A closer look at the milk people say they would choose

A study published in September examines which liquid milk attributes appeal to household shoppers in Sri Lanka’s Northern Province. Rather than asking only whether consumers like milk, the researchers present combinations of product characteristics and prices. The aim is to understand the trade-offs shoppers make when more than one feature changes at the same time.

Across the modelled choices, respondents generally preferred low-fat milk, pasteurisation or UHT processing, enhanced food safety measures and reusable packaging. The paper also finds differences between consumer groups. For processors, that variation is as relevant as the overall ranking: a single product specification is unlikely to represent every shopper’s priorities.

Five attributes and a choice to keep the status quo

The experiment covered milk fat, processing technology, food safety measures, packaging and price. Fat levels included whole, low-fat and non-fat milk. Processing options included pasteurisation, sterilisation, UHT and no processing. Participants evaluated eight choice sets, each with two product options and a third option representing the status quo.

Researchers first used a pilot involving 40 people to develop the experimental design. The main survey involved face-to-face interviews with adult primary household shoppers across the province’s five districts. Of 417 questionnaires, 410 remained after incomplete responses were removed. Interviewers explained the attributes and reminded respondents to choose according to their real preferences, including the option not to select a new product.

The results do not support treating all reduced-fat milk alike

Low-fat milk and non-fat milk did not attract the same response. In the model, low-fat milk was preferred relative to whole milk, while the result for non-fat milk went in the opposite direction. That distinction matters for product development: combining the two into a broad reduced-fat category would lose information that the experiment was designed to reveal.

The paper also estimates willingness to pay for individual attributes and finds variation linked to demographic characteristics. Overall support for reusable packaging coexists with differences between groups in their preference for single-service packs. The findings therefore point towards segmentation, rather than establishing one universally preferred package or a price premium that every retailer could collect.

A regional survey, not a national sales forecast

The authors explicitly caution that their sample does not represent Sri Lanka as a whole. It is confined to the Northern Province and differs from national demographics, including in its proportion of female respondents. Concentrating on primary household shoppers is useful for the research question, but it does not capture every milk consumer or every purchasing occasion.

A second boundary concerns the choice task itself. Participants selected hypothetical combinations rather than spending money on products in a shop. Reminders intended to reduce hypothetical bias do not eliminate that difference. Nor does the experiment measure nutritional outcomes or demonstrate the safety of a processing system. The study gives processors a structured starting point for consumer testing; product trials and observed purchases would be needed to establish whether those preferences translate into a viable offer.

SOURCES & TRANSPARENCYHow we work ↗
dairyconsumer preferencesSri Lanka