Why It Works: Dishoom and its secret sauce of surprise

Published in Marketing Week, April 2025.

Dishoom is known for two things: fabulous food and its matka game of chance. More than a mere PR trick, this offer is underpinned by behavioural science.

Do you live in one of the four lucky UK cities to be blessed with a Dishoom? If so, the chances are you’ll have sampled, or at least heard of, the tasty Bombay comfort food on offer.

The chain was founded in 2010 and has since grown to 13 eateries UK-wide. It’s won too many awards to count, and their black daal is legendary. Aside from the special food, Dishoom holds a few other delights up its sleeves. One is its mysterious matka.

This is a game of chance — roll a six at the end of your meal and you get it for free. The name comes from the Hindi for pot, in reference to the tiny urn you’ll use to shake the dice, and also fittingly refers to an illegal underground gambling game in 1960s Bombay.

To play, you show your special keyring when you ask for your bill, a nice touch that adds an element of exclusivity to the offer. You’ll need to have asked for a keyring on a previous visit so you can immediately see how it works to draw you back. But that’s not the only thing that brings you back.

The thrill of the game

The matka gamble is exciting because the prize is far from certain. There’s just a one in six chance you’ll win. And there’s very good evidence from psychological experiments that we are highly motivated by this element of chance — it’s the whole basis of gambling as entertainment.

One such study comes from Luxi Shen and her colleagues at the University of Chicago in 2015. They studied how uncertain rewards affect willingness to pay.

For the study, they asked 138 participants to make an offer to buy a bag of chocolate truffles. Half were told the bag would contain 4 truffles (certain reward), while half were told that it contained either 2 or 4 truffles, with a 50/50 chance (variable reward).

Participants bid an average of $0.66 for the certain reward – the 4 truffles – and $1.49 for the variable reward — an uplift of more than double. They preferred the variability by far.pe.

Results adapted from Shen et al.’s (2015) study

The researchers theorised that this was because participants focused more on the thrill of the bidding in an uncertain situation. The value comes in the chance to be a winner, not just the reward.

But this theoretical experiment is a far cry from any commercial setting. Can we reasonably extrapolate to food purchasing behaviour in the real world? Another study suggests so.

Again, in 2015, Nina Mazar from the University of Toronto led a study to test the influence of uncertain rewards in a real setting.

Using a vending machine, the researchers introduced a sales promotion. Snack-buyers were given a choice between paying a discounted price of $0.50 per item (fixed reward) or paying the regular price of $0.75, but with a one in 3 chance of receiving their chosen treat for free (uncertain reward). These are mathematically the same — 3 snacks equates to $1.50 in both options.

The researchers observed the sales of each pricing set-up.

In the fixed reward option, 84 snacks were purchased over two weeks. However, in the uncertain reward option, 120 snacks were purchased — a 43% difference. So, despite the two offers being of equivalent value, customers preferred the chance to win over the fixed price.

The Shen and Mazar studies explain the draw of Dishoom’s matka. We’re captivated by the thrill of the gamble. Even though the offer equates to a 16.6% discount — not much considering it can only be used at off-peak times — the impact is much greater than this for diners.

So much so that it’s quite likely we’ll choose to eat at Dishoom again, just to get a roll of that dice. Because, next time, who knows? A delicious meal — and it could be free. We’re bound to spread the word.

You might not work in the restaurant business but you can still learn from Dishoom. If you run a loyalty scheme, why not swap the standard fixed points approach and add a dash of randomness. That’s what Pret does. They empower their staff to give out free cups of coffee every so often rather than ask drinkers to collect points.

Or next time you run a promotion, rather than offering every person 10% off, could you give 1 in 10 customers their order for free? The cost to the business is the same, but the customer buzz is far higher.