Since November we’ve been interviewing people who work for aid agencies in a variety of functions – project management, procurement, logistics, innovation. We’ve been showing visual mock-ups of three systems we think might help them to buy more of what they need from local manufacturers, and using this process to also get insights into their current procurement processes. The interview materials we used are published here .
We’ve conducted 22 interviews so far, and these are some of the patterns we’ve noticed in the responses:
Concept 1: Manufacturer Map
This is basically a system that shows where manufacturers are on a map, with some additional structured information about them. Broadly speaking, the people we’ve interviewed from procurement functions didn’t like this idea, and people in operational roles (project management, supply chain & logistics) did like it.
A common view from procurement was that this system wasn’t useful because they want to buy things ready made, rather than to commission things to be made for them. Several also made the point that they don’t have people who know about manufacturing processes. The operational people saw benefit in the idea because they tend to have experience of situations where standard items from a catalogue either don’t do the job, or just take too long to arrive. And it seems that the further the operations are from a supply hub, the more they have these problems.
So our preliminary conclusion on this type of system is that it may not be particularly useful at headquarters / regional headquarters level, but is potentially very useful in more remote or less developed areas, particularly those which have a large humanitarian response far from an industrial base. Northern Uganda, currently dealing with huge numbers of refugees from South Sudan, and Somalia have been suggested as places where this type of system would be extremely useful.
An interesting discovery we made was the Kenyan startup Lynk, who have created a platform to help people engage the services of informal sector workers. There is a lot of crossover with small-scale bespoke manufacturing, and their platform already includes carpenters. They are thinking about how to add furniture making in future.
This system would provide information on availability of specific items, so it would be far more difficult to keep up to date and reliable, and would have to be extremely useful to justify the effort involved. Interestingly, the reaction to this was to some degree reversed – all the procurement people we spoke to said it would be useful, and fewer of the operational people did. Some operations people also liked it, saying they could use it to give them info to prepare budgets – but then not actually buy from it. The procurement people liked the fact that it provides them with all the information needed to evaluate different options, and to justify their decisions. But a preference was expressed for imported products even when a local alternative was cheaper and available faster – because of a perception of quality. This re-emphasizes our earlier finding that establishing quality and the perception of quality are fundamental requirements for local manufacturing to make significant inroads.
The preliminary conclusion here is that this type of system may not really drive a shift to purchasing from local manufacturers. As the data maintenance requirements are so high, we are not currently convinced that this type of system is worth the investment. In any case, it could not be done without first knowing where the manufacturers are, so Concept 1 is a precursor to this and we will focus on that for now.
Concept 3: Transaction Engine
This is a different type of system, not focussed on making a connection to enable a contract but on managing the execution of the contract itself. It is a blockchain-based smart contracting system with two key sets of features: those that enable a one-to-many relationship so one buyer can manage a single contract with multiple small manufacturers, and those that enable quality checks on the output of each manufacturer and disburse payment based on results.
Similar to Concept 1, this system was less popular with people in procurement than with those in more operational roles. Procurement people raised concerns over how it would fit in with current working practices. In particular the requirement for the buyer to pay at least some of the contract value up front when placing the order was thought to be a barrier by some. We do think it is essential for quality that small-scale manufacturers have enough working capital to buy the right raw materials, but there are other ways to achieve that including by working with commercial finance providers or possibly microlending institutions.
Of the people we have spoken to so far, interest has been expressed in the contract distribution features but in quite an abstract way – it is solving a problem that has not yet really been felt by most organisations in the sector, because they are not dealing with many small manufacturers at present. On the other hand, we were surprised by the degree of enthusiasm for the quality management features. In fact we have been told that because of the quality control functionality built in, the system would be useful even to manage a contract with one supplier “This would have saved us millions if we had been able to use this in Somalia” was the reaction of one project manager.
So overall, the message we have received so far is that each of these systems could be useful in certain contexts. Concept 2, the marketplace option, could not be implemented without first mapping the manufacturers, so the first step for both of those options is to select a situation where it seems the mapping would be most useful by itself, and do that for a particular area. This would allow us both to learn from the process of doing the mapping and also then test what impact making the info available has on the purchasing habits of aid agencies operating in that area – from which we will be able to work out whether there is a business case to continue. For Concept 3, the transaction engine, the most important next step is to test the system’s impact on quality, and we are currently applying for funding to run a trial of that.
One comment that stood out for me was a tale about signposts for aid projects in South Sudan. Every project had its own signboard, saying which agency had implemented it and who the donors were. There were two types – rough, painted signs that had been made locally, and fancy imported ones. I was told that what drove organisations to choose one type versus the other type was not function or cost but rather what image they thought the project’s funders wanted to convey. Although image may not always be as central to a purchasing decision as it is with a sign, I think the key message still holds. This and several other conversations I’ve had through this project have made me realise how essential donor education will be in supporting the shift to local manufacturing. When donors tell the projects they fund that they want to know they are supporting local industry, it will happen.
There is still time to take part! If you would be willing for us to interview you or if you know someone who you think would be interested, please let us know by e-mailing connect (at) makernet (dot) global