Substantial efforts are devoted to increasing consumption of preventive health products such as mosquito nets and chlorine tablets, including through door-to-door marketing. This method is likely to increase sales by leveraging pressure to purchase due to guilt or reciprocity. But without the intent to use, these purchases can represent a loss for poor households in developing countries.

A field experiment in promoting water chlorination tablets in urban slums of India finds evidence of such purchasing pressure and underconsumption. But the study by Camille Boudot and Anita Mukherjee, to be presented at the Royal Economic Society''s annual conference at the University of Sussex in Brighton in March 2018, also finds that co-selling of a well-known side good can provide considerable gains by serving as both a positive frame and opportunity for opting out.

Access to clean drinking water is a daily concern for nearly 800 million people, and diarrhoea from unsafe water kills 800,000 children under five years old each year. Chlorination of water offers one possible solution to this problem.

Water purification tablets, like many other preventive health products, are known to have considerable positive externalities. But they are underconsumed by poor households, perhaps due to informational or cost constraints. Increasing consumption is therefore the goal of numerous agencies.

Informational campaigns can be costly and ineffective, and interventions that work typically require large subsidies or free distribution. Many organisations therefore devote substantial energy to pressured sales through door-to-door marketing, relying on the assumption that purchase leads to usage. Pressure, however, can backfire if it encourages purchases among individuals who only care about the guilt of saying ''no'' and do not plan to use the product.

Through a field experiment conducted in urban slums of India to promote water chlorination tablets, the researchers evaluate the effect of door-to-door marketing techniques on purchase and usage decisions. In particular, they assess the impact of co-selling a side good that can affect both marketing pressure and the inferred quality of the main product.

The researchers find the same proportion of purchases for the chlorination product among individuals in the treatment condition having received the main product alone or with the side good sold at retail price. Usage however is significantly higher in the latter condition, with approximately double the free chlorine content detected in the households'' drinking water. This provides evidence for the presence of a dominating positive framing effect from co-selling a well-known item.

Importantly, the price of the second good is critical. A third treatment condition received the side good on a promotional offer – ''buy one get one free''. Individuals in this condition where 10% more likely to purchase the chlorination product, compared with those in which the product was sold alone. But when measuring usage, this group showed no additional levels in drinking water free chlorine content.

The discount appears to have ramped up marketing pressure, perhaps via feelings of guilt or reciprocity toward the marketer, and undone nearly all of the welfare gains from co-selling a side good at retail price.

This study shows that marketing techniques, such as co-selling of a well-known side good, offers organisations a low-cost intervention that can positively affect product purchase and usage decisions by poor households. The authors also find that a commonly used option for many well-intentioned organisations, discounts and promotional offers, can generate behavioural responses from pressure that result in unused purchases of the main product.