
A resource-flow map of one cohort is a dated record of everything that reached its founders or left them during the program year: cash, credit, goods, introductions, customers, and money sent home. Each entry names the hand it passed through and what it became. We know of no published map of this kind for a Kampala cohort, and AVODA has never kept one either. What follows is a method. It joins two real studies, a network map of one St. Louis grant cohort and a year of weekly cash-flow visits to Kenyan small firms, and adjusts both for Kampala.
Key Takeaways
- A cohort flow map records each resource that reached or left a cohort’s founders in a program year, with its date, source, carrier and outcome.
- Researchers mapped the ties around the twenty 2013 Arch Grants winners in St. Louis. Five firms without a tie to a fellow winner each connected to the ecosystem through a support organization.
- Kenya’s Small Firm Diaries visited small firms weekly for a year. About 47 percent used supplier or customer credit, a share the authors call a lower bound.
- In Uganda, 92.8 percent of mobile money transactions in the year to June 2023 were UGX 50,000 or less, so a Kampala map has to capture many small amounts.
- Being surveyed can change behavior, so the diary is part of the program and should be disclosed.
- The Cohort Flow Diary sets the boundary first, keeps two records in different hands, reconciles them monthly and reads them three times a year.
What the book asks a cohort record to hold
Entrepreneurial Ecosystems: Resource Acquisition, Flow and Management (Springer, 2025), edited by Parisa Baig and four colleagues, names a gap in method in its introduction. Studies of how ecosystems change over time are scarce, and of the few that exist, “most rely on retrospective historical data rather than real-time longitudinal observations.” A program meets its founders every week, closer to real time than any researcher, and usually reports on them once, at the end, in totals.
Two chapters say what a real-time record has to catch. João Almeida and Ana Daniel studied three rural ecosystems in Portugal through 36 interviews and six site visits between 2019 and 2024, and they argue that an ecosystem runs past its geographic boundary. The editors turn their agenda into a question every mapper must answer: “which actors are on the other side of the boundary.”
Melissa Bejjani-Cazaré interviewed 14 people in Jena, Germany, in 2024 and found coordination failing amid plenty. Founders and support staff struggled to find the right contact person. One director put the referral problem plainly: “I would like to see more startups from all players being directed to us.”
Between them, the passages ask a cohort map to record flows as they happen and to follow each one across the city’s edge, referrals included.
Twenty grant winners in St. Louis, drawn as a network
In 2013 Arch Grants, a St. Louis nonprofit, gave each of twenty early-stage companies $50,000 in equity-free cash, on condition that they locate in the city for a year. Yasuyuki Motoyama of the Kauffman Foundation and Karren Watkins of Washington University began interviewing the recipients in late 2013 and reached sixteen of them. They then interviewed the support organizations those sixteen named and drew the ties as network maps, published in September 2014.
Twelve of the sixteen described active relationships with other recipients. Five companies on the peer map lacked any tie to another company, yet each connected to the ecosystem through at least one support organization, so a map of peers alone would have filed them as isolated. Four local mentors with founding experience served six of the companies between them. One founder traced a path in a single sentence: “the Arch Grants got us to ITEN, which got us to a business journal, which put us in touch with some of our client contacts.”
That sentence is a resource flow written as a chain of hands, and the study’s design is the first half of a cohort map: start from a defined cohort, ask each founder who helped, then interview the helpers. The authors called their work “exploratory and descriptive.” Its maps show who is tied to whom, without amounts, dates or what each introduction became, and they rest on one retrospective round of interviews. Whole-ecosystem maps are weighed against other records in the trust density essay; the St. Louis study is narrower, and closer to what a program can run.
A year of weekly visits to Kenyan small firms
The Small Firm Diaries, led by the Financial Access Initiative at New York University, sent field researchers to visit small firms weekly for a year and record their cash flows. It ran in seven countries from 2021 to 2023, Uganda among them; the country report we found is the Kenyan one, funded by FSD Kenya.
The Kenyan team recruited 166 firms in low-income neighborhoods of Nairobi, Kisumu and Kwale and collected data from November 2021 to November 2022. Eleven dropped out. The analysts set aside 13 more whose owners, in the field researchers’ judgment, were holding back complete and reliable cash-flow information, and reported on 142.
Their working habits carry straight over to a cohort diary. Owners bundled several days of transactions together and were often unsure of exact dates, so the team summarized by month. They dropped the first two months of data in cleaning. And they asked about separating business and household money at intake, before the diary could prompt the practice it was measuring.
The findings show who allocates around a small firm. About 47 percent of firms used supply chain finance, which the authors define to include in-kind transfers and treat as a lower bound. Firms outside services gave more credit to customers than they received from suppliers, and most owners started on their own savings. The diaries capture the amounts and timing the St. Louis maps lacked, but only inside the firm: a supplier’s credit appears, and the introduction that produced the supplier stays invisible.
Ninety-three in a hundred transfers under fifty thousand shillings
Three conditions in Kampala change how the two methods combine. The first is size. The Bank of Uganda’s annual report for the year to June 2023 found that 92.8 percent of mobile money transactions were for UGX 50,000 or less. Each St. Louis winner received one $50,000 award; resources around a Kampala founder are likely to arrive as many transfers small enough to forget by the next call. A ties-only map would miss them, and a diary demanding exact amounts would wear the founder out. Size bands, totaled by fortnight, sit between.
The second is the hands. The Kenyan diaries found suppliers, customers and owners’ savings carrying much of the load. A Kampala cohort’s suppliers, savings groups and customers are allocators in their own right, and a map that lists only programs, lenders and mentors will look thin for the wrong reason.
The third is the edge. Bank of Uganda figures reported in April 2026 put 61 percent of inbound remittances through mobile money, onto the same phones many founders run their businesses from. Money also leaves for school fees and relatives in the home district, a pull set out in Household Economics as a Startup Constraint. Almeida and Daniel’s boundary point comes from rural Portugal, so it needs a Ugandan check. Inbound, it passes: remittances cross the border by definition, and most land on mobile money. Outbound is untested, since we found no Ugandan figure for what founders send home, so the map must measure it.
The book’s Jena finding does not transfer to Kampala. Jena’s interviewees described resource abundance, an overwhelming offer of events and funding; Uganda’s mobile money figures point to many small amounts, so a Kampala map should look for gaps in a thinner stock. The coordination finding does transfer: a founder sent to the wrong door loses time either way.
Surveys that change the surveyed
Alix Peterson Zwane of the Bill & Melinda Gates Foundation, Jonathan Zinman and twelve co-authors, Esther Duflo and Michael Kremer among them, ran five field experiments in which they randomly chose who would be surveyed and then measured later behavior from records other than the subjects’ own reports. In three health studies, being surveyed raised the use of water treatment products and the take-up of medical insurance, and frequent surveys about diarrhea biased the estimated effect of cleaner water. Their conclusion reaches a cohort diary as surely as a health survey: researchers “should reconsider whether, how, and how much to survey their subjects.”
A fortnightly call asking a founder what arrived is a prompt, and the diary is part of the program. A program that maps its cohort cannot then present that cohort as an untouched benchmark, and it owes founders and funders a statement that the record is itself an input.
Two findings narrow the objection. In the same paper’s two microlending studies, being surveyed had no effect on borrowing, the behavior closest to a founder’s credit. And the Kenyan diarists guarded against the effect by asking about practices at intake, before any cash flows were logged. A cohort map can do the same, and keep each call short enough that it records more than it coaches.
The Cohort Flow Diary
The Cohort Flow Diary is a year-long record of one cohort, kept in two hands and read three times. It extends the four quarterly measures in Resource Flow: Why Counting Programs Misses What Reaches Founders to flows out, flows between founders, and flows across the city’s edge.
It starts with a boundary drawn before the first entry: three lists, dated. The cohort is everyone admitted, including founders who stop attending, because a map of finishers flatters the program. The resources are cash; credit from lenders, suppliers and savings groups; goods in kind; introductions; customers; and money that left the business. The edge names the outside flows to record: family in the home district, diaspora transfers, suppliers in another town.
Then come two records in different hands. The founder’s record comes from a fortnightly call or visit by someone outside her mentoring and grant relationships, asking what arrived, from whom, through whom, and what left. The program’s record is its own dated log of what it sent: each introduction, seat, disbursement and referral.
Each month the two are reconciled. An entry in the program’s log that is missing from the founder’s record is a dead end: an introduction that went nowhere, or money still waiting. An entry in the founder’s record absent from every program log is the ecosystem working on its own, as St. Louis support organizations did for the five firms without peer ties. Mark entries the recorder doubts, as the Kenyan team did, and keep them visible.
At three, six and twelve months someone reads the map. Four readings carry the most information: how many introductions pass through one person, whether flows out to customers and family outweigh flows in, how long promised resources take to arrive, and which founders have empty pages. Treat the first month or two as settling. Return each founder’s pages to her before anyone else reads them, the practice set out in Measuring Inclusion Without Turning Founders Into a Diversity Statistic.
Each diary row has seven fields. The row below holds placeholders only:
| Date | From | Through | To | What | Size band | Became |
|---|---|---|---|---|---|---|
| [fortnight] | [supplier, lender, relative, program, peer] | [person who carried it] | [founder, or her customer] | [credit, cash, goods, introduction] | [a band you set, such as up to UGX 50,000] | [stock, wages, school fees, a second introduction] |
The “Became” column is filled at the next call.
Ithamar’s account of the tabernacle silver
When the tabernacle was finished, Exodus gives its materials a separate account, “recorded at Moses’ command by the Levites under the direction of Ithamar.” Bezalel, who made everything, appears in a parenthesis. The account runs from source to use. The silver came from the census at half a shekel a head, and its 100 talents became 100 bases, one talent each. Even the remainder is followed: 1,775 shekels made the hooks, overlays and bands for the posts (Exodus 38:21-28). Two features of the method already sit in that chapter. The record is kept by someone other than the builder, and small contributions are traced to what they became instead of being folded into a total. Why keep such an account when the funder asks only for the total? The Levites had a reason. Every half shekel had come from a person counted in the census, and the account answered to the God the tabernacle was built for, so the hooks were owed the same care as the gold.
Frequently Asked Questions
What is a cohort resource-flow map?
It is a dated record of every resource that reached or left a cohort’s founders during the program year: cash, credit, goods, introductions and customers. Each entry names the source, the person it passed through, the receiving founder and what the resource became, following it past the program’s own books.
How is a flow map different from an ecosystem network map?
A network map shows who is tied to whom, usually from one round of interviews; the St. Louis study of twenty Arch Grants winners is a careful example. A flow map adds a size band, a date and what each resource became, and it is kept through the year instead of reconstructed afterward.
How often should founders be asked about resource flows?
Fortnightly suits most cohorts. The Kenyan Small Firm Diaries visited weekly and still found owners bundling several days of transactions together, so the team summarized by month. A fortnightly call using size bands instead of exact amounts catches most flows and stays short enough to avoid coaching the founder.
Does tracking a cohort change how founders behave?
It can. Zwane and colleagues found that being surveyed raised the take-up of health products in three field experiments, while borrowing in two microlending studies stayed unchanged. Treat the diary as part of the program, tell founders and funders so, and keep the mapped cohort out of any untouched comparison group.
Next Step
Write your cohort’s boundary down: who is in, which resources count, and which flows across the city’s edge you will record, then name a recorder who mentors none of the founders. Then set that boundary beside your funder’s reporting template on a strategy call with Joshua, so the monthly reconciliation feeds the report you already owe.
Related Reading
- The Accelerator That Learns From Its Own Cohorts
- Cohort Peer Effects: The Evidence for the Room
- Cash Flow When Everyone Pays Late: The 90-Day Economy
- Impact Measurement Grew Up: 60 Decibels and Verified Data
Sources and Evidence
- Baig, Montoya-Martinez, Belitski, Theodoraki and Godley, “Introduction,” in Entrepreneurial Ecosystems: Resource Acquisition, Flow and Management, Springer, 2025: scarcity of real-time longitudinal studies of ecosystems.
- Almeida and Daniel, “Mitigating Local Bottlenecks in Rural Entrepreneurial Ecosystems through Trans-Local Linkages,” chapter 6: three Portuguese rural ecosystems, 36 interviews, 2019 to 2024; flows across ecosystem boundaries.
- Bejjani-Cazaré, “Challenges and Opportunities: Resource Flows and Digitalization in Entrepreneurial Ecosystems,” chapter 11: 14 interviews in Jena, 2024; abundance, coordination failure, referrals.
- Baig et al. (eds.), Entrepreneurial Ecosystems: Resource Acquisition, Flow and Management, Springer, 2025: the volume tested here, including the editors’ summary of chapter 6’s research agenda in chapter 14.
- Motoyama and Watkins, Examining the Connections within the Startup Ecosystem: A Case Study of St. Louis, Kauffman Foundation, September 2014: the 2013 Arch Grants cohort, interviews with sixteen of twenty recipients and their support organizations, network maps, the ITEN path quote.
- Financial Access Initiative and FSD Kenya, Kenya Small Firm Diaries country report, 2023: weekly diary method, sample of 166 to 142 firms, bundled dates, cleaning rules, intake questions, supply chain finance as a lower bound.
- Business Times Uganda, “Mobile Money, Agent Banking Transactions Surge as EAC Regional Transactions Decline”: Bank of Uganda Annual Report 2022/23 figures, including 92.8 percent of mobile money transaction volume at UGX 50,000 or less.
- Nile Post via allAfrica, “Mobile Money Dominates Uganda’s Shs 9.6 Trillion Remittance Market,” 9 April 2026: Bank of Uganda figures putting 61 percent of inbound remittances through mobile money.
- Zwane et al., “Being Surveyed Can Change Later Behavior and Related Parameter Estimates,” PNAS 108(5), 2011, Europe PMC record: five field experiments on survey effects; the quoted conclusion.
- Exodus 38:21-31, NIV, Bible Gateway: the Levites’ account of the tabernacle materials under Ithamar.
