The best participant doesn’t have to be the most visible one.
FromVisibility→ToTrusted fit
In an open network there is no single ranking that every buyer is measured against. The question moves from who is most visible to who can fulfil this need under conditions the buyer is willing to accept, with protection, evidence and track record visible at the point of choice.
A buyer describes the job, not a brand
Kavita runs a small packaging unit in Pune. A part in her sealing machine has worn out, and she needs a replacement cut to her drawing within ten days illustrative. She doesn’t want a brand. She wants someone who can make it, who can show they have made things like it, and who will work on terms she is comfortable with.
On her buyer app she describes the job and adds three conditions: delivery by the 30th, payment released only after she has inspected the part, and a workshop that has completed similar custom orders. The network returns three fabricators. Each shows what it can make, what has been verified about it, how it handles payment and what it has delivered before.
One of them, Imran’s workshop in Rajkot, is small and little known beside the other two. It is on her list because it fits the job and meets her conditions, not because it is the most visible.
Discoverability is more than visibility
In a marketplace, being found often depends on being visible, and visibility is built from a history on that marketplace. An open network has no single place where a participant’s history lives. Discovery there needs a different recipe.
Discoverability = capability × evidence × trust × transaction fit
Capability: can this participant do the job? Evidence: what has been demonstrated, and who says so? Trust: under what conditions will they transact, such as protection, delivery terms and resolution? Transaction fit: does it match this buyer’s need, place, timeline and budget?
Think of it as a product rather than a sum. A gap in any one factor should show, not be averaged away.
Protection you can search for
Search today captures what a buyer wants. Trust-aware search would also capture the conditions under which they are willing to buy: protected payment, inspection before release, delivery evidence, a resolution provider they accept. Those conditions become filters, and a seller who offers them becomes easier to find for the buyers who care. Trust stops being a badge on the side and becomes part of the match.
Buyers weigh these conditions differently, so the network’s job is to return a relevant set with reasons, not one list that claims to be best for everyone.
Go deeper
What has already been proposed
ONDC’s 2022 consultation paper on trust proposed that gateways forward a search to all relevant seller apps and return every response without filtering, that buyer apps publish their listing criteria and general ranking guidance and not favour their own captive sellers, and that sponsored or boosted results be clearly labelled[1]. It treated scoring and badging as a separate layer, with scores that move with the seller between apps (Shift 7 picks this up)[1].
These were proposals put out for comment. What the network has adopted since needs checking against its current policy before anyone builds on it. This shift starts where they leave off: what a buyer sees when the conditions of a transaction, and not only the product, are part of discovery.
Four design principles
A fair start for newcomers. A participant with a verified identity and demonstrated capability, but no history, should be visible and labelled as new. Otherwise the record of those already established becomes a wall. The Hyperlocal teardown meets the same cold-start problem from the other side.
The buyer decides. The network can suggest conditions. It never imposes them.
No paid placement inside trust signals. Anything promoted is labelled, and promotion never changes the trust information shown beside it.
Explainable. For every participant shown: why this one, what has been verified, and what has not.
What could go wrong
- A filter bubble of the proven.Buyers who filter for the strongest evidence may never see newcomers.
- Too many conditions.If a buyer must set five conditions to get a good result, the idea fails. Defaults matter.
- Signals that can be bought.If protection or evidence badges can be acquired without being earned, they stop meaning anything.
- Adoption.Buyer apps would have to read trust signals from other participants in a common format.
How we’d test it
One category on one network. Results will be directional, so every kill threshold is written down before the test starts.
Buyers ignore trust conditions, or newcomers never get enquiries even when shown.
| The test | What we would measure | What would end the idea |
|---|---|---|
| Protection as a search condition | Offer protected payment and delivery evidence as search conditions. Measure how often buyers use them, and whether those who do complete more orders with fewer problems | Buyers ignore the conditions, or using them changes neither completion nor problem rates |
| A fair start | Show newcomers with a visible “new” label beside established participants. Measure whether they receive and convert enquiries | Newcomers receive almost no enquiries even when shown |
Sources
- [1]ONDC, Building Trust in the ONDC Network: Consultation Paper (29 September 2022)Checked; proposals for comment, not adopted policy
Spotted a source we missed or a rule we misread? Tell us what we got wrong.
Once a buyer has found someone they are willing to trust, the next question is what happens to the money.