Ubuntu Advisory Group Send a question Send one
Send a question

About

A market research team, founded on Ubuntu.

We are undergraduates at Williams College who research African markets for the companies operating in them. Ubuntu is the idea that none of us gets there alone, and it is the arrangement rather than the decoration: a company brings a question it cannot answer, students answer it with the methods they are being taught, and both sides end up with something they could not have had on their own.

Bring us a question Join the desk
  • Williams College undergraduates

The gap

The number a decision turns on has often never been written down.

Not hidden, not proprietary, not behind a paywall. Not written down by anybody, in any language, ever.

Fewer than

10%

of African startups reach a Series A.

Source Condia, State of Startup Funding in Africa, in Forbes Africa, May 2026

Capital is the part of that everybody names, and it is real. What gets named less is that a founder raising in Accra or Nairobi is asked for evidence that does not exist yet. How many kiosks are on this corridor. What a clinic actually pays for a test. Which regulator decides, and how long it took the last company. In a market with good data those are a morning's work. Here nobody has published them, so the founder either guesses in the deck or spends a month of runway finding out.

That is the gap this desk was built for. It is unglamorous, it is slow, and it is the reason the group exists: the work is boring enough that it does not get done, and important enough that the decision waits on it.

What we do

Classroom method, aimed at a question that has no answer key.

Nobody here is claiming twenty years in the sector. The claim is narrower and it is checkable: these are the research methods we are being taught, applied in the open, with the working shown.

Sourcing
The registers, gazettes, regulator circulars and statistics offices where an answer would be filed if it were filed anywhere. Read in the language they were published in, and dated, because a licensing rule can change by circular without notice.
Grading
Every source in a brief carries an A, a B or a C, and the grade is on the page beside the figure it supports. A reader can therefore disagree with one number without discarding the brief.
Two routes
A second researcher reaches the central figure without being shown how the first one got there. Where the two disagree, the disagreement is printed rather than settled quietly.
Operator calls
The third source is somebody’s memory and it is reached by telephone. Route salespeople, depot managers, clinic administrators and the consultants who took a company through licensing know the collection rate rather than the published one.
What we could not establish
The last page of every brief. The questions that stayed open, the people who did not call back, and the figures nobody would stand behind. It is the page that makes the rest of it worth reading.

A problem set has an answer at the back. A question from a company in Kumasi does not, and the person who called the depot manager is the person who has to defend the number. That is the part of this that cannot be taught in a seminar, and it is the part the students are actually here for.

The six stages, in order

The desk

There is no bench behind this desk.

No analyst pool, no research assistants, nobody senior on call from a larger firm.

A question that arrives in September is worked by the people who are on the desk in September, and the person who made the call to a distributor is the person who has to defend the number in the brief. Everybody is named on it, next to the part of it they reached, so a figure that turns out to be wrong has an author.

  • Lia Navon, of the desk
    Lia Navon Board Member
  • Arielle Levy, of the desk
    Arielle Levy Board Member
  • David LaPrade, of the desk
    David LaPrade Board Member
  • Nimoh Acheampong, of the desk
    Nimoh Acheampong Board Member
  • Asa Silverman, of the desk
    Asa Silverman Research Analyst
  • Fisher Stites, of the desk
    Fisher Stites Research Analyst
  • David DeRoo-Kury, of the desk
    David DeRoo-Kury Research Analyst
  • Milhan Sarfraz, of the desk
    Milhan Sarfraz Research Analyst
  • Zak Faris, of the desk
    Zak Faris Research Analyst
  • Joseph Jeon, of the desk
    Joseph Jeon Research Analyst
  • Santiago Paul, of the desk
    Santiago Paul Research Analyst
  • Jamie Evarts, of the desk
    Jamie Evarts Research Analyst
  • Sagata Das, of the desk
    Sagata Das Research Analyst
  • Alex Bethencourt Bohoyo, of the desk
    Alex Bethencourt Bohoyo Research Analyst
  • Jack Cheng, of the desk
    Jack Cheng Research Analyst
  • Ethan Miller, of the desk
    Ethan Miller Research Analyst
  • Henry Faulkner, of the desk
    Henry Faulkner Research Analyst
  • Declan Nayak, of the desk
    Declan Nayak Research Analyst
  • Femi Anthony, of the desk
    Femi Anthony Research Analyst
  • Nico Orazi, of the desk
    Nico Orazi Research Analyst
  • Vincent Kim, of the desk
    Vincent Kim Research Analyst
  • Arye Unger, of the desk
    Arye Unger Research Analyst
  • Christine Pak, of the desk
    Christine Pak Research Analyst
  • Hunter Justice, of the desk
    Hunter Justice Research Analyst
  • Odeng'el Ikwang'a, of the desk
    Odeng'el Ikwang'a Research Analyst
  • Henry Gold, of the desk
    Henry Gold Project Manager
  • Henry Hilton, of the desk
    Henry Hilton Research Analyst
  • Ella Densen, of the desk
    Ella Densen Project Manager
  • Clara Son, of the desk
    Clara Son Board Member
  • William Shepardson, of the desk
    William Shepardson Research Analyst
  • Divya Sijwali, of the desk
    Divya Sijwali Research Analyst
  • Aydin Aly Ajani, of the desk
    Aydin Aly Ajani Research Analyst
  • Zach Borzio, of the desk
    Zach Borzio Board Member
  • Nathan Zekarias, of the desk
    Nathan Zekarias Research Analyst

How somebody joins the desk

How a project is staffed

Two roles, and both of them do research.

A Project Leader and four or five researchers per project, with several projects running in a semester, so a question is scoped to a team rather than queued behind the others.

Project Leader
Scopes the question with the company, allocates the research, checks the evidence as it comes in, and writes the brief. Also makes the calls. Nobody reports to them.
Project team
Four or five per project. Each takes a piece of the question outright: a market, a regulator, a competitor set, a round of operator calls. They source it, grade it, and write the finding notes the brief is built from.

A first year sources, grades and writes finding notes. By third year the same person is holding a sector and training whoever arrives after them. Two roles staff a project, and there is no third.

Governance

Registered with the College. Reviewed by nobody in it.

This is the part a fund has to be able to repeat to a partner, so it is set out in full rather than described.

Status
A registered student organisation at Williams College.
Faculty advisor
College policy requires one of every registered organisation, so the group has one. The advisor does not read briefs.
Review
Nobody outside the student team reads a brief before the company that asked for it does. A fund that made the introduction is not an exception. It reads a brief when the company sends it, and not before.
Constitution
Written. It sets out the offices that run the group and what it was founded to do.
Terms
The research is free for the company. Nobody on the desk is paid.
The College
The group speaks for itself. Nothing on this site is a statement by Williams College.

That is the whole of it. There is no committee above the desk, no department signing anything off, and no outside firm standing behind the students. Registration means the College recognises the group and applies its policies to it. It does not mean anybody in the building reads the research.

So there is no signature further up the building to point at, and the group is answerable for its own work. The check that does exist sits inside the team: a second researcher has to reach the central figure without being shown how the first one got there, and where the two routes disagree, the disagreement goes in the brief.

The six stages and the grading scale

After the desk

Where people went next.

Firms and programmes, never a person.

Eleven organisations. Each one employed or admitted somebody who came through the desk, and those eleven are every one of them. None is connected to Ubuntu Advisory, none endorses it, and none of them knows this page exists. It is not a placement record and there is no pipeline behind it: where somebody went afterwards was their own doing.

Where the work is done

The work is done out of the public record, by phone.

You will ask how a desk in Massachusetts answers a question about a market in Dakar. It is the right question, and the answer is that the number is not sitting in Dakar either. It is in a register, a gazette or somebody's memory, and two of those three are public.

Most of what a decision turns on has already been filed and never read. A certified extract from the commercial register in Egypt is what a counterparty there asks for before it signs anything. One BCEAO list covers the licensed electronic money institutions of eight West African countries at once, so the question is not who licenses in Dakar or Abidjan but whether a company is on it at all. Tanzania's regulator publishes indicative fertiliser prices year by year, in Swahili. ARPTC, the telecoms regulator in Kinshasa, has published quarterly counts of mobile subscribers and operator revenue for years, and in a country where distribution and payment both run over those networks that series is the closest thing to a published market size. The Namibian Agronomic Board posts a monthly notice closing specific products to import when local supply covers demand.

Somebody still has to open the document, read it in the language it was written in, and find out what has replaced it since. The third source, the memory, is reached by telephone, which is why the mapping stage of a project ends in a list of ten people to ring. All of that is hours. None of it is a plane ticket. In a market where nobody has published the answer, that is roughly how it would be done from Lagos too.

What a telephone will not do is put somebody in a warehouse. Questions that need a person standing in the market are the ones the desk turns down, and where a question makes that plain the no comes back in the reply rather than after a call.

Where the desk works, sector by sector

Next

Bring the question, or come and work on one.

A company gets a Project Leader, four or five researchers and a brief with every source graded. A student gets a live question with no answer key, and is grading sources on it inside a semester.

Scope a question Join the desk