Millions go into radio, activations and trade promotions every month. Afterwards nobody can say whether anyone heard it, whether the product reached the shelf, or whether the wholesaler passed the price cut on.
Field reps file visit reports nobody verifies. Research houses charge tens of millions and take two months. So the decision gets made on a feeling.
“The ads must not be working.” Often they are. In one launch, 54% of people who wanted it had never seen it in their duka. The media budget was about to be cut for a distribution failure.
“Our price promotion is live.” On paper. The wholesaler had absorbed the Ramadan cut on a 5L oil before it reached a single shopper.
“Distribution is at 80%.” That is what the reps reported. Shelf checks by consumers standing in the shop said otherwise — ward by ward, with a photograph attached to each one.
Every figure above came out of a real study run on this platform.
You write the question. We do the rest.
Pick what you need to know — did they hear the campaign, is it on the shelf, what is it selling for. Standard question sets are one click, already written in Kiswahili.
We invite panelists in the wards you chose. They dial a short code and answer in ninety seconds, on any handset. No app, no data bundle. They are paid airtime instantly.
Every answer is checked for duplicate photographs, impossible completion times, straight-lining and location before it counts. Rejected answers are still paid for — and never counted.
The findings arrive as a report you can hand to a board: what people said, ward by ward, with the sample size and every exclusion stated openly.
Is it on the shelf, at what price, next to which competitor — with a photograph from inside the shop. Ward by ward, worst first, so you know where to send someone.
Did anyone outside the newsroom hear it, did they know whose advert it was, and did they go looking for it afterwards.
Anything else you need from ordinary buyers: what they pay, where they shop, what decides it.
The national average tells a brand they have a problem. This tells them the address.
One person with a drawer of SIM cards. A recruiter enrolling relatives who never leave the house. The same shelf photographed twice and sent from two numbers. This is the part of the product you are actually buying, and it runs on every single response before it counts.
A shelf photograph already submitted — on any study, for any client, at any time — is rejected. Recycling a picture across two clients is the most common fraud in crowdsourced retail work anywhere in the world.
An answer completed faster than the questions can physically be read is a script or a memorised key sequence, not a person who walked into a shop.
Answers from outside the wards you paid for, panelists below the trust floor you set, and accounts too new to have any record are all flagged before a single figure is calculated.
Individual fraud is expensive to chase. Recruitment fraud is not — it arrives in blocks, because one bad recruiter produces a group of panelists that all fail the same way.
Rejected answers are still paid for. A panelist whose honest answer trips a check keeps their airtime — that cost is ours, not theirs. It is the only way a panel keeps answering. And every report states how many were excluded and why, because a client who cannot see what was thrown out has no reason to believe what was kept.
Salaries, transport, supervisors. Accurate, and priced accordingly — which is why most brands commission it once a year and guess for the other eleven months.
Because we don’t send anyone. The shopper is already in the shop, already holds a phone, and will answer four questions for airtime. That is the entire difference.
Cheap enough to run monthly — and a monthly tracker is worth more than any one-off study, because the comparison between months is the finding.
Your own workspace. Run a campaign, commission the fieldwork against it, and the report says what you published and what consumers actually heard — one document, not two.
Open a business accountCommission fieldwork on a client’s behalf, mark it up, and invoice from the same screen. You keep the client relationship and the margin; we run the panel.
Open an agency account
Their own reps had reported eighty per cent distribution. Nobody had checked. The media budget was about to be cut for what turned out to be a distribution problem.
That gap — between what a company believes and what is actually happening in a duka in Temeke — is where a great deal of money in this country quietly disappears.
We could not close it the way it gets closed elsewhere, by hiring an army of enumerators. Too slow, too expensive, and the numbers arrive two months after the decision was made. So we did the only thing that works at Tanzanian prices: we asked the person already standing in the shop.
Ten thousand ordinary Tanzanians with a phone in their pocket can tell you more about your distribution in ten days than a field force can in two months. They only need to be asked properly, in their own language, and paid on time.
That last part matters more than anything technical we built. If the airtime does not arrive, there is no panel. So we pay on completion — even for the answers we go on to reject.
The plan covers your workspace, campaigns and reports. Fieldwork is charged per verified response — you are never billed for an answer that failed our checks. 14-day free trial, no card required.