The interesting move here is the comparison, not the conclusion. Permission and provision are different products. A patio that tolerates a leash asks the owner to bring the water bowl; a listed item for the dog says the house planned for the animal before it arrived. What the Toronto Metropolitan team reports is an interaction: patio access does more work when it comes with dog menu items [3]. That is a claim about complementarity, and it is exactly the sort of claim a survey experiment can carry, because you can hold the restaurant constant and vary one line on a menu.
The outcomes are where care is needed. What was measured is intention to visit, reported spending and stated likelihood of recommending [1]. Those are attitudes about a described restaurant, and stated intention runs ahead of observed behaviour more often than not. The release's own language drifts across that boundary, promising a boost to customer traffic in one breath and reporting intention to visit in the next [12]. The thing this does not tell you is whether a dog menu changes party size, table turn time, or how often the same household comes back inside a quarter.
Nor does the published summary give the magnitude. No sample size, no effect size, no description of the comparison condition [9]. So the arithmetic an operator actually wants stays out of reach: payback is implementation cost divided by incremental margin per cover, and neither term appears [10]. Lead author Rishad Habib, of the Ted Rogers School of Management, frames small additions like a dog menu as translating into measurable business outcomes [2][4]. Measurable is fair. Measured, on a till, is a different sentence.
The context claim is easier to accept. Searches for dog-friendly restaurants have climbed since the pandemic-era rise in pet ownership, and pet-related spending has kept growing [6][7], while evidence on how pet-friendly features actually shift customer behaviour has been thin [5]. Thin evidence on a decision most venues make by feel is worth a paper.
It is also worth a better test, and this is one of the rare hospitality questions where a real control is cheap. Two comparable sites, or four weeks on and four weeks off against a matched location, with covers, average cheque and repeat visits as the outcome instead of a rating scale. Count the other side too: the release reports nothing about diners who do not own a dog or would rather not sit beside one [11], and they are on your covers as well.
My reading, with its conditions stated: the direction is plausible and probably holds for casual, patio-forward rooms in neighbourhoods with high dog ownership, where the marginal guest is an owner deciding between two similar terraces. Where the patio is small and turns fast, a dog at the table costs capacity, and I would expect the effect to sit near zero or below it. Nothing in the summary distinguishes those two cases, which is the number to ask the authors for.
