A field experiment on carbon footprint information provision in restaurants

Eight restaurants will introduce the “climate information” menus, while the remaining eight restaurants serve as the ‘control’ restaurants. In the ‘climate information’ restaurants, we will introduce table menus which will include carbon footprint information for all vegan swap meals on offer. A vegan swap meal includes a standard meat meal which can be made vegan and climate friendly by substituting meat and dairy ingredients with meat-free and vegan alternatives. Each menu contains four swap offers, across different food categories (e.g. Starters, Mains, Deserts). The default option is always the meat-based option, however, customers can ask for the vegan swap when ordering.

The “climate information” menus include the following message: “Make this Climate Friendly and Vegan with [details on swap ingredient], [Calories], reducing the carbon footprint by [X]%*”. The asterisk directs the customer to an additional info box, located in the bottom right corner of the menu, which contains the following message: “*Trying to cut down on your carbon footprint? – Small choices like what we eat can have a big impact on the planet. That’s why we’ve been working with [company name] to understand the environmental impact of some of our dishes and where we can make a difference. Keep an eye out for the “Make this Climate Friendly and Vegan” swaps on our menu to see how you can reduce the carbon footprint of your meal by making a simple switch. Want to know more? Visit [Website link].”

In the control restaurants, no information about the carbon footprint of the low-carbon alternative is provided. Sales data for both the default and vegan swap alternatives will be recorded through the restaurant point of sales system. Any data which can be obtained on the sales of “vegan swaps” since their introduction and prior to the intervention will serve as baseline data. These were introduced in all restaurants in July 2021.

We will test the following main hypotheses:
H1: Provision of carbon footprint information for swap meals will increase the probability that customers order low-carbon swap meals.
H2. Provision of carbon footprint information for swap meals will not affect the total sales of these items in the treated restaurants.

The new menus will be introduced on 8th November and will be displayed until 21st March 2023, except between 21st November and 2nd January, during which a special Christmas Menu without any carbon footprint information will be used (this applies to both treatment and control restaurants).

2022-11-08

2023-03-20

The main outcome variable is the sales share of vegan swaps relative to the total sales of the default alternative.

We will also record total sales of the vegan swaps to assess whether the intervention has an effect on revenue/profit as secondary outcome variable.

The field experiment will be conducted in 16 restaurants at different locations across the UK, belonging to the same restaurant chain. All restaurants offer the same menu with respect to availability of meals, prices and cuisine.

While the restaurant chain operates over 50 restaurants across the UK, eight pairs of restaurants were selected for the experiment, where each pair is comparable on observable macro-level characteristics (e.g. customer base, socio-demographic profile, number of customers).

Treatment was assigned at the restaurant level: Eight restaurants will introduce the “climate information” menus, while the remaining eight restaurants serve as the ‘control’ restaurants.

Not available

The retail chain provided the researchers with a list of restaurants eligible for treatment and control status. In collaboration with the restaurant management, eight matched pairs were selected (comparable on observable characteristics) where one restaurant in each pair was assigned to the treatment group.

The retail chain provided the researchers with a list of restaurants eligible for treatment and control status. In collaboration with the restaurant management, eight matched pairs were selected (comparable on observable characteristics) where one restaurant in each pair was assigned to the treatment group.

No

N/A

Baseline period (13 weeks) + Treatment period (13 weeks): Daily sales data: N ≈ 3000 Weekly sales data: N ≈ 420

Daily sales data: N ≈ 1500
Weekly sales data: N ≈ 210