Personalising food recommendations on GrabFood

Learning what users want

User behaviour on the app such as order history and clicks gives us a good sense of the likes and dislikes of the user.

Over time, the recommendation system will form a more accurate profile of a person to give us deeper insight into their culinary preferences at different times of the day, budgets and willingness to wait for orders.

What users want to eat varies throughout the day. About two in five consumers snack at least once a week across the region, with cravings hitting most frequently at teatime between 3 and 5pm, according to data from our report.

(Read more: How Southeast Asia ate in 2022)

The top teatime food order in Vietnam last year was fresh spring rolls while that in Indonesia was dim sum. Across the region, drinks like milk tea and coffee were popular at teatime.

We take such information into consideration to generate a list of personalised options for each user on the home screen of the app.

On the other hand, users who already have a dish or restaurant in mind should be able to cart out their items as seamlessly as possible.

For example, those who frequent the same restaurant can repeat the order quickly through an “Order Again” section, without having to start a search from the top.

Our search algorithms also ensure that users can find what they are looking for with ease by ranking search results based on historical data.

For instance, if a user consistently patronises the same restaurant every time she searches for Japanese food, that particular restaurant could emerge as the top search result for the user when searching for Japanese food in the future.

It’s different in every country

Even as we continue to fine-tune our algorithms and recommendation system, we have to take into account that we serve a vast, non-homogenous region with diverse tastes and preferences.

That’s why respective country teams also play an important role in crunching data and customising sections on the app in order to curate options that cater to the local palate.

For instance, we observed that local food still dominated most markets last year. In Vietnam, bahn mi emerged among the top trending food orders for the year. In Malaysia, people popularly ordered roti canai and nasi lemak.

Helping users find more

While having the app know you is more convenient, users have told us that part of the fun of looking through a delivery app is in getting ideas for new options.

So we also increase the chances of discovering new food options on our platform. One way is to present options that deviate slightly from their go-tos.

We also recommend substitutes or similar options when a search for a particular dish or restaurant is unavailable, or has very few matches.