Whopper Detour: the work behind the joke

The business problem behind the detour

Burger King needed adoption of an updated ordering app in a category where order-ahead was already familiar. The harder task was persuading someone to interrupt an established buying routine. A download could remove one barrier while leaving payment setup, first order and repeat use unresolved.

The useful strategic question is therefore: which customer behavior should the campaign make easier to repeat? Here, the promotional purchase passed through the app. That made the acquisition offer a rehearsal for a future ordering habit. Whether that habit persisted requires evidence beyond the initial response.

Three audiences, three different jobs

Treat location as eligibility, not proof of preference: standing near McDonald’s does not establish that someone is its customer. These segments suggest different measurement questions, rather than claims about the campaign’s actual media targeting.

• Existing Burger King buyers without app habits: the opportunity is channel adoption. Count app activation, but check whether total purchases increased or merely moved from the counter.
• Lapsed or competitor-oriented buyers: the offer can justify reconsideration. Ask whether a second, unsubsidized purchase occurs after the novelty ends.
• Deal seekers and spectators: one group may redeem without returning; another may share without visiting. Their contribution belongs in separate acquisition and attention measures.

A long build, then a short invitation

Machado’s account shows an important planning distinction: nine days described the public promotion, not the development effort.

• September 2017: FCB New York and Waze proposed rewarding drivers who passed McDonald’s en route to Burger King.
• Later in 2017: the idea moved toward the BK app. Burger King chose its regular Whopper rather than a special product. Development took roughly a year.
• December 4–12, 2018: customers could unlock a one-cent Whopper within 600 feet of eligible McDonald’s locations, order in the BK app, then follow directions to Burger King. The launch release specified one redemption and participating restaurants.

The interface had to finish the story

Technology partner mParticle describes a chain connecting deep links, a location event, messaging, ordering and redemption analytics. Its account names Branch, Radar, Braze, Tillster and Amplitude alongside mParticle. The creative promise depended on those handoffs working together.

An analogous team should map failures before buying attention: installation succeeds but location fails; eligibility appears but ordering is unavailable; payment completes but collection is unclear. Each needs an understandable recovery path. Sky News reported customer complaints about app access and nearby restaurants lacking mobile ordering. These are anecdotes, not a failure-rate estimate, but they reveal how the detour could send a disappointed buyer to the rival.

Paid, owned and earned had different jobs

Machado describes modest digital promotion, print and outdoor activity, without a disclosed spend total. He reports no television, radio or influencer program. Describing the campaign as free virality would erase production, technology, fulfilment and food subsidy.

Owned social could make the contradiction noticeable; the app had to make the offer usable. Paid placements and promoted film could seed encounters beyond existing followers. News coverage and customer discussion could carry the premise further, with less control over explanation. These are plausible channel roles, not measured channel-level contributions.

The practical lesson is to assign two distinct communications jobs: make the invitation memorable, then explain the conditions accurately. A headline that spreads faster than its instructions can generate both demand and confusion.

What changed, and what remains unresolved

The One Show entry reports more than 1.5 million downloads and 500,000 redemptions, with app sales value tripling during the promotion and remaining twice the earlier level afterward. These are entrant claims; the public entry supplies neither customer-level data nor a controlled comparison.

A counter-explanation is substantial price subsidy. Others include publicity, improved app functionality, existing customers switching ordering channels and temporary curiosity. The campaign combined these factors, so the evidence cannot isolate the effect of the competitor joke or geofencing.

Measurement wording also matters. The award entry reports 27:1 ROI; the brand retrospective reports 37:1. No reconciliation is supplied. A later vendor page’s six-million download figure is cumulative, not the nine-day increment. Do not merge these into a stronger headline.

Nor can the two rounded lower bounds establish an exact redemption rate: downloads and redemptions are different events and may involve different populations. Long-term value needs defined cohorts, repeat behavior, net revenue and costs, with an explicit comparison period.

A small-team test without geofencing

For a booking tool, test whether a “calendar rescue” invitation improves first completed bookings. Use a fictional scheduling example and let users enter their own availability. Avoid importing private messages or adding location collection.

First instrument two weeks of ordinary traffic to establish booking completion, errors, support contacts and contribution per eligible visitor. Then randomly assign new eligible visitors to a plain invitation or challenge framing. Keep the reward, eligibility, interface, traffic sources and fulfilment identical. Retain assignment across repeat visits.

Primary outcome: a first confirmed booking within seven days per assigned visitor. Include nonstarters in the denominator. Report absolute lift and an uncertainty interval. As a planning example, a 10% baseline and a four-percentage-point detectable increase require roughly 1,040 visitors per arm at conventional 80% power and two-sided 5% significance. Recalculate for actual conditions. This assumes independent visitor-level assignment; account for missing follow-up, clustered exposure or repeated observations before choosing the final sample.

Set the cash ceiling and minimum worthwhile lift before launch. Observe 30-day repeat bookings and contribution after credits, acquisition and support costs. Pause for a privacy incident, booking errors above 2%, or offer complaints above three per 100 participants. Choose stricter limits where needed.

Decide at the planned endpoint, not after a favorable daily spike. Scale only if the lift is sufficiently precise and economics and guardrails hold. If traffic cannot support the sample, run a usability pilot and report feasibility, not effectiveness. This comparison tests framing; a later no-reward holdout would test the subsidy.

Three exercises for the next brief

• Draw the handoffs, 15 minutes. Sketch invitation, eligibility, account, action and fulfilment. Mark one failure and recovery at each boundary. Deliverable: a journey map with a named operational owner for every step.
• Challenge the headline, 10 minutes. Rewrite “the stunt created loyal customers” into one supported observation, one hypothesis and one missing measurement. Deliverable: a three-line claim audit a colleague can verify from the sources.
• Design the fair comparison, 20 minutes. Write a plain invitation and a surprising one for the same product action and reward. Specify assignment, denominator, success threshold and stop rule. Deliverable: a one-page experiment brief that explains what the result cannot prove.