Closing the Loop

The data was available the entire time. The question being asked of it was wrong.

The measurement trap

In 2016, the prevailing logic in digital advertising was that a click was the goal. For e-commerce businesses, this made sense. The next action you want from someone who sees your ad is to visit your online store.

For a QSR brand whose customers decide where to have lunch based on proximity and habit rather than online research, it did not.

McDonald's Canada had $11.8 million a year being optimised toward website clicks. The ROI looked poor, and this was being used as evidence that digital advertising simply did not work for McDonald's. The measurement was wrong, and the wrong measurement was producing the wrong conclusions.

The quick fix

We changed what we measured. The job of McDonald's digital advertising was identical to its TV, radio, and outdoor: keep the brand at top of mind. There was no reason to hold digital to a different standard.

We implemented MOAT and shifted to visibility as the optimisation goal, bringing digital into alignment with every other channel. Digital ROI doubled quickly. Not because we had invented anything new, but because we had stopped asking the wrong question of the data.

The closed loop

In 2017 Meta offered beta test partnerships for their new Store Visits product. Using location data from mobile devices, they could identify whether someone exposed to a McDonald's ad had subsequently visited a McDonald's location. For the first time, we could directly connect advertising exposure to a physical store visit.

The trap was obvious. Goodhart's Law: when a measure becomes a target, it ceases to be a good measure. Without sufficient historical data, we had no way of knowing whether an ad had inspired a visit or whether we had simply served an ad to someone already heading there. Optimising toward Store Visits would guarantee the latter.

We used Store Visits as a measurement of ad effectiveness, not a targeting signal. As we expanded the capability across the broader digital buy with the Weather Network, that discipline held. The data told us what was working. It did not tell us who to chase.

The outcome

Over two years, by changing what we measured and how we interpreted it, digital ROI improved more than fivefold.

The data required to reach that outcome was available throughout. The question being asked of it was wrong.

Ensuring alignment between the metrics used to understand performance and the goals you actually want to achieve is where marketing success begins.