Buyers of handmade goods face the quality uncertainty that Akerlof (1970) analysed: the seller knows how an item was made and the buyer, at a distance, does not. Research on online shopping places trust beside usefulness and ease of use as a determinant of intention to buy (Gefen et al., 2003). Social commerce supplies its own trust-building devices. Ratings and reviews, recommendations and referrals, and communities in which buyers talk to each other have been shown to shape trust and purchase intention (Hajli, 2015), and reviews of the field describe these social cues as its distinguishing feature (Zhang and Benyoucef, 2016). Sellers also benefit from links to one another, since connections among shops make each easier to find (Stephen and Toubia, 2010).

Three theories of adoption are relevant. Diffusion theory holds that an innovation spreads faster when potential adopters perceive relative advantage, compatibility with existing practice, low complexity, and the chance to try it and observe results (Rogers, 2003). The technology acceptance model reduces individual adoption to perceived usefulness and perceived ease of use (Davis, 1989). The technology-organisation-environment framework locates a firm's decision in the technology available, the organisation's resources and structure, and its external setting (Tornatzky and Fleischer, 1990); it is among the most used models at the level of the firm (Oliveira and Martins, 2011).

Work on developing economies adds two cautions. Molla and Licker (2005) showed that electronic commerce adoption there depends on the organisation's own perceived readiness and on external readiness, including the support of market forces, government and supporting industries such as payment and logistics. Heeks (2002) attributed the frequent failure of information systems in developing countries to the gap between the assumptions built into a design and the realities of the place where it is used. A framework for rural cooperatives must therefore begin from local conditions and not from the features of a platform.

A cooperative differs from a small firm in ways that matter for adoption. It is owned and democratically controlled by its members, who also share in its economic results (International Co-operative Alliance, 1995). A decision to sell through social media is a collective one, and the gains and the work it creates must be divided by rules that members accept. Ostrom (1990) found that durable self-governing institutions share certain design principles, among them clear membership boundaries, rules made by those affected, monitoring, graduated sanctions and accessible means of resolving disputes. These principles apply as well to a shared online shop as to a shared irrigation canal.

The first stage establishes whether the cooperative can and wishes to proceed. Readiness has an internal side, covering members' devices, network coverage in the village, literacy and language, and any prior experience of selling online, and an external side, covering the reach of courier services and the payment methods that urban buyers use (Molla and Licker, 2005). The findings are put to the membership for decision. A recorded mandate matters because the later stages commit shared money and reputation, and because members who doubt the usefulness of the venture are unlikely to supply products or labour for it (Davis, 1989).

The second stage chooses what to sell and prepares the evidence of its origin. A narrow opening range is preferable: items that photograph well, survive transport, can be made to a consistent standard and are priced so that delivery cost is a small part of the total. For each item the cooperative records who makes it, from what materials and by what technique. This record is the raw material for the content on which social commerce depends, and it answers the buyer's uncertainty about authenticity directly (Akerlof, 1970). Where a craft holds a geographical indication or similar certification, that fact belongs in the record.

Platforms differ in what they are good for. The functional building blocks proposed by Kietzmann et al. (2011), which include identity, conversations, sharing, relationships, reputation and groups, give a vocabulary for the comparison. A visual service suits display and discovery. A messaging application suits negotiation, custom orders and repeat buyers. A marketplace with social features adds payment protection and logistics at the cost of commission. The framework recommends beginning with no more than two channels, chosen by where the intended urban buyers already spend time and by what members can operate, and holding the account in the cooperative's name so that it survives changes of personnel.

A buyer who has never met the seller needs reassurance at three points: paying, receiving and complaining. The cooperative should offer payment methods familiar to its buyers, state delivery times honestly, pack to a written standard and publish a simple policy on damage and returns. Each completed order is also an opportunity to gather the reviews and referrals on which later buyers rely (Hajli, 2015). Trust built this way accumulates in the account's visible history, which is one more reason for the account to belong to the cooperative and not to an individual.

Online selling creates tasks that craft production does not: photography, writing, replying to messages, packing and bookkeeping. The fifth stage assigns them, decides whether they are paid, and fixes how prices are set and how proceeds are divided among the maker, those who do the selling work and the common fund. Following Ostrom (1990), the rules should be made by the members they bind, be simple enough to monitor, and provide a way of settling disagreements. Records of orders and payments that any member may inspect are the practical form of monitoring. Education and training of members, itself a cooperative principle (International Co-operative Alliance, 1995), spreads the new skills beyond a few younger members.

At fixed intervals the cooperative reviews what has happened: enquiries received, orders completed, complaints, costs, and the time members have spent. The review follows the logic of action research, in which a group diagnoses a problem, plans, acts, evaluates and records what it has learned before the next cycle (Susman and Evered, 1978). Its results may send the cooperative back to an earlier stage, to change the product range, drop a channel or revise the revenue rules. The attributes of trialability and observability are served by this design: a modest first cycle lets members see results before committing more (Rogers, 2003).