Dinner kit company Gousto has developed technology to help tailor its recipes to individual customer tastes.

The online start-up, which sells boxes containing the raw ingredients and recipes needed to make meals from scratch, has started using predictive modelling to help it work out its customers’ preferences and tailor its kits accordingly.

“If you choose 10 recipes, we should be able to tell what you like to eat and be able to present you with more of what you want,” said Gousto founder Timo Schmidt.

At present, it is only sending personalised suggestions via email, but it is working on personalising the website.

Gousto head of operations Ricardo Hummel and chief intelligence officer Tim Haines have also created an algorithm that can determine the optimum order in which ingredients should be packed. Taking into account variables such as weight, size, shape and appearance, the algorithm helps prevent breakages and spillages and ensures the food looks attractive.

Gousto has also overhauled its error checking technology and can now identify the cause of errors in real time so they can be corrected quickly.