The following is a guest post we’re pleased to share by Mireya Almazan, Associate Program Officer and Ignacio Mas, Deputy Director from the  FSP program at the Bill & Melinda Gates Foundation.

Building and managing a cash merchant network is the least scalable component of a mobile money scheme, as stores need to be sought out, evaluated, contracted, trained, branded and supervised individually. And driving sufficient transaction volumes through that channel is critical in order to make sure that the cash merchants continue to invest sufficient working capital to maintain liquidity and promote the service. We use a metric of 50 transactions per agent per day as the kind of metric that makes a store truly dedicated to the mobile money proposition, or rather a point-of-no-return.

Achieving this is especially challenging in markets where several operators have launched their mobile money offering, but none has a dominant share. No operator has the muscle to demand exclusivity of the channel, and stores have little loyalty to mobile money schemes that simply don’t deliver enough to their bottom line on a daily basis.

In such cases, providers should consider joining forces to cut their losses by exploring commercial arrangements to share cash merchants.

Giving competitors access to anything a company has invested heavily in may sound like a controversial proposition, but there are in fact precedents for collaborating at an industry level in both the banking and the mobile phone sector. Over 21,000 banks around the world have access to the VISA platform, which enables sharing of costly physical infrastructure (branches and ATMs). And most mobile operators have come to the conclusion that sharing towers makes good business sense.

But why share cash merchants, the engine of a mobile money scheme and often a unique selling point for leading service providers?

Pooling the transactions of all mobile money providers into a common set of stores helps achieve cash merchant sustainability faster (those 50 transactions per store per day). It also gives customers a broader set of cash in/out points to choose from. And it permits most efficient use of cash pools in each community, avoiding a situation where one store has run out of cash while a nearby store has plenty of cash – except that it works with a different provider. In smaller rural communities, the cumbersome task of going to the branch to rebalance can be delegated to only one store – serving all.

In some markets, cash merchants are already servicing multiple mobile money schemes. This is the case in Tanzania, where four mobile money schemes are vying for increased market share. The mobile money providers compete to claim a marketing presence at the store, which in the best of cases becomes cluttered with flags and posters from three or four providers, and in the worst case the store simply refuses to hold much signage from any. As the mobile wallets are not interoperable, cash merchants need to hold e-value for each of the schemes. This is sub-optimal not only because it ties up scarce working capital, but because it fragments the liquidity pool and diminishes the likelihood that any one will have sufficient float. Additionally, there’s a risk that the cash merchants play the mobile money schemes against each other, driving commissions up.

The most compelling reason to share cash merchants is for the long-term development of the mobile money ecosystem. More and more merchants will accept e-money in fulfillment of payments, more employers will pay wages in e-money, and the government—often a country’s largest micro-payer—will begin to disburse civil servant salaries, pensions and other government-to-person payments through e-money. Thus, there will be less of a need for cash-in/ cash-out, and therefore less of a need for cash merchants. In the longer term, the cash merchant network is not where the action will be. Providers will instead want to differentiate on service attributes that are core to their DNA.

Sharing cash merchants will happen in the long-run. If operators don’t provide some degree of interoperability, stores and customers will take it in their own hands.  Why not do this in a commercially attractive manner, with terms set out by providers?