Study types

What you can commission us to run.

Research is commissioned under names like baseline survey, KAP study, facility assessment or focus group discussion — not under the name of the design underneath. This page lists the study types VoiceInsights runs, what each one produces, which collection routes suit it, and what it cannot do. Every entry states its own limits, because the limits are the part you plan around.

How to read this page

Two words carry most of the weight, and they promise different things. A third scale describes the collection routes.

Proven

VoiceInsights runs this study type end to end, and a repeatable proof of that run exists in the platform’s own release evidence.

Supported

VoiceInsights runs this study type on the platform’s shared research structures. It works, and there are limits stated on the row that you plan the study around.

How to read the routes

Every study type rates each collection route on the same four-point scale, because “available on every channel” is not true of any research design worth running. The ratings are our recommendation for the design, not a restriction the platform enforces — and they can only narrow the live routes, never add one.

Primary

The route this study type is normally run on.

Supported

A sound choice for this study type.

Bounded

Workable within a stated bound — usually instrument length or the forms an answer can take on that route.

Not suitable

We do not recommend this route for this study type, and would talk you out of it.

A study type is not a research mode

These are different axes, and confusing them is how a platform ends up claiming thirty products when it has seven designs. A study type is what the work is called when it is commissioned. A research mode is the design the platform runs to deliver it. A study phase is when it happens in a programme. A collection route is how an answer travels.

So baseline, midline and endline are phases of one study, not three engines: the same instrument, the same indicator identity, a new wave and a new dataset version each time. Run them here and the comparison stays inside one governed workflow, with the change reported as a value, a later value, and the difference between them. Whether a programme caused that change is a question of evaluation design, and we will not answer it for you automatically.

Key informant interviews, in-depth interviews and focus group discussions are formats within qualitative research, not modes of their own. They are listed on this page because that is what people search for, and each one names the mode that carries it.

The seven research modes The five collection routes

Browse by what you are commissioning

Eight groups. Jump to the one that matches the brief on your desk.

Every study type we run

Each entry names the research mode that executes it, the phase where a phase applies, the forms of evidence it collects, what it produces, and how each collection route rates for it.

Surveys

A defined questionnaire put to a defined population, reported as numbers with a stated method.

Household survey

Supported

Measure living conditions, access, income, food security or programme reach across a population, and defend the figure.

What you should know

One record per unit on every collection surface. Member-level rosters — a household enumerated person by person — are held relationally by the platform, and there is no roster control on the field app or the web link, so a roster instrument is configured with your implementation team rather than built and fielded unaided.
Skip logic is one condition on one earlier question — “if they said no, do not ask the follow-ups” — evaluated by the server as well as the device. Multi-condition expressions, arithmetic across questions and cross-repeat references are refused when the instrument is imported rather than accepted and ignored.

Research mode that runs it: Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Written · Structured · Spoken
Typical outputs: Dataset and codebook · Analytical workbook · Research report · Indicator dashboard
Who commissions it: NGO and INGO programmes · Government statistics · Research and evaluation firms · Donors

Which routes suit it

Phone Call Bounded
Workable for a short module; a long instrument is a long call.
WhatsApp Bounded
Workable for a short module.
SMS Not suitable
A household instrument is too long to run one message at a time.
Web link Supported
Offline app Primary

Population survey

Supported

Estimate a characteristic across a defined population with a stated sampling design and a stated margin of error.

What you should know

One record per unit on every collection surface. Member-level rosters — a household enumerated person by person — are held relationally by the platform, and there is no roster control on the field app or the web link, so a roster instrument is configured with your implementation team rather than built and fielded unaided.
The sampling frame is yours. The platform records the frame, the design, the strata and the achieved sample, and computes the target size — it does not supply a population register.

Research mode that runs it: Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Written · Structured · Spoken
Typical outputs: Dataset and codebook · Statistical report · Analytical workbook
Who commissions it: Government statistics · Research firms · Donors

Which routes suit it

Phone Call Bounded
Suits a short instrument on a reachable frame.
WhatsApp Bounded
Suits a short instrument.
SMS Not suitable
Instrument length and the single answer form make this unsuitable.
Web link Supported
Offline app Primary

KAP survey

Proven

Measure knowledge, attitudes and practices on a health or behaviour topic, usually to target and then re-measure a campaign.

Research mode that runs it: Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Written · Structured · Spoken
Typical outputs: Research report · Indicator dashboard · Dataset and codebook · Presentation
Who commissions it: Public health · NGO and INGO programmes · Ministries of health
How this is proven: An end-to-end proof of this study type runs against the deployed platform on every release, and the release gate refuses the build if it is missing.

Which routes suit it

Phone Call Supported
WhatsApp Supported
SMS Bounded
Workable for a short closed-question battery; not for open practice questions.
Web link Supported
Offline app Primary

Agriculture & farmer survey

Supported

Measure production, inputs, market access, prices or extension reach at farm or farm-household level, usually by season.

What you should know

One record per unit on every collection surface. Member-level rosters — a household enumerated person by person — are held relationally by the platform, and there is no roster control on the field app or the web link, so a roster instrument is configured with your implementation team rather than built and fielded unaided.
Photographs, GPS points and file attachments are not question types a respondent or enumerator answers. Location is recorded about a submission, not asked as a question.
Quantities are collected as numbers. The unit is carried in the question wording — there is no per-question unit register, so an instrument must fix the unit for the respondent rather than let them choose one.

Research mode that runs it: Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Written · Structured · Spoken
Typical outputs: Dataset and codebook · Analytical workbook · Research report
Who commissions it: Agriculture ministries · Agri-development programmes · Value-chain and market-systems projects

Which routes suit it

Phone Call Bounded
Suits a short seasonal round.
WhatsApp Bounded
Suits a short seasonal round.
SMS Not suitable
Quantity-heavy instruments do not survive one message at a time.
Web link Supported
Offline app Primary

Assessments

A structured judgement about a place or a service — a facility, a school, a water point, a district.

Facility assessment

Proven

Assess what a service point has and does — staffing, stock, equipment, availability — against a defined checklist.

Research mode that runs it: Facility assessment. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Structured · Written · Observed
Typical outputs: Assessment report · Indicator dashboard · Analytical workbook · Dataset
Who commissions it: Health systems · Education · WASH · Government service delivery
How this is proven: An end-to-end proof of this study type runs against the deployed platform on every release, and the release gate refuses the build if it is missing.

Which routes suit it

Phone Call Not suitable
An assessor is standing in the facility. A call is the wrong instrument.
WhatsApp Not suitable
An assessor is standing in the facility.
SMS Not suitable
An assessor is standing in the facility.
Web link Supported
Offline app Primary

Health facility assessment

Supported

Measure service availability and readiness at health facilities — the standard input to a district or national health-system review.

What you should know

Photographs, GPS points and file attachments are not question types a respondent or enumerator answers. Location is recorded about a submission, not asked as a question.
Facility-level and respondent-level evidence are held apart, and analysis runs at the level the study declares. A design that needs both in one model — facility readiness explaining patient experience — is configured with our team.
Nothing in the platform produces a clinical or diagnostic conclusion. Health findings are descriptive statistics over what was collected.

Research mode that runs it: Facility assessment. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Structured · Written · Observed
Typical outputs: Assessment report · Indicator dashboard · Dataset and codebook
Who commissions it: Ministries of health · Health programmes · Donors

Which routes suit it

Phone Call Not suitable
An assessor is on site.
WhatsApp Not suitable
An assessor is on site.
SMS Not suitable
An assessor is on site.
Web link Supported
Offline app Primary

Needs assessment

Supported

Establish what an affected population needs, at what scale and where, quickly enough for the answer to still be actionable.

What you should know

One record per unit on every collection surface. Member-level rosters — a household enumerated person by person — are held relationally by the platform, and there is no roster control on the field app or the web link, so a roster instrument is configured with your implementation team rather than built and fielded unaided.
Safeguarding wording, referral pathways and the handling of a disclosure are governed by your own protocols. The platform carries consent state and access control; it does not run a referral service.

Research mode that runs it: Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Study phase: rapid assessment — carried explicitly on the study as a wave, not inferred from the report title or from the date the fieldwork ran.
Evidence it collects: Selected · Written · Structured · Spoken
Typical outputs: Assessment report · Humanitarian brief · Dataset · Presentation
Who commissions it: Humanitarian response · INGOs · UN agencies · Disaster management authorities

Which routes suit it

Phone Call Supported
WhatsApp Supported
SMS Bounded
Workable for a short closed battery in a connectivity-limited response.
Web link Supported
Offline app Primary

Rapid needs assessment

Supported

A short instrument and a compressed fieldwork window, run in the first days of a response.

What you should know

Nothing in the platform structurally distinguishes a rapid assessment from any other structured survey. What makes it rapid is the instrument and the fieldwork window you set, not a different engine.

Research mode that runs it: Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Study phase: rapid assessment — carried explicitly on the study as a wave, not inferred from the report title or from the date the fieldwork ran.
Evidence it collects: Selected · Structured · Spoken
Typical outputs: Humanitarian brief · Assessment report · Dataset
Who commissions it: Humanitarian response · UN agencies · First responders

Which routes suit it

Phone Call Supported
WhatsApp Supported
SMS Bounded
Suits a very short closed battery.
Web link Supported
Offline app Primary

WASH assessment

Supported

Assess water points, sanitation facilities and hygiene practice, at household or facility level or both.

What you should know

Photographs, GPS points and file attachments are not question types a respondent or enumerator answers. Location is recorded about a submission, not asked as a question.
One record per unit on every collection surface. Member-level rosters — a household enumerated person by person — are held relationally by the platform, and there is no roster control on the field app or the web link, so a roster instrument is configured with your implementation team rather than built and fielded unaided.
A water-point inspection is recorded as structured answers with the submission’s location. A photograph of the water point is not a question type an enumerator can answer.

Research modes that runs it: Facility assessment, Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Structured · Written · Observed
Typical outputs: Assessment report · Indicator dashboard · Dataset
Who commissions it: WASH programmes · Humanitarian response · Local government

Which routes suit it

Phone Call Bounded
Household hygiene questions only; a water-point inspection is not a call.
WhatsApp Bounded
Household hygiene questions only.
SMS Not suitable
Not suitable for an inspection instrument.
Web link Supported
Offline app Primary

Protection assessment

Supported

Understand protection risks and access to services for an affected or vulnerable population, under strict handling rules.

What you should know

Sensitivity is governed at the study and access level — who may read the evidence, what may leave the tenant, and what a published product may contain. There is no per-question sensitivity flag that changes how one answer is stored.
Safeguarding protocols, referral pathways and the response to a disclosure remain yours. The platform holds evidence under access control; it does not make a protection decision.

Research modes that runs it: Structured survey, Qualitative research. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Written · Spoken
Typical outputs: Assessment report · Evidence appendix · Humanitarian brief
Who commissions it: Protection programmes · Humanitarian response · Human rights organisations

Which routes suit it

Phone Call Bounded
Only where a call to that number is safe for the participant.
WhatsApp Not suitable
A message thread persists on the participant’s device. We do not recommend it.
SMS Not suitable
A message persists on the participant’s device. We do not recommend it.
Web link Bounded
Only where the link can be given to a participant safely.
Offline app Primary

Gender & inclusion assessment

Supported

Measure how differently a programme, a service or a market works for women, for people with disabilities, and for groups a design might otherwise average away.

What you should know

Disaggregation is only as good as the questions asked. The platform cross-tabulates and tests differences between exactly two groups; it does not impute a characteristic nobody was asked about.

Research modes that runs it: Structured survey, Qualitative research. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Written · Structured · Spoken
Typical outputs: Assessment report · Analytical workbook · Presentation
Who commissions it: Gender programmes · Disability inclusion · Donors · Evaluation firms

Which routes suit it

Phone Call Supported
WhatsApp Bounded
Where the participant already uses it and it is safe to do so.
SMS Bounded
Short closed batteries only.
Web link Supported
Offline app Primary

Education & school assessment

Supported

Assess a school’s conditions, staffing, resources and learning environment against a defined checklist.

What you should know

One record per unit on every collection surface. Member-level rosters — a household enumerated person by person — are held relationally by the platform, and there is no roster control on the field app or the web link, so a roster instrument is configured with your implementation team rather than built and fielded unaided.
Photographs, GPS points and file attachments are not question types a respondent or enumerator answers. Location is recorded about a submission, not asked as a question.
School, teacher and learner are separate studies or separate instruments, linked by the identifiers you carry on them. The platform does not model a school → class → teacher → learner hierarchy as one nested record.

Research mode that runs it: Facility assessment. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Structured · Written · Observed
Typical outputs: Assessment report · Indicator dashboard · Dataset
Who commissions it: Ministries of education · Education programmes · Donors

Which routes suit it

Phone Call Not suitable
An assessor is at the school.
WhatsApp Not suitable
An assessor is at the school.
SMS Not suitable
An assessor is at the school.
Web link Supported
Offline app Primary

Monitoring & evaluation

The phases of a programme’s own evidence: what was true at the start, what changed, and what the change was measured against.

Baseline survey

Proven

Establish the value of a programme’s indicators before it starts, on the instrument the endline will repeat.

What you should know

One record per unit on every collection surface. Member-level rosters — a household enumerated person by person — are held relationally by the platform, and there is no roster control on the field app or the web link, so a roster instrument is configured with your implementation team rather than built and fielded unaided.
Change is reported descriptively — a value, a later value, and the difference between them. No causal or attribution claim is produced, because the platform implements no identification strategy and refuses to imply one.
Baseline is a PHASE, carried on the study as a wave. It is not a separate product, a separate engine or a separate research mode, and a report is a baseline because the study says so — never because the title says so.
A published Baseline report names the study and the wave it came from, so a Baseline and an Endline of the same study are distinguishable without reading their titles. Where a publication declares no wave, its phase reads as “not declared” rather than being guessed.

Research mode that runs it: Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Study phase: baseline — carried explicitly on the study as a wave, not inferred from the report title or from the date the fieldwork ran.
Evidence it collects: Selected · Written · Structured · Spoken
Typical outputs: Baseline report · Dataset and codebook · Analytical workbook · Indicator dashboard · Presentation
Who commissions it: NGO and INGO programmes · Donors · Evaluation firms · Government programmes
How this is proven: An end-to-end proof of this study type runs against the deployed platform on every release, and the release gate refuses the build if it is missing.

Which routes suit it

Phone Call Bounded
Suits a short indicator battery.
WhatsApp Bounded
Suits a short indicator battery.
SMS Not suitable
A baseline instrument is too long for one message at a time.
Web link Supported
Offline app Primary

Midline study

Supported

Re-measure the baseline indicators part-way through a programme, while there is still time to change something.

What you should know

One record per unit on every collection surface. Member-level rosters — a household enumerated person by person — are held relationally by the platform, and there is no roster control on the field app or the web link, so a roster instrument is configured with your implementation team rather than built and fielded unaided.
Change is reported descriptively — a value, a later value, and the difference between them. No causal or attribution claim is produced, because the platform implements no identification strategy and refuses to imply one.
A midline is a wave on the same study, with the same indicator identity and its own dataset version.

Research mode that runs it: Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Study phase: midline — carried explicitly on the study as a wave, not inferred from the report title or from the date the fieldwork ran.
Evidence it collects: Selected · Written · Structured · Spoken
Typical outputs: Monitoring report · Analytical workbook · Indicator dashboard
Who commissions it: NGO and INGO programmes · Donors · Evaluation firms

Which routes suit it

Phone Call Bounded
Suits a short indicator battery.
WhatsApp Bounded
Suits a short indicator battery.
SMS Not suitable
Instrument length.
Web link Supported
Offline app Primary

Endline survey

Supported

Repeat the baseline instrument at the end of a programme and report what changed.

What you should know

One record per unit on every collection surface. Member-level rosters — a household enumerated person by person — are held relationally by the platform, and there is no roster control on the field app or the web link, so a roster instrument is configured with your implementation team rather than built and fielded unaided.
Change is reported descriptively — a value, a later value, and the difference between them. No causal or attribution claim is produced, because the platform implements no identification strategy and refuses to imply one.
Change is reported as the baseline value, the endline value, and the difference — in percentage points where the indicator is a proportion. Whether the programme caused that change is a question of evaluation design, and the platform will not answer it for you.

Research mode that runs it: Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Study phase: endline — carried explicitly on the study as a wave, not inferred from the report title or from the date the fieldwork ran.
Evidence it collects: Selected · Written · Structured · Spoken
Typical outputs: Endline report · Dataset and codebook · Analytical workbook · Presentation
Who commissions it: NGO and INGO programmes · Donors · Evaluation firms · Government programmes

Which routes suit it

Phone Call Bounded
Suits a short indicator battery.
WhatsApp Bounded
Suits a short indicator battery.
SMS Not suitable
Instrument length.
Web link Supported
Offline app Primary

Post-distribution monitoring

Supported

Check what people actually received, whether it was usable, and what it displaced — after a distribution has happened.

What you should know

Linking a submission back to a specific distribution record is done through an identifier you carry on the instrument. The platform does not read your distribution list.

Research mode that runs it: Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Study phase: post distribution — carried explicitly on the study as a wave, not inferred from the report title or from the date the fieldwork ran.
Evidence it collects: Selected · Written · Structured · Spoken
Typical outputs: Monitoring report · Indicator dashboard · Dataset
Who commissions it: Humanitarian response · Cash and voucher programmes · INGOs

Which routes suit it

Phone Call Supported
WhatsApp Supported
SMS Bounded
Suits a short closed battery to a list you already hold consent for.
Web link Supported
Offline app Primary

Programme monitoring

Supported

Keep a results framework current — routine measurement against targets, reported as status and variance rather than as a study.

What you should know

Targets, achieved values, variance and status are carried on the indicator. Forecasting a trend forward is not implemented and is refused rather than approximated.

Research mode that runs it: Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Study phase: continuous monitoring — carried explicitly on the study as a wave, not inferred from the report title or from the date the fieldwork ran.
Evidence it collects: Selected · Structured · Spoken · Imported
Typical outputs: Monitoring report · Indicator dashboard · Board pack
Who commissions it: NGO and INGO programmes · Government programmes · Donors

Which routes suit it

Phone Call Supported
WhatsApp Supported
SMS Bounded
Short closed batteries only.
Web link Supported
Offline app Primary

Evaluation study

Supported

Answer evaluation questions about a programme — relevance, coverage, effectiveness as measured — on evidence with a traceable line back to who said it.

What you should know

Change is reported descriptively — a value, a later value, and the difference between them. No causal or attribution claim is produced, because the platform implements no identification strategy and refuses to imply one.
Treatment and comparison groups, and the design that defines them, are recorded as study metadata and analysed as a two-group comparison. Difference-in-differences, matching, regression discontinuity and interrupted time series are not implemented, and a request for one is refused with the reason rather than approximated.

Research modes that runs it: Structured survey, Qualitative research, Mixed methods. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Written · Structured · Spoken · Imported
Typical outputs: Evaluation report · Evidence appendix · Analytical workbook · Presentation
Who commissions it: Evaluation firms · Donors · INGOs · Government

Which routes suit it

Phone Call Supported
WhatsApp Bounded
For a short quantitative module within the design.
SMS Bounded
For a short quantitative module within the design.
Web link Supported
Offline app Primary

Programme review

Supported

Bring the quantitative record, the qualitative evidence and the programme’s own data together for a periodic review.

What you should know

Mixed-methods evidence is assembled with your implementation team rather than from a single review screen. Every finding carries the source and form of the evidence behind it, so the strands stay distinguishable rather than being concatenated.

Research mode that runs it: Mixed methods. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Structured · Spoken · Written · Imported
Typical outputs: Programme review report · Board pack · Presentation · Evidence appendix
Who commissions it: NGO and INGO programmes · Donors · Government programmes

Which routes suit it

Phone Call Supported
WhatsApp Bounded
For the quantitative module.
SMS Bounded
For the quantitative module.
Web link Supported
Offline app Supported

Qualitative research

What people say in their own words, recorded as evidence and coded rather than summarised away.

Key informant interviews (KII)

Proven

Interview people whose position makes them authoritative on the question — and keep what they said as evidence, not as a paraphrase.

Research mode that runs it: Qualitative research. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Spoken · Written
Typical outputs: Research report · Evidence appendix · Thematic summary
Who commissions it: Evaluation firms · Policy research · Programme evaluation · Government
How this is proven: An end-to-end proof of this study type runs against the deployed platform on every release, and the release gate refuses the build if it is missing.

Which routes suit it

Phone Call Supported
WhatsApp Bounded
For a short structured follow-up, not the interview itself.
SMS Not suitable
An interview cannot be conducted one message at a time.
Web link Bounded
Where the interview is typed rather than spoken.
Offline app Primary

In-depth interviews (IDI)

Proven

Long-form interviews with individuals, coded against a codebook so themes are traceable to the passages that produced them.

Research mode that runs it: Qualitative research. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Spoken · Written
Typical outputs: Research report · Evidence appendix · Thematic summary
Who commissions it: Research firms · Health research · Social research
How this is proven: An end-to-end proof of this study type runs against the deployed platform on every release, and the release gate refuses the build if it is missing.

Which routes suit it

Phone Call Supported
WhatsApp Not suitable
Not suitable for long-form interviewing.
SMS Not suitable
Not suitable for long-form interviewing.
Web link Bounded
Where the interview is typed rather than spoken.
Offline app Primary

Focus group discussions (FGD)

Supported

Group discussions recorded as one session, coded against the same codebook as the interviews around them.

What you should know

An FGD recording is one audio stream carrying several voices. The platform records the participant count and refuses to attribute a statement to an individual participant, because speaker separation is not implemented and a confident wrong attribution is worse than none.

Research mode that runs it: Qualitative research. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Spoken · Written
Typical outputs: Research report · Evidence appendix · Thematic summary
Who commissions it: Evaluation firms · Community research · Market research

Which routes suit it

Phone Call Bounded
Only where the group is convened around one line.
WhatsApp Not suitable
A group discussion is not a message thread.
SMS Not suitable
A group discussion is not a message thread.
Web link Not suitable
A group discussion is not a form.
Offline app Primary

Stakeholder consultation

Supported

Structured consultation with the institutions and groups a decision affects, kept as evidence rather than as meeting notes.

What you should know

Only three qualitative session formats are distinguished — KII, IDI and FGD. A consultation is recorded as an unspecified qualitative session, so it codes and reports like the others but does not carry a format of its own.

Research mode that runs it: Qualitative research. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Spoken · Written
Typical outputs: Consultation report · Evidence appendix · Policy brief
Who commissions it: Government · Policy research · Donors · Programme design

Which routes suit it

Phone Call Supported
WhatsApp Bounded
For short written submissions.
SMS Not suitable
Not suitable for a consultation submission.
Web link Supported
Offline app Primary

Community dialogue

Supported

Facilitated community discussion recorded as evidence, usually alongside a scorecard or a feedback round.

What you should know

Recorded as an unspecified qualitative session, with the same attribution limit as a focus group: several voices in one recording, and no statement attributed to an individual.

Research mode that runs it: Qualitative research. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Spoken · Written
Typical outputs: Community feedback report · Evidence appendix · Thematic summary
Who commissions it: Community programmes · Social accountability · Local government

Which routes suit it

Phone Call Bounded
Only where the dialogue is convened around one line.
WhatsApp Not suitable
A facilitated dialogue is not a message thread.
SMS Not suitable
A facilitated dialogue is not a message thread.
Web link Not suitable
A facilitated dialogue is not a form.
Offline app Primary

Citizen & community feedback

Evidence from the people a service is for, gathered on the routes they already use.

Citizen feedback survey

Proven

Ask the people a service is for what it is actually like to use, on the routes they already have.

Research mode that runs it: Citizen feedback. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Spoken · Written · Selected
Typical outputs: Community feedback report · Indicator dashboard · Presentation
Who commissions it: Local and national government · Utilities · Social accountability programmes · NGOs
How this is proven: An end-to-end proof of this study type runs against the deployed platform on every release, and the release gate refuses the build if it is missing.

Which routes suit it

Phone Call Primary
WhatsApp Primary
SMS Supported
Short closed questions, one at a time.
Web link Supported
Offline app Supported

Community scorecard

Supported

Score a service against criteria the community set, and publish the result back to the people who scored it.

What you should know

A scorecard is a PRODUCT that citizen feedback produces, not a separate way of collecting it. The scoring criteria are questions on your instrument; there is no scorecard designer.

Research mode that runs it: Citizen feedback. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Structured · Spoken
Typical outputs: Community scorecard · Community feedback report · Presentation
Who commissions it: Social accountability · Local government · Community programmes

Which routes suit it

Phone Call Supported
WhatsApp Supported
SMS Bounded
Scores only, not the discussion around them.
Web link Supported
Offline app Primary

Market & enterprise research

Customers, employees and markets — the same governed pipeline, without the humanitarian frameworks.

Customer satisfaction & CX

Proven

Measure what customers think of a service or product, including in the language and on the route they prefer.

Research modes that runs it: Structured survey, Citizen feedback. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Written · Spoken · Structured
Typical outputs: Research report · Indicator dashboard · Presentation · Analytical workbook
Who commissions it: Banking and financial services · Telecoms · Utilities · Retail and consumer
How this is proven: An end-to-end proof of this study type runs against the deployed platform on every release, and the release gate refuses the build if it is missing.

Which routes suit it

Phone Call Supported
WhatsApp Primary
SMS Supported
Web link Primary
Offline app Supported

Employee engagement

Supported

Measure how people experience working somewhere, with the anonymity that makes the answers worth having.

What you should know

Small-group results are suppressed below a minimum cell size so a team of four cannot be identified from a cross-tab. The floor can be raised for your organisation and cannot be lowered past one.

Research mode that runs it: Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Written · Structured
Typical outputs: Research report · Indicator dashboard · Presentation
Who commissions it: Enterprises · Public sector employers · NGOs

Which routes suit it

Phone Call Bounded
Where anonymity can still be assured.
WhatsApp Supported
SMS Bounded
Short closed batteries only.
Web link Primary
Offline app Supported

Market research

Supported

Understand a market — awareness, usage, attitudes, segments — including in markets where a panel does not exist.

What you should know

Ranking, constant-sum allocation and conjoint or pricing exercises are not question types the platform collects. A design that needs them is fielded as closed questions instead.

Research modes that runs it: Structured survey, Qualitative research. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Written · Spoken · Structured
Typical outputs: Research report · Analytical workbook · Presentation · Dataset
Who commissions it: Consumer goods · Financial services · Telecoms · Research agencies

Which routes suit it

Phone Call Supported
WhatsApp Supported
SMS Bounded
Short closed batteries only.
Web link Primary
Offline app Supported

Financial inclusion survey

Supported

Measure access to and use of financial services — accounts, credit, digital payments, and what stops people using them.

What you should know

One record per unit on every collection surface. Member-level rosters — a household enumerated person by person — are held relationally by the platform, and there is no roster control on the field app or the web link, so a roster instrument is configured with your implementation team rather than built and fielded unaided.
Amounts are collected as numbers, with the currency fixed by the question wording. There is no per-question currency register and no conversion between currencies.

Research mode that runs it: Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Written · Structured · Spoken
Typical outputs: Research report · Dataset and codebook · Analytical workbook
Who commissions it: Financial regulators · Development finance · Banks and fintechs · Research firms

Which routes suit it

Phone Call Supported
WhatsApp Supported
SMS Bounded
Short closed batteries only.
Web link Supported
Offline app Primary

Government & public sector

Service delivery, consultation and policy evidence, with the hierarchy a public institution actually reports through.

Government service delivery survey

Supported

Measure how a public service performs where it is delivered, reported through the administrative hierarchy that owns it.

What you should know

Results roll up through the country → region → district hierarchy the platform holds. Nothing in the platform draws a legal or regulatory conclusion from that evidence.

Research modes that runs it: Citizen feedback, Structured survey. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Written · Spoken · Structured
Typical outputs: Government brief · Performance report · Indicator dashboard · Presentation
Who commissions it: National and local government · Public service commissions · Donors funding governance

Which routes suit it

Phone Call Supported
WhatsApp Supported
SMS Bounded
Short closed batteries only.
Web link Supported
Offline app Primary

Public consultation & policy research

Supported

Gather submissions and views on a proposed policy, in the languages the public actually uses, with the record kept.

What you should know

A submission is evidence, not a vote. The platform does not deduplicate submissions by identity, and it draws no legal or policy conclusion from what was submitted.

Research modes that runs it: Citizen feedback, Qualitative research. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Spoken · Written · Selected
Typical outputs: Consultation report · Policy brief · Evidence appendix · Government brief
Who commissions it: Ministries and agencies · Regulators · Parliaments · Policy institutes

Which routes suit it

Phone Call Supported
WhatsApp Supported
SMS Bounded
For a short structured response, not a submission.
Web link Primary
Offline app Supported

Advanced & combined designs

Designs that combine methods, repeat over time, or draw on records the platform did not collect.

Mixed-methods study

Supported

Run a structured survey and qualitative work as one study, so the numbers and the accounts answer the same question together.

What you should know

The strands are combined with your implementation team rather than from a single mixed-methods screen. Every finding carries the source and evidence form behind it, so the design remains inspectable — but the integration is a piece of analytical work, not a button.

Research mode that runs it: Mixed methods. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Selected · Written · Structured · Spoken · Imported
Typical outputs: Evaluation report · Evidence appendix · Analytical workbook · Presentation
Who commissions it: Evaluation firms · Donors · Research institutes · Programme evaluation

Which routes suit it

Phone Call Supported
WhatsApp Bounded
For the quantitative strand.
SMS Bounded
For the quantitative strand.
Web link Supported
Offline app Primary

Longitudinal & panel study

Supported

Follow the same units across rounds, and report change with the attrition stated rather than hidden.

What you should know

Change is reported descriptively — a value, a later value, and the difference between them. No causal or attribution claim is produced, because the platform implements no identification strategy and refuses to imply one.
Panel identity is an opaque per-study linkage token, never a phone number, so a participant can be followed across waves without their identity travelling into the analysis.
Waves, linkage tokens and attrition states are configured with your implementation team. There is no panel-management workspace.

Research mode that runs it: Longitudinal and panel. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Study phase: repeated wave — carried explicitly on the study as a wave, not inferred from the report title or from the date the fieldwork ran.
Evidence it collects: Selected · Written · Structured · Spoken
Typical outputs: Research report · Analytical workbook · Dataset and codebook · Indicator dashboard
Who commissions it: Research institutes · Evaluation firms · Long-running programmes · Government statistics

Which routes suit it

Phone Call Supported
WhatsApp Supported
SMS Bounded
For a short recontact battery.
Web link Supported
Offline app Primary

Observation & field assessment

Supported

Record what a trained observer sees — a site, a service being delivered, a market, a condition — as structured evidence with a time and a place.

What you should know

Photographs, GPS points and file attachments are not question types a respondent or enumerator answers. Location is recorded about a submission, not asked as a question.
Observed evidence is recorded by an identified observer through the field surface — it is not a question type on a public link, and we do not present it as one. The observer, the time and the location travel with the record.

Research mode that runs it: Observation and field assessment. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Observed · Selected · Structured · Written
Typical outputs: Assessment report · Indicator dashboard · Dataset
Who commissions it: Monitoring visits · Infrastructure and compliance · Market monitoring · Environmental programmes

Which routes suit it

Phone Call Not suitable
An observer is present at the site.
WhatsApp Not suitable
An observer is present at the site.
SMS Not suitable
An observer is present at the site.
Web link Supported
Offline app Primary

Imported & administrative data study

Supported

Bring programme records, facility registers or an existing dataset into the same governed analysis as the evidence you collected.

What you should know

Imported is an INGESTION PATH, not a collection route and not a question a respondent answers. It has no channel, and it is counted separately everywhere it appears.
Every imported record keeps its source system, the import event, the file identifier, the tenant and project it was scoped to, and its dataset version — so an imported figure can never be presented as something a respondent said.
Imports are staged and validated before anything is committed. A file that fails validation is rejected whole rather than partially loaded, because a dataset silently missing its failed rows has a denominator nobody can reconstruct.

Research modes that runs it: Structured survey, Mixed methods. A study type is what the work is commissioned as; the mode is the design the platform runs. See the mode.
Evidence it collects: Imported
Typical outputs: Analytical workbook · Dataset and codebook · Research report
Who commissions it: Government statistics · Health information systems · Programme M&E · Research firms

Which routes suit it

This is an ingestion path, not a collection route. It has no channel rating because nobody answers it.

What we do not offer, and why

Three study types are deliberately absent from the list above. They are named here rather than left out, because an omission with no explanation reads as an oversight.

Not implemented

Impact evaluation (causal)

A causal impact claim requires an identification strategy and a counterfactual. The analytical engine implements neither and refuses the request by name: CAUSAL_INFERENCE, DESIGN_BASED_VARIANCE, REGRESSION, ANOVA and TIME_SERIES are all registered as not implemented, with reasons. The platform can record a randomised or quasi-experimental DESIGN as study metadata and report a two-group comparison; it cannot license the word "impact", so it is not offered.

What we do offer instead: Evaluation study.

Not implemented

Census enumeration

A census enumerates every member of every unit, which is a repeat group — and repeat groups are deliberately not supported, because flattening one into ordinary questions turns a household into a single row. A complete-coverage survey at unit level is a household survey with a census sampling design and is offered as one.

What we do offer instead: Household survey.

Not implemented

Conjoint & pricing research

Conjoint, MaxDiff and constant-sum allocation are question types the collection surfaces do not implement, and the estimation behind them is not in the analytical engine. Offering the label while collecting closed questions instead would be a study nobody could analyse.

What we do offer instead: Market research.

Bring us the brief

Tell us what you need to find out, in which countries and in which languages, and we will map it onto the study types above — including saying plainly where the platform would leave work to your team.

Request a demo Check your languages See offline collection