Training

Five tracks, 9 sessions, about 9 hours in total if a whole organisation is trained at once. Each session is run against the deployed product, on the devices the team will actually use.

Training material and documentation are the same material Every session below points into a published role guide rather than restating it. That is deliberate: separately maintained training decks and documentation always drift, the deck is always newer, and the trainee always learns the deck. Here, deleting or renaming a documented section breaks the training build.

What a session is

Not a presentation. Each one names the sections it reads, a task the trainee does with their own hands, and what competence looks like at the end — because “we trained them” is not a fact until somebody can say what the trainee can now do.

Field enumerator

105 min across 2 sessions

People who will collect responses, on a phone, often with no connection.

Two sessions, on the devices the team will actually use. Not a classroom.

Before you start: An account, invited and assigned to a project. The device the enumerator will work on, charged.

1Your day, from sign-in to a submitted interview

60 min

Reads: Before your first day · Signing in · Your workspace

They do: Complete one full interview in the demonstration project, from opening an assignment to submitting it.
Competent when: The trainee completes an interview without prompting, and can say aloud when consent was taken and what it covered.

2Working with no signal, and what you are holding

45 min

Reads: Working with no connection · Consent, privacy and what you are collecting

They do: Put the device in flight mode, complete two interviews, restore the connection and watch the queue drain.
Competent when: The trainee can explain why signing in before leaving signal matters, and what a queued interview is.

Open the Enumerator guide

Organization administrator

120 min across 2 sessions

The person who will run the organisation inside the platform.

Two sessions against the organisation’s own structure, not a demonstration one.

Before you start: The organisation provisioned. A list of the people to invite and what each will do.

1People, roles and what each can reach

75 min

Reads: What you see, and why · Adding people · Narrowing what someone can see

They do: Invite one person, assign a role, and predict out loud what they will and will not see before signing in as them.
Competent when: The trainee’s prediction matches what the invited account can actually reach.

2Oversight without taking someone else’s job

45 min

Reads: Watching collection without doing the M&E job · Removing someone

They do: Remove a test account, then confirm what happened to the work it had already submitted.
Competent when: The trainee can state what removal does to a person’s access and what it does not do to their data.

Open the Organization Admin guide

M&E officer

135 min across 2 sessions

The person between raw collection and anything anybody would act on.

Two sessions with a dataset that is genuinely in progress.

Before you start: A project with responses already collected.

1Monitoring collection and data quality

75 min

Reads: Watching the field · Reviewing quality

They do: Find the assignment furthest behind, then find one response that needs review and act on it.
Competent when: The trainee can state what a review decision changes downstream and what it leaves alone.

2Evidence, and why the platform refuses things

60 min

Reads: Evidence · Handing off to analysis

They do: Attempt an export on a finding whose evidence does not resolve, and read the refusal.
Competent when: The trainee can explain the refusal to a colleague without describing it as a bug.

Open the M&E Officer guide

Data analyst

150 min across 2 sessions

People producing results that will be challenged.

Two sessions, then a supported first analysis on real data.

Before you start: A reviewed dataset. The study’s sampling design, whatever it is.

1What the numbers are allowed to say

90 min

Reads: The analysis workspace · What the platform will compute · What it will refuse, and why that is the feature

They do: Run one crosstab on a non-probability sample and read what the platform reports about representativeness.
Competent when: The trainee can say why post-stratification is unavailable for that dataset without looking it up.

2From a result to something publishable

60 min

Reads: Connecting results to evidence · Getting results out

They do: Attach one result to the evidence beneath it, then export it.
Competent when: The trainee can walk a figure back to a response in front of somebody else.

Open the Data Analyst guide

Executive and decision-makers

45 min across 1 session

People who will read the outputs and record decisions against them.

One short session. Their time is the constraint, and the product should not need more of it.

1Reading a finding and recording a decision

45 min

Reads: What reaches you · Management response · The decision record

They do: Open one published finding, walk it back to a response, then record a decision against it.
Competent when: The trainee checks the weighting state and n before quoting a figure, unprompted.

Open the Executive and decisions guide

What we do not claim about training

Stated precisely

There is no certification attached to any of this. Completing a track means somebody watched a trainee do the task; it is not a credential and we will not describe it as one.
These materials are in English. The interface runs in English, Kiswahili, French, Portuguese and Arabic; the guides are not translated, and we would rather say so than machine-translate instructions somebody follows in a village.
Training is run against a demonstration organisation with synthetic data unless your own project is already collecting, in which case it runs against yours.

All documentationHelp Center Where this fitsImplementation approach After trainingEnterprise support