Ads, April 2025)

The discount came second. Ocado first decided what it wanted each customer to do differently, then used the coupon to trigger that behaviour.

The shopper who bought the coffee is four different people in your system.

Your loyalty platform knows a member. Your CRM knows the coupon recipient. Your till records a jar of coffee. Finance records the discount, while Commercial records the supplier’s contribution.

Each file is accurate but none follows the shopper all the way through.

One shopper today

Loyalty, CRM, till and finance each hold their own view: a member, a coupon, a jar, a discount. Four accurate records, no shared ID. The control group is only a list of names.

One shopper, measurable

Loyalty, CRM, till and finance all resolve to one customer ID. Test and control become comparable. Incremental margin becomes provable.

The member ID in loyalty has to survive the handoff into CRM and appear again on the receipt. If it doesn’t, Finance can’t attach the discount or margin to the same customer either.

Without that link, your control group is only a list of names. You can’t compare what those shoppers bought with the shoppers who received the coupon.

In Malaysia, promotions already drive 28.6% of FMCG value sales. Worldpanel says retailers and manufacturers must determine whether those promotions create growth or merely move full-price sales into discounted formats. (Worldpanel by Numerator, May 2025)

Without one customer ID across loyalty, CRM and the till, you can’t separate the two.

Three questions your key account manager will use to judge the next campaign.

Put these three answers beside the sales total before your next supplier review.

01 · The sales that didn’t need the discount

What share of the shoppers who redeemed the offer had bought the same brand at full price on each of their previous three category visits?

These customers may be loyal and valuable, but the campaign didn’t win them. If you label their discounted purchase incremental, you’re charging the supplier to subsidise demand it already had.

02 · The margin the promotion actually added

Once you remove the sales that would’ve happened anyway, how much extra gross margin remained after the coupon and campaign costs?

A promotion can beat its sales target and still destroy margin. This is the number that tells your key account manager whether the campaign paid back, not whether discounted stock moved.

03 · The new buyers who came back

Of the shoppers who were genuinely new to the brand or reactivated, how many bought it again within 90 days without another coupon?

A first discounted purchase creates a redemption. A second full-price purchase shows a changed habit. Without the second, the supplier bought one cheap transaction, not a new customer.

See if your next promotion can follow one shopper to checkout.

Open the brief for your next large brand promotion.

Find the person who owns the audience file and the person who owns the till receipts. Ask both to show you the customer ID they use.

If the same ID appears in both places, ask the team to keep a comparable control group before the offer goes out. Agree that the final report will show incremental gross margin, new-to-brand buyers and 90-day repeat.

If the IDs don’t match, stop polishing the campaign deck. Before launch, connect the audience file and the receipt to one customer profile.

In five minutes, you’ll know whether the campaign is ready to measure.

Thanks for reading this far. Your next supplier cheque won’t be won by a larger sales total. It’ll be won by showing the key account manager which customer behaviour that money bought.