01
Find a mission
The shopper opens the app and picks a nearby mission, a task at a store they are already going to or passing.
Nearby→
Case study · Snooper
Turn a normalshop into cash.
A mystery-shopping mobile app that pays everyday shoppers to complete quick in-store missions, like snapping a few photos while they shop, and pays the reward straight to their bank. For brands, that crowd becomes a live stream of real, on-shelf retail data from stores they cannot stand in themselves.
What we shipped
Location-tied missionsIn-store photo captureBank payoutsSubmission review3 languagesRetail intelligence
01 — The product
Everything is a mission
A mystery-shopping app that pays everyday shoppers to complete quick in-store missions, and turns the results into a live stream of retail data for the brands that cannot stand in those stores themselves.
Missions, not surveys
Everything is a mission: a small, well-scoped task tied to a place and a reward. Scoping the work this tightly is what lets a stranger do it correctly on the first try, without training.
Photo capture, not opinions
Shoppers document what they see, most often photos of shelves, displays and products, so the data is grounded in a real store rather than a survey answer.
Cash to a bank account
Completed missions pay cash directly to a bank account, or convert to redeemable rewards, so participation is genuinely worthwhile.
02 — The problem
Why is a two-sided earning app hard to get right?
You are building two products that have to stay in balance: an earning app that shoppers will actually open, and a data pipeline that brands will pay for. Neither works without the other.
A mission has to be small and unambiguous enough for a stranger to complete correctly the first time, with no training — and it has to fit inside something people already do, a normal shop, not ask them to make a special trip.
Trust runs both ways: shoppers need a payout they trust enough to bother completing unglamorous tasks, and brands need submissions that are verified, not just uploaded.
2
products that have to stay in balance — an app shoppers actually open, and a data pipeline brands will pay for.
03 — Goals & challenges
What Snooper had to balance
Neither side of the model works without the other, so the goals and the hard parts are really the same list, seen from opposite sides.
Goals
Make both sides work
- Scope every mission tightly — one place, one clear ask, a photo or two.
- Capture real evidence — photos of shelves, displays and products, not survey answers.
- Pay straight to a bank account, so participation is genuinely worthwhile.
- Verify before it counts, so a brand receives trustworthy fieldwork.
Challenges
Why it was hard
- Two products have to work at once — an app shoppers want to open, and a pipeline brands will pay for.
- Verification has to be real, not a checkbox, or the whole model's credibility collapses.
- Trust in the payout is what keeps a crowd active, so cash had to beat points or vouchers on day one.
- Running in English, Japanese and Spanish means adapting mission content, the app and payouts per market, not just translating text.
04 — Two sides of one loop
Who is Snooper actually built for?
A mystery-shopping app is really two products sharing one loop — a shopper who earns, and a brand who buys what the crowd sees.
Shopper
Everyday shoppers
Opens the app before a normal trip, picks a nearby mission, completes it in a few photos, and gets paid straight to their bank account or in redeemable rewards.
Brand
Brands & retailers
Never sets foot in the store themselves, but receives the aggregate of thousands of verified missions as a live picture of how their products actually appear on shelves.
05 — How it earns
How does a shopper turn a trip to the shop into money?
02
Shop and capture
They complete the task in store, often taking a few photos of a shelf, a display or a product, in the normal course of shopping.
In-store→
03
Submit the proof
The photos and answers are submitted in the app and checked, so a brand can trust the data came from the real aisle.
Verified→
04
Get paid
The reward is paid straight to the shopper's bank account, or redeemed, turning a normal shop into supplementary income.
Paid✓
keep scrolling — the deal glides left →
06 — What we shipped
The scope, in one panel
Every number in this panel is delivered scope — what was designed, built and handed over for Snooper. It leaves out traffic, conversion and revenue on purpose: those figures belong to the client, and we don’t publish numbers we cannot stand behind.
4
Steps: find, capture, submit, get paid
3
Languages the app runs in
2-sided
A shopper earning app and a brand data pipeline
6
Core capabilities, mission to retail intelligence
- The model
- Mystery shoppingCrowdsourced retail data
- What shoppers do
- In-store photo missions
- The reward
- Cash to a bank accountRedeemable rewards
- Reach
- English, JapaneseSpanish
07 — Visual design
What does the identity signal?
This is a consumer earning app, so it has to feel light, trustworthy and instantly legible in a shop, not corporate. A bright, optimistic cyan carries the brand and every reward moment against a clean canvas, and Bricolage Grotesque, a characterful contemporary sans, gives it personality without sacrificing the readability a task app needs mid-aisle.
Colour
Snooper cyan
#00C7FC
Ink 900
#0F1B2D
Slate 800
#1E293B
Grey 500
#6F6F6F
Off-white
#FAF7F5
White
#FFFFFF
Typography
Turn a shop into cash
How it earns
Find a mission
Shoppers document what they see, most often photos of shelves and displays.
HOW IT EARNS
Iconography
08 — Transferable
What does a crowdsourced-work app have to get right?
These came out of building a two-sided app where strangers do paid fieldwork for brands. They apply to any platform that turns a crowd into reliable data.
01
Scope the task so a stranger can nail it
A mission has to be small and unambiguous enough to complete correctly first time, with no training, or the data is worthless.
02
Make earning fit real life
The work has to slot into something people already do, a normal shop, not ask them to make a special trip, or the crowd never shows up.
03
Pay in a way people trust
Cash straight to a bank account is a stronger, clearer incentive than points, and trust in payout is what keeps a crowd active.
04
Verify before it counts
Submissions have to be checked, because the value to the brand is that the data is real, from the actual aisle, not guessed.
05
Ground answers in evidence
A photo of a shelf is proof in a way a survey tick is not, so capture, not just questions, sits at the centre of the task.
06
Localise to scale
Running in multiple languages lets one platform serve very different retail markets without rebuilding the product for each.
FAQ
Building a crowdsourced-data app
How do you get reliable data out of a crowd of untrained shoppers?
By engineering the task, not the worker. Each mission is scoped so tightly, one place, one clear ask, a photo or two, that a first-time user can complete it correctly without any training, and every submission is verified before it counts. The reliability comes from the design of the mission and the review step, not from assuming the crowd is expert. Get that scoping right and thousands of strangers produce clean, comparable data.
Why pay shoppers in cash rather than points or vouchers?
Because trust in the payout is what keeps a crowd active, and cash to a bank account is the clearest, most trusted form of reward. Points and vouchers add friction and doubt, exactly what you do not want when the whole model depends on people bothering to complete unglamorous tasks. A simple, direct payout is a feature, not just an expense.
What makes a two-sided app like this hard to build?
You are building two products that have to stay in balance: an earning app that shoppers will actually open, and a data pipeline that brands will pay for. Neither works without the other, and the hard part is the connective tissue, mission distribution, verification, payouts and turning raw submissions into intelligence a brand can act on. The consumer screens are the visible tenth of the iceberg.
How do you support multiple countries and languages?
By treating localisation as core, not an afterthought. Running in English, Japanese and Spanish means the mission content, the app and the payout all have to adapt per market, and the platform has to serve very different retail environments from one codebase. Built in from the start, that lets a single product scale into new markets without a rebuild each time.
How long does an app like this take to build?
A two-sided platform with a consumer earning app, mission distribution, in-app capture, a verification workflow, bank payouts and a brand-facing data layer is a multi-month build. The visible shopper app is the fast part; the pipeline that turns crowd submissions into trustworthy, sellable retail intelligence is where most of the engineering, and most of the value, actually sits.
Let's build
Building a platform that turns a crowd into data?
Tell us how your business actually works and we'll tell you honestly what it takes to build — the app, the verification, the data pipeline. No deck, no pitch.