The part of Tanzania’s ecosystem we haven’t learned to measure
Tanzania has been building its entrepreneurial ecosystem in earnest, and it shows. From 2020 to 2024, the number of known startups grew from 247 to 1,041. In 2024 alone, these startups created over 138,000 jobs, and foreign investment more than doubled from the previous year (Tanzania Startup Association, 2024). The annual Ecosystem Status Report from the Tanzania Startup Association is one of the most comprehensive datasets of its kind on the continent, and it does the job it was built to do. What we have far less of, and I include my own work here, is a systematic account of how the actors behind those numbers connect and collaborate. We know how many startups there are, in which sectors, at what stage, with how much funding behind them. We have counted the parts carefully. We have looked far less closely at what runs between them. That gap matters because of a specific problem. Startups in Tanzania are being created in growing numbers, but comparatively few of them convert early traction into durable commercial relationships: a signed pilot that becomes a renewed contract, a first corporate customer that becomes a reference for the next three, an investment that leads to a follow-on round. Call this the conversion problem. Ecosystem support is currently designed as though it is solved by more capital and stronger formal institutions. My working view, and the reason I am starting this research, is that a good deal of it sits in the relational layer instead, and that trust is the part of that layer most worth measuring first. Nobody has tested either account, and if the second has force, the first is quietly expensive: every programme cycle aimed at the wrong bottleneck is a year of founders not converting.
Why this is the moment to ask
Tanzania’s Development Vision 2050, launched in July 2025, sets out a twenty-five-year path towards a trillion-dollar economy, with private sector dynamism among its central pillars and science, technology, research and digital transformation among its strategic enablers (United Republic of Tanzania, 2025). The National Planning Commission has described the vision as placing the private sector at the centre of economic development, with government acting as facilitator, enabler and derisker. The first Medium-Term Development Plan under that vision, covering 2026 to 2030, is being finalised now. A vision of that scale succeeds or fails on whether firms can transact with one another at volume, and the plans being written this year will set how ecosystem support is sequenced for the next five. Derisking in particular is a relational instrument: it works by making a counterparty willing to commit to someone they would otherwise pass over. Designing it well means knowing where that hesitation currently sits and what causes it.
Why Tanzania, and not simply any young market
The argument that measurement tools miss relationships is general. What makes it bite in a particular place is the mechanism underneath it, and here I can offer a hypothesis rather than a finding. It comes from what I have observed working with this ecosystem since 2021, and every part of it needs testing. My working proposition is that verification is expensive and unreliable in this market, and that relational proximity is what people substitute for it. Four things appear to drive that. There is no shared picture of who is actually operating in the ecosystem, broken down by sector and by position in a supply chain, so a firm looking for a partner often cannot identify the candidates. Reputation is thinly signalled: for most young firms there is no track record a counterparty can look up. Contract enforcement through the courts is slow enough that, for deals of the size startups do, it is not a realistic remedy. And formal due diligence runs through lawyers, which puts its cost out of proportion to a first contract. Take those together, and the rational response is to transact with people you already know something about, or whom someone you trust can vouch for. If that is right, it explains why bonding connection is strong here and bridging connection is harder, and it locates the problem somewhere addressable. It is not a claim about anyone’s character or willingness to do business. It is a claim about the cost of finding out whether a stranger is worth doing business with. That cost is measurable, and so is what people do in response to it.
Counting the parts is not the same as seeing the connections
The main tools we use to read ecosystems, such as the World Bank’s Doing Business index, Isenberg’s Entrepreneurship Ecosystem Model and Stam’s Entrepreneurial Ecosystem Index, assess whether the essential elements are present: finance, talent, markets, policy, culture, support organisations. Networks appear in several of these frameworks, Stam’s included, but as a stock to be counted rather than as relationships whose quality and reach can be traced. Knowing that an ecosystem has networks is not the same as knowing who can actually work with whom, and on what basis. These frameworks were developed where much of the relational work is already handled by something else: contracts routinely enforced, credit histories visible, reputation a matter of public record. There, whether two parties can rely on one another is largely settled before they meet. The tools never had to measure relationships, because the institutional environment was doing it in the background.
The practical consequence is easy to state. Startups do not convert capital into growth on their own; they do it through partners, in supply chains, hiring, distribution and sales. A founder in Arusha with a working product and money in the bank can still spend a year failing to convert a single corporate buyer, not because the product is wrong or the capital is short, but because nobody in the room has a reason to take the risk on her. The obvious objection is a simpler explanation. Tanzania raised under $15 million in startup funding in 2025; on that basis, the conversion problem needs no relational account at all, because a thin capital base and a young market explain it. That explanation is real, and it is part of the picture. But it has never been tested against an alternative, because the alternative has never been measured. Resources are easier to count than connections, so resources are what we count, and an explanation that is easy to evidence crowds out one that is not. The two also compound: where few deals close, few relationships get built, and where few relationships exist, capital has nowhere to travel. Treating the first as the whole story is what keeps the second invisible.
The lens I use for this
If the standard tools cannot see relationships, the first thing needed is a vocabulary that can. The most useful one comes from social capital, and specifically the distinction Robert Putnam and James Coleman drew between two kinds of connection. Bonding social capital is the strong, trusting relationship within a group: a founder community where people already know one another. Bridging social capital spans different groups and sectors, linking people who do not yet share a network. An ecosystem can be rich in one and poor in the other while looking, in aggregate, entirely healthy.
The literature does not speak with one voice, and the disagreement shapes how the question gets asked. Macro-level work treats trust largely as a good in short supply: more of it, better outcomes. Welter (2012) complicates that directly, showing that strong ties within a group often arise as a practical response to the environment people operate in, and that those same ties can narrow the range of partners a venture will engage. A diagnostic assuming more trust is simply better would be unable to detect one of the more interesting things it might find: connections that are dense, functional and yet contained.
Scholarship grounded in African markets makes the same point with evidence. Fafchamps (2001), researching across Sub-Saharan Africa, showed the weight relational conditions carry where formal contract enforcement is still developing: personal relationships take on the role usually filled by formal mechanisms, serving as a form of assurance in business dealings. Relational ties are not a soft factor in that account. They are doing hard economic work, standing in for enforcement infrastructure that firms elsewhere take for granted. The implication for measurement is uncomfortable. If relationships are carrying enforcement, an instrument counting only formal provision will read a functioning arrangement as an absence. The instrument fails, not the ecosystem. Fafchamps and Welter together define the question: not how much trust there is, but which kind is doing the work and how far it reaches beyond the circles where it already exists. That is answerable from things that leave a trace: whether a pilot with a corporate buyer becomes a renewed contract, whether an investment relationship began through an acquaintance or a formal channel, whether a supplier relationship survives its first dispute. Each marks the point where a relationship either extends beyond an existing circle or does not. None of it requires asking anyone how much they trust anyone else, which is fortunate, since self-reported trust is among the least reliable things one can ask about. One scoping note: Khlystova, Kalyuzhnova and Belitski (2022), studying institutional trust across six cities in transition economies, found it mattered clearly in some and not at all in others within a single country. Sub-national variation is real, and a national picture assembled without it would be misleading.
What would change if this is right
Knowing which relationships are load-bearing and which are missing would let TSA and its partners sequence support differently: convening around the specific interfaces where deals stall rather than adding another general-purpose programme, and investing in the shared information and reputation infrastructure that lowers the cost of checking out a stranger. It would give funders a way to tell a relational bottleneck from a capital one before committing to a three-year instrument. And it would make visible the founders currently invisible to ecosystem support: those with a product and no route to the people who could buy it. The approach is not new to me. My research on Lombardy used foreign education as a visible trace of something otherwise hard to see, a proxy for whether a founder had ties reaching beyond the region. Across the funded innovative startups I studied, 93.6% had at least one founder who had studied abroad; in manufacturing that share fell to 12.5%, roughly half the rate in ICT. A single headline figure suggested a thoroughly connected region, while disaggregation showed a whole sector sitting outside those flows. Tanzania’s headline numbers may be concealing something similar.
The argument, plainly
I am not arguing that trust explains everything about Tanzania’s ecosystem. I am arguing that the relational layer is currently assumed rather than examined, that the assumptions being made are consequential, and that trust is the most tractable place to start testing them. If Fafchamps is right about contexts like this one, relationships are carrying weight that a component map records as absence. If Welter is right, an ecosystem can be densely connected and still constrained. Neither possibility is visible in the measures currently in use, and both change what the sensible next move is. This work will not produce straightforward answers, and it has to be willing to conclude that the constraints lie elsewhere. But the sequencing decisions being taken now will hold for years, and they are being taken without this. That is the opportunity, and the window for it is narrow.
References
Coleman, J.S. (1988). Social capital in the creation of human capital. American Journal of Sociology, 94(Supplement), S95–S120.
Fafchamps, M. (2001). Networks, communities and markets in Sub-Saharan Africa: Implications for firm growth and investment. Journal of African Economies, 10(Supplement 2), 109–142.
Khlystova, O., Kalyuzhnova, Y. and Belitski, M. (2022). Towards the regional aspects of institutional trust and entrepreneurial ecosystems. International Journal of Entrepreneurial Behaviour & Research.
Putnam, R.D. (1993). Making Democracy Work: Civic Traditions in Modern Italy. Princeton: Princeton University Press.
Tanzania Startup Association (2024). Tanzania Startup Ecosystem Status Report 2024. Dar es Salaam.
United Republic of Tanzania (2025). Tanzania Development Vision 2050 (Dira 2050). Dodoma: President’s Office, Planning Commission. Available at: https://www.planning.go.tz
Welter, F. (2012). All you need is trust? A critical review of the trust and entrepreneurship literature. International Small Business Journal, 30(3), 193–212.
