Petro-Points: inside the launch

In the early nineties, fuel retail in Canada looked like a commodity fight with a car wash attached. Price boards set expectations, brand loyalty was thin, and the customer decided at the corner based on which side of the road they were driving on. Inside Petro-Canada's downstream business, with roughly 3,500 locations and about 200,000 transactions a day, tiny shifts in visit frequency were worth a great deal of money.

The strategic question was not how to discount. It was how to earn a reason to be chosen when the price on both corners was the same. That is where Petro-Points began.

First, we anchored the program to behaviour rather than to price. Points were earned for coming back, for filling rather than splashing, for using the site's full offer. The economics were built so that the reward funded a habit rather than subsidising a transaction that would have happened anyway.

Second, we insisted the program be explainable in a single sentence at the pump. Every additional rule costs comprehension, and comprehension is what drives enrolment and repeat use. Complexity is the most common cause of death for loyalty programs, and it usually arrives as a well intentioned improvement.

Third, we treated data as the actual product. A loyalty card turns anonymous volume into identified behaviour. That meant we could see frequency, basket, daypart, drift and lapse. It also meant we had an obligation to handle it with care, because trust is the currency the whole program runs on.

Fourth, we built for the field. If a site manager cannot explain the program, staff will not promote it, and customers will not join. Every operational decision was tested against the reality of a busy forecourt on a Friday afternoon.

Enrolment is not adoption. A signed up member who never presents the card is a cost with no signal. The metric that mattered was identified transactions as a share of total transactions, because that is the number that determines whether your data is representative enough to act on.

Redemption is a retention event, not a cost event. Finance teams instinctively want to suppress redemption. In practice, the member who redeems is the member who stays. The right conversation is about reward relevance and breakage assumptions, not about making redemption harder.

Partnerships multiply relevance but dilute clarity. Every partner added earning occasions and complexity in the same breath. We learned to add partners only where the member could describe the benefit without a brochure.

Field incentives beat advertising. A month of manager-level focus on enrolment quality moved the numbers more than a national campaign, and it cost a fraction as much.

Reward the behaviour you want. Treat the data as the dividend. Keep it simple enough to explain in one sentence. Those three rules have survived every technology cycle since, including the shift to apps, digital wallets and now AI-driven personalization.

What changed is the speed of feedback. In 1995 we waited a quarter to read a cohort. Today the same read takes an afternoon. That is an advantage only for teams with the discipline to decide something based on it. Faster data with unchanged decision-making produces nothing except better documented indecision.

What did not change is the human part. Loyalty is a posture toward your customer, expressed consistently over years. Programs communicate that posture. They do not substitute for it.

Most companies asking me about loyalty do not need a new program. They need to make the existing one legible and act on what it already knows. I usually start with three questions: what behaviour are you paying for, what decision have you changed in the last quarter because of member data, and can a front line employee explain the program without notes?

If the answers are uncomfortable, that is the work. It is also usually the fastest available margin in the business.