Skip to content
saanjha
In practice

Four worked examples, not four success stories.

Saanjha is in private pilot, so there are no customer results to report and we are not going to invent any. What follows is the mechanics, followed end to end — including, for each one, where it comes apart.

What these are, and what they are not

Every scenario below is illustrative. There is no real household, shop, electrician or society being described, no quotes from anybody, and no claimed outcome. The quantities exist to make a mechanism legible — they are not results.

Nothing here says how much anyone saved. When there are validated figures from the pilot, they will be published with their source and these examples will be replaced by the real thing — described as what it is, rather than relabelled.

A household

Eleven flats, one trip to the wholesale market.

The mechanic Saanjha is built around, followed end to end. Nobody in this example does anything they would not already do — one person was going to the market anyway.

  1. 1

    Thursday. A member of the community posts that she is going to the wholesale market on Saturday morning, and sets the order cutoff at Friday 4am — three hours before she leaves.

  2. 2

    Thursday–Friday. Eleven households add what they need. Rice, atta, oil, onions. The app shows the running total against the lot size each item needs to clear.

  3. 3

    Friday 4am. The cutoff passes. Two items did not reach a workable quantity and are dropped, and the households who wanted them are told why rather than quietly charged.

  4. 4

    Saturday morning. She buys at the counter price, which is not knowable in advance — fresh produce is weighed, and the shop’s price is the shop’s price.

  5. 5

    Saturday midday. Handover at an agreed point in the building. Each household takes its share and pays her directly, by UPI or cash, against the real bill rather than the estimate.

  6. 6

    Saturday afternoon. The trip closes. What each household actually paid is recorded against the order, so a disagreement later has something factual to start from.

What made this work

The cutoff was real, so she could actually leave on time. The community had already passed the threshold where a trip can fill a lot. And payment happened at handover, so nobody fronted money to a neighbour on trust alone.

Where this goes wrong

In a community of six households the lot never clears, the trip is cancelled, and every one of those six is less likely to try again. That is precisely why a new community cannot host a trip until enough verified households have joined — the gate exists to prevent this scenario, not to slow anyone down.

A local store

A shop that turns walk-ins into a customer base it keeps.

The loop a neighbourhood store runs, from the QR on the counter to knowing what to buy before buying it.

  1. 1

    The shopkeeper registers a storefront in the app — name, category, address, community — and submits it for verification. It is not visible to buyers until that completes.

  2. 2

    A QR code goes on the counter and on a flyer. A customer already standing in the shop scans it and joins that store’s community; the join is attributed to the store.

  3. 3

    The store sets a welcome offer on its own terms and funds it, so it controls the value, the conditions and when it stops.

  4. 4

    Orders start arriving inside a collection window the store defines, rather than as a stream of individual pings through the day.

  5. 5

    At the end of the window, the store sees consolidated demand — not eleven separate requests, but a quantity worth negotiating on.

  6. 6

    The store buys against that number, allocates what arrived back to the orders that generated it, and tells anyone whose item was short rather than substituting silently.

What made this work

Every customer here was already walking into the shop. Nothing in the loop depends on Saanjha spending money to buy the store a customer, which is why it does not stop working when a marketing budget does.

Where this goes wrong

A store that treats the announcement channel as a broadcast list gets muted, and past a threshold loses promotional rights for a period. The limit is not there to frustrate the shop — it is there because attention in a few hundred households is finite, and a channel everybody mutes is worth nothing to the next store.

A service provider

An electrician whose second job is not somebody else’s lead fee.

What changes for an independent professional between the first job and the fifth.

  1. 1

    The electrician registers, completes the checks for the category, and publishes the slots he can genuinely serve.

  2. 2

    A household in a nearby community posts a job. He sees it because he covers that area, not because he outbid anyone for the placement.

  3. 3

    He quotes. The household compares his quote with others on price, timing and availability — and on whether a neighbour has used him before.

  4. 4

    On site the job turns out to need a part that was not in the original scope. He records a revised quote before continuing, so the change is agreed rather than sprung at the end.

  5. 5

    The work finishes and the household pays him directly. Nothing is deducted, and there is no settlement cycle to wait out.

  6. 6

    Three weeks later the same household needs something else. They book him again — and because they are already recorded as his customer, that second job is not resold to him as a fresh lead.

What made this work

The revised quote was recorded before the extra work, which is where most doorstep arguments actually start. And the repeat booking cost him nothing, which is the whole difference between a platform he uses and a platform he rents access from.

Where this goes wrong

A provider who publishes slots he cannot serve, or who quotes low to win and revises up on arrival, damages the one thing that makes a neighbourhood marketplace work at all. Fulfilment against accepted quotes is tracked, and the privilege of quoting is losable.

A community

A society going from nothing to a working local market.

The slowest of the four, and the one everything else depends on. A community is not useful on the day it is created.

  1. 1

    A resident creates the community, draws the boundary around the actual buildings rather than a circle that sweeps in three towers and a market, and sets the join policy.

  2. 2

    Neighbours are invited by link, by a QR poster on the noticeboard, and through WhatsApp. Each join is checked against the boundary and, where the policy requires it, approved by an admin.

  3. 3

    The community sits in a forming state. Members can chat, invite and organise — but community purchases stay locked, and the progress towards unlocking them is visible to everyone.

  4. 4

    The threshold of verified households is reached. Community purchases unlock.

  5. 5

    A local store joins and brings its own customers with it, which grows the community faster than the residents alone were managing.

  6. 6

    Service providers join, because the demand is now visibly there. Selection improves, demand steadies, and sourcing gets easier for every seller in it.

What made this work

The boundary matched a real building rather than a radius, so membership meant something. And the forming state was honest about what was not yet possible instead of letting the first purchase fail.

Where this goes wrong

An admin who approves anyone who asks hollows out the one guarantee the community offers its members. A geofence alone does not prove residence — anybody standing outside the gate passes it — which is why a flat number, an admin who recognises their residents, and neighbour vouching all sit on top of it.

What comes next

What we will publish when we can.

A real case study needs a real outcome, a real participant who agreed to be named, and a number somebody can check. All three are coming; none of them are here yet.

Measured, not estimated

Figures from settled orders rather than from a model, with the method stated alongside them.

Named with permission

No participant appears without agreeing to it, and a community is never identified in a way that exposes its members.

Including the ones that failed

A community that never reached its threshold teaches more than one that did. We would rather publish both than curate.

Want to be one of the real ones?

If you run a shop, offer a service, or want your society on Saanjha, the pilot is how these examples stop being illustrative.