The open link
A public URL plus an airtime prize invites the fastest finger, including people who are not in Kenya. Send a tracked invite. Count the finish, not the click.
Surveys
Survey participants in Kenya: completions you can stand behind
A cheap Kenya complete is easy to buy and hard to defend. Speeders, people outside the country, and respondents who qualify for every brief will give you a chart. They will not give you a decision. A short survey of screened people will.
Use a survey when you already know the question and the words. “Of people who paid a boda with M-Pesa this week, how many would rather pay the rider than an app?” can be a survey. “What do Kenyans think about our brand?” cannot — not until interviews have told you which thoughts exist. If you are still discovering the category, start with market research participants, not a 40-item questionnaire.
Afritesting fills the completes from a Kenyan panel. You keep Typeform, Google Forms, SurveyMonkey or whatever you already use. We screen, send people in, and count a response when it is actually finished. Surveys start at KES 500 a completion, incentive included. Fifty consumer responses are typically about KES 40,000.
Fifty is a pattern, not a population.
Say “among screened people who matched this brief”. Do not say “Kenyans”. Afritesting is not a probability sample of the country, and pretending otherwise is how survey slides get thrown out.
Fraud is a design problem
A complete is not automatically a person.
The obvious screener
“Do you use our category at least weekly?” is an instruction. Ask which ordinary things they did this week, and disqualify on behaviour. The same rule as any other participant brief.
The professional
Someone who completes surveys for a living will pass a sloppy screen and flatten every scale toward the middle. Cap how often a person is invited, and throw out finishes that are faster than a human can read.
The duplicate
One phone, two accounts, three panels. Ops checks exist for this. If a batch looks cloned — same open-ends, impossible combinations — do not average it into the chart. Replace it.
The questionnaire
Write it for a matatu, not a desk.
Your respondent is on a phone, in Kiswahili or English, with a bundle that makes images expensive and a conductor who will interrupt question seven.
Under six minutes
If you need 25 minutes, you need an interview. Long surveys in Kenya select for people with time to sell. Cut anything you will not put on a slide.
One question, their words
No grids of twelve brands. No stock photo. Use the phrases interviews already heard. If you have not heard them, you are not ready — see what to study first.
Language without a trap
Offer English and Kiswahili if the task is bilingual in real life. Do not mix both inside one sentence. Sheng in an open-end is a valid answer, not a data-quality error.
Reading the number
What 50 responses can change, and what they must not.
Fifty screened completes can tell you that a problem is common among people like your brief, or that it is rare enough to stop arguing. They can kill a feature nobody uses. They cannot estimate market size, vote share, or “Kenya’s” opinion. County quotas are a wish unless you have the budget and the time to wait for the thin cells. Ask for a mix — Nairobi and one other place — and report the mix you got.
When the chart says 40% are confused, you still do not know the confusion. Stop fielding and book interviews or a usability test. The survey did its job by telling you the question is open.
Price stays simple: no subscription, KES 500 and up per response, the payout included, the quote visible before you pay. A harder screen — a profession, a recent loan, a specific county outside the obvious cities — moves the quote. You are paying for people who match, not for a larger vanity sample of whoever clicked.
Questions
Before you field.
How much does a Kenya survey cost on Afritesting?
Surveys start at KES 500 per verified completion, recruitment and incentive included. A 50-response consumer survey is typically about KES 40,000. You use Typeform, Google Forms, SurveyMonkey or a similar tool. Harder audiences are quoted before you pay.
Is a 50-person Kenya survey nationally representative?
No. It is a screened group of people who match the brief you wrote. That can tell you whether a pattern shows up among those people. It cannot size the Kenyan market or stand in for a random national sample.
How long should a mobile survey in Kenya be?
Aim for under six minutes, one question at a time, with no large images. People are on a phone, often on a bundle, often interrupted. A 25-minute questionnaire selects for people who survey for a living.
When should I interview instead of surveying?
When you do not yet know the words people use, or when the survey would only tell you that many people are confused. Confusion needs a conversation. Run interviews first, then a short survey to see whether the pattern repeats.
Keep going
If a survey is the wrong shape.
What do you need a count of?
One behaviour, a short form, a number of completes. We’ll screen Kenyan participants into the survey tool you already use.