Method
How a question becomes a brief.
Six stages, in the same order every time, run by a Project Leader with four or five researchers. Before a figure leaves this desk, a second researcher has to reach it without being shown how the first one got there.
Six stages
Most of a project happens before anybody writes a sentence.
Stage three takes the longest. Stage two is the one that decides whether the question can be answered at all, because an answer has to exist somewhere before anybody can go and read it.
- 1
Scope
A twenty-minute call to pin down the decision, and what evidence would move it.
- 2
Map
Where an answer could exist at all. A company register, which is a conservatória in Luanda, the RCCM in Dakar and the Corporate Affairs Commission in Lagos, and none of the three works the same way. A regulator that publishes, or one that moves by circular. And which ten people to ring. Often the hardest step.
- 3
Research
The longest stage by far. Each researcher owns a piece of the question outright. Every source graded as it goes in.
- 4
Synthesise
The Project Leader separates what we know from what we are inferring.
- 5
Challenge
A second researcher has to reach the central figure without being shown how the first one got there. Where the two routes disagree, the disagreement goes in the brief.
- 6
Brief
The brief, the graded evidence log, and a page listing what we could not establish.
The two parts of this you can check for yourself are both below: the grade printed beside every source, and the second route to every central figure.
Evidence
A, B or C, printed next to the claim it supports.
A grade is a statement about where a claim came from. You can see which parts of an argument are solid and which are the best available guess, without taking our word for either.
- Grade A
- A primary or official source we read ourselves: a Bank Al-Maghrib circular, a Zambian PACRA entry, a customs record, a contract, an interview we conducted and named.
- Grade B
- Credible secondary reporting we could not verify independently but have reason to trust, such as established trade press or a research house that states its method.
- Grade C
- A single unverified claim we are carrying because nothing better exists. It appears in the brief flagged, and no conclusion rests on it alone.
The last page of every brief lists what we could not establish, and every Grade C claim carried in it.
An official source is not automatically an A. Tunisia's directory of industrial firms counts only active establishments with ten employees or more, and the Agence de Promotion de l'Industrie et de l'Innovation has said its own database does not reflect real closures. A competitor count built on it is a floor and not a census, and it is graded C where it appears.
The test runs the other way as well. Cameroon's National Cocoa and Coffee Board posts the price exporters are paying producers, day by day, next to its own list of active exporters. A figure read straight off that is an A, and it is the whole reason the grade is a statement about the source and not about the country it came from.
The check
Undergraduates wrote it. Someone else had to get the same number.
No professor reads a brief before you do. Before a figure goes in, a second researcher has to reach it without seeing the first route.
One may build it out of the weekly export tables Rwanda's National Agricultural Export Development Board publishes, commodity by commodity and week by week, the other out of what four exporters say they were paid. Where the two land apart, both routes go into the brief.
When the second researcher cannot get there at all, nobody splits the difference. The number comes out of the brief and onto the list of what we could not establish.
There is no senior reviewer above that stage and no signature at the end of it. Stage five is the thing standing between a wrong figure and the decision you are about to make with it.
For a fund, that is the stage an introduction rests on. You do not read the brief before the company you sent us does, so what stands behind your name is the second route and not a look at the draft.
Where we stand
The shortest rule on this list is the one about paperwork.
Only what the desk actually does is on this list. There is no corrections policy on it, because nothing has been published for anyone to correct.
- Confidentiality
- Only the named analysts on your project see your material. Nothing about your company is published unless you have approved it, and nothing is kept after the engagement without your written agreement.
- Paperwork
- Ask on the first call. We say what we can sign.
- Conflicts
- We will not work both sides of the same question. If a direct competitor of a current client asks us something, that client hears it from us before we take the work.
- Attribution
- The analysts are named on the brief, and so is the researcher who had to reach the figures independently. If a number is wrong, you know whose it was.
- Quality
- A brief ships when its central figure has been reached twice, not on the date we gave you. If those two conflict you get an email moving the date, and not a thinner brief on time.
Limits
Five kinds of question that come back as a no.
- Questions that need somebody standing in the market. We do not have that, and a phone call is not a substitute for it
- Questions that have to be answered before the end of the month. The desk is staffed by students on a college calendar: reading period stops the work and January starts it again, so the unit to plan in is months
- Competitive intelligence dressed as market research
- Diligence a fund is contractually required to perform itself
- Questions somebody has already answered in print. We will send you the link instead
A question that is plainly one of these comes back as a no in the reply, before any call. Where it takes a conversation to tell, the call is twenty minutes and a no can still come out of it. The reason arrives with the no either way.
The first call
Bring the decision you are stuck on.
Twenty minutes, and you find out whether the answer exists in a public record anywhere, and whether this is the desk to go and get it. The research is free.