Out of home businesses - marketing strategies: research

2. Method

Overview

This research involved 316 'mystery shopper' visits to 158 OOH premises and 20 online visits to a range of outlets.

Sampling of outlets

Overview

The study used a purposive sampling approach for both in-premise and online visits. The aim of this sampling approach was to achieve a spread of types and locations on the Scottish mainland of OOH outlets rather than a representative sample.

This was achieved by sorting a purchased sample of 1,000 OOH outlets in Scotland by brand (or a separate category for independent businesses) and then selecting individual outlets within each brand, plus independents, to achieve a spread across four deprivation and urban/rural categories. The sample included more from urban centres because many brands have far fewer locations in rural and semi-rural locations.

In-premise

A sample of names and locations of 1,000 OOH outlets falling into the categories of interest for the study was purchased from the specialist data company, Local Data Company. Outlets were then selected for inclusion based on the sample plan agreed with Scottish Government. The categories of outlets and specific brands included in the research were based on the most used types of OOH outlets in Scotland. In 2019, the six most visited categories of OOH outlets in Scotland were convenience stores (17.1% of visits), bakery and sandwich shops (16.1% of visits), quick service restaurants (QSR) (12.6% of visits), supermarkets (12.2% of visits), workplace and education canteens (10.3% of visits), and cafes (9.2% of visits) (The Out of Home Environment in Scotland, Food Standards Scotland, 2019). It was agreed workplace and education canteens would not be included given they are only accessible for those who work or study in the premises. Cafes were therefore included instead, which made up 9.2% of OOH visits in 2019, compared with 10.3% for canteens.

Around three fifths (57%) of visits to OOH businesses across Great Britain in 2020 were made to chains and around two fifths (43%) to independents. Independent businesses were therefore included in the study (they were not in the 2017 FSS study). Table 2.1 shows the five outlet categories included in the in-premise part of the study. Three market leading brands were chosen within each of these categories. Independent businesses were included for all categories except supermarkets. The QSR outlets selected included a mix of the most popular types of takeaway food including burgers, fried chicken and pizza. The supermarket OOH businesses comprised the smaller high street branches of the largest supermarket brands, and a small number of their outlets in petrol stations. We aimed to achieve around the same number of visits for each brand across each category. However, in some rural locations some brands were not available – in this scenario, outlets with similar characteristics were used instead.

Table 2.1 also shows the number of outlets visited as well as the purchases shoppers were asked to make. Food and drink purchases were agreed with the Scottish Government and Food Standards Scotland. This was informed by existing data on the most frequently consumed foods outside the home in Scotland, in addition to being relevant to the particular business offering of that brand. They were also selected to allow for an opportunity for promotions and upsizing to be offered by staff.

Every outlet in the in-premise research was visited twice within the space of a week. In Visit 1 shoppers were asked to purchase typically HFSS products and in Visit 2 they were asked to buy healthier options. See below under Data Analysis for description of food classification. In both scenarios, shoppers were asked to buy specific items or equivalents.

The in-premise sample was also designed to include a spread of levels of deprivation (using SIMD 2020 deciles) and levels of rurality using the Scottish Government Urban Rural Classification 2020 six-fold scale of: large urban areas, other urban areas, accessible small towns (bands 1-3); and remote small town, accessible rural, and remote rural (bands 4-6).

Businesses were categorised in four ways: rural and deprived; rural and non-deprived; urban and deprived; and urban and non-deprived. These relatively broad categories of deprivation and rurality were necessary due to the very small numbers in some of the categories. Table 2.2 shows how these categories were defined, and the number of premises visited in each of these. While the research aimed to achieve a roughly representative spread of levels of rurality, larger brands having fewer premises in rural areas meant the final sample included a higher proportion of businesses in urban locations. Where we were unable to find an outlet in a rural area, we identified and used the nearest match in our sample. There was an even split of deprived and non-deprived locations (exactly 50/50 with 79 premises in each).

Online

The online sample was also taken from the purchased sample list. The 20 online visits included large brands and independents in the three categories where it is possible to make OOH purchases online: bakery / sandwich shops; QSR; and cafes and coffee shops. For the online research, only one visit was made, as the experience was considered to be very similar whether a person orders HFSS products or not. We are aware that online shopping experiences can be more personalised than in-premise experiences due to the use of algorithms, but in this instance, it was felt there would be no opportunity for OOH websites or apps to 'learn' the preferences of shoppers.

The visits were made on either the businesses' own apps (e.g. the Starbucks app) or on an aggregator app (e.g. Just Eat, Deliveroo, Uber Eats). Overall, 14 visits were made on the aggregator apps and six were made directly on the outlets' apps. Just two out of the seven visits to independent businesses were made on outlets' apps (as the majority do not have their own app or website to order from). Table 2.3 provides details of the number of visits made in each category and the purchases made.

Table 2.4 shows the levels of deprivation and rurality according to the locations of the online sample. It was not possible to visit many online OOH outlets in rural locations because they do not tend to have an app or website for online ordering and delivery services, such as Deliveroo, do not service rural locations. Only one online visit was for an outlet in a rural location. The sample contained a fairly even split of businesses in deprived and non-deprived locations (11 in deprived areas and 9 in non-deprived areas).

Questionnaire design and fieldwork

Two questionnaires were developed by the research team – one for in-premise visits and one for online (these can be found in Appendix A). The in-premise questionnaire had two parts, for the two separate visits. This approach was taken to ensure a manageable amount of data was collected during each visit; and also to explore any differences in promotions and marketing strategies when shoppers purchased HFSS and non-HFSS food and drinks. Visit 1 covered price promotions, including meal deals, and staff interactions. Visit 2 covered staff interactions, nutritional information and any further promotions the shopper noted. Both in-premise visits were undertaken by the same mystery shopper.

The online questionnaire was designed to be as similar as possible as the in-premise one, to allow comparisons to be made. The staff interaction section was replaced by an 'ordering' section which asked about whether the app prompted them with any price promotions or deals.

All fieldwork was undertaken by a team of specialist mystery shoppers, using the Ipsos mobile phone app 'Ipsos Shopmetrics' to complete the questionnaire. The use of a mobile app to collect the data allowed shoppers to fill out the survey while in-premise, or soon after leaving the outlet, or while completing the online order - and therefore potentially record more accurate data than they would have done if they had completed the survey later that day.

The shoppers were fully briefed on the aims of the research and were provided with instructions on how to conduct the survey (Appendix B). Careful consideration was given to the in-premise instructions, in order to allow opportunity for upsizing and upselling and also to ensure comparability between visits. Shoppers were instructed which products to purchase (as detailed in the sample table above) and (where relevant) exactly how to ask for it within each outlet they visited. This included asking those visiting coffee shops and cafes to specifically ask for a 'latte', being careful not to mention the size initially, to then see whether they were encouraged to buy a larger size by staff. The online visit instructions were similar and also specified which app to use.

The main fieldwork was undertaken between 4 July and 21 August 2022. A total of 316 in-premise visits were made (two visits in 158 outlets), and 20 visits were made online – making a total of 336 mystery shopping visits to outlets across mainland Scotland.

Data analysis

The raw data from the survey was analysed using SPSS Statistics. The analysis of the survey comprised question by question analysis examining both the frequency of response to each survey question, as well as the extent to which responses varied by outlet category. It also involved looking for any differences or patterns according to deprivation and rurality. Throughout the report, we have only commented on any differences by SIMD/urban/rural category where a clear difference exists.

Food and drink categories were grouped into three types to provide a further level of analysis (including examining what types of products were promoted in Visit 1 compared with Visit 2). These were:

• Less healthy products including savoury pastries (sausage rolls, steak/chicken/cheese bakes), burgers, chips, fries, potato wedges, fried chicken, chicken wings, chicken nuggets, pizza with stuffed crust and/or meat, savoury sides (e.g. hash browns and onion rings), crisps, savoury snacks, cakes, biscuits, desserts, sweet pastries, chocolate, confectionery, cereal bars, shakes/frappes, soft drinks with added sugar, coffees (latte, cappuccino, coffees with syrups) hot chocolate.
• Healthier products including salads, pasta salads, vegetable sides, soup, fresh and dried fruit, nuts, water, low/no sugar soft drinks, pure fruit juice and smoothies, plain milk, and tea.
• Unclassified products including sandwiches, pasta, yoghurts.

This categorisation is necessarily a generalisation and pragmatic decisions were taken on the classifications of products. It is worth noting that within some of the healthier codes, that there may be items which are less healthy (such as smoothies with a high amount of sugar) and vice versa – and it was not possible to conduct a full assessment of each product in store to refine the classification, for example, for sandwiches.

Limitations of the research

There were some limitations which should be considered when interpreting the findings of this study:

• The number of outlets visited for each brand was small (seven to 10 outlets) and the outlets were not randomly selected from all outlets of that brand in Scotland (as discussed above, they were purposively sampled to provide a range across different area types). The results are therefore not statistically robust estimates of the prevalence of promotions across all outlets or across all outlets within each brand. We cannot say, for example, that 'the majority of price promotions in bakery outlets in Scotland are X' or '15% of QSR outlets use X marketing strategy'.
• Due to the small numbers of each brand, the urban/rural and deprivation categories used were necessarily very broad. A larger sample, with the scope to analyse area differences in more detail, may have revealed more differences. In addition, the deprivation classification of an outlet location may not reflect the customer profile.
• Because of the small sample size, the online data should be viewed as indicative of the types of promotions now being used by large brands and independent OOH outlets. That said, while only 20 visits were made online, the data reflects the online experience of consumers using the same apps across Scotland. For example, the user journey and promotions offered on the Costa app in Aberdeen, will be the same for someone using it elsewhere in Scotland.
• The research focused on the marketing strategies that were evident. It did not explore the consumer response to those promotions.
• The mystery shoppers were provided with detailed instructions on how to undertake the survey and, as far as possible, the questionnaire was designed to minimise variation in the way that promotions were recorded. Inevitably, however, there was potential for variation in some instances in the way the same or very similar promotions were recorded by different mystery shoppers. This issue has been mitigated as far as possible during data validation and cleaning.
• The online research undertaken recorded promotions that appeared to shoppers while they were making a purchase, it did not capture any promotions received as push notifications, texts or emails from OOH businesses.