A lead source can't be judged on its first delivery. An opening batch may be exceptionally good — or disappointing — without saying anything about what that source will produce month after month. Measuring a source's return over time means watching how consistent its quality stays across a window long enough to tell durable performance apart from plain luck, a seasonal effect, or a one-off spike. On a marketplace like leads-qualifie.ch, this measurement isn't a financial dashboard reserved for the buyer: it's an internal mechanism of the platform, applied symmetrically to every active source regardless of how long it's been around.
This dossier explains how a marketplace measures a source's return over time, without ever falling back on cost-per-contact reasoning. It sets out why a source is judged on a series rather than a snapshot, how traceability ties each request back to its origin, what signals receiving companies send back after attempting contact, how those signals condense into a source rating that shifts over time, and what this ongoing measurement concretely changes for both sides of the market.
Why a source is judged over time, not on a single batch
A source's first batch is a poor estimator of its real quality. Across a small number of requests, chance dominates: ten reachable contacts in a row don't prove structural quality, and one bad week doesn't condemn a source that turns out solid over the long run. Only through accumulation, as requests pile up, does a source's underlying quality finally emerge from the noise. Judging too early means mistaking a random fluctuation for an underlying trend — an error that would unfairly penalise a good source that started out unlucky, or wrongly keep a mediocre source that happened to open with a flattering batch.
On top of this come seasonality and drift. A source may be excellent at one time of year and weaker at another, or see its quality slowly shift as its own capture methods change. That's why the platform never issues a fixed verdict: it watches a rolling window that tracks the source continuously, keeping in mind the principle of regression to the mean — an extreme batch, whether very good or very bad, tends to be followed by batches closer to the source's usual level. Measuring over time is precisely what makes it possible not to overreact to these occasional extremes.
Traceability: tying every request back to its source
No measurement is possible without traceability. Every request that enters the platform carries, from the moment of capture, an identifier for the source that produced it, together with a timestamp, the channel of origin, and proof of consent. This end-to-end history — capture, validation, distribution, feedback from the receiving company — is what later makes it possible to aggregate outcomes by source and isolate each one's own contribution. Without this anchoring to origin, a converted or lost request would stay anonymous, and there would be no way to know which source to credit or correct.
Traceability also serves to detect the ambiguous cases that would distort the measurement if ignored. The same contact surfaced by two different sources reveals an attributable duplicate: it's clear which of the two submitted it first, and which reintroduced a request already in circulation. Likewise, a need captured outside the declared zone or category ties unambiguously to the source that misrouted it. This ability to trace the chain precisely, request by request, is the first condition of an honest assessment: you only measure well what you can tie back to its starting point.
The signals gathered from receiving companies
Measuring a source's return draws first and foremost on the signals receiving companies send back after attempting contact. These returns are about concrete facts: was the request reachable, did it genuinely match the announced category and zone, did it lead to an appointment, or did it prompt a complaint over a duplicate, wrong contact details, or a need outside scope? It's these field observations, accumulated request after request and tied back to their source of origin, that make up the raw material of the measurement.
Signal still has to be separated from noise. A single complaint isn't a verdict: what carries meaning is the repetition of the same problem across several requests from the same source. For fairness, the platform also accounts for the receiving company's own behaviour — a request left untreated for several days can't be blamed on the source, since the freshness lost is on the recipient, not the supplier. This symmetry is essential: you can only honestly measure a source's quality by neutralising the factors that don't depend on it.
From signal to score: the source rating that evolves
The signals gathered condense into a source rating, which isn't a label pinned on once and for all but a living indicator. This rating is computed on a moving average: recent requests weigh more than older ones, so a recent improvement or deterioration shows through fairly quickly without the whole history being erased. The platform also requires a minimum volume before issuing a first stabilised opinion, so as not to freeze a judgement on too few requests — the concrete expression of the principle stated above: a source is judged on a series, not a snapshot.
This rating has direct, reversible consequences. A source whose quality holds keeps, or even widens, its place in the distribution queue; a source that deteriorates sees its flow downgraded until its returns improve again; a source that recovers gradually regains ground. These adjustments follow the same rules for every source, with no regard to seniority or commercial relationship with the operator. It's this ongoing assessment — the permanent arbitration described in the overview dossier — that makes a source rating reflect actual behaviour over time, rather than a reputation earned once and never re-examined.
What measuring over time changes for both sides
Measuring a source's return over time aligns the interests of both sides of the market. For a source, ongoing assessment rewards consistency over one-off volume: it becomes more worthwhile to submit steadily good requests, month after month, than to flood the platform with one large mediocre batch followed by disengagement. A source that knows its quality is watched over a rolling window has predictable visibility on its place in the queue, and a structural reason to preserve what makes it valuable over time.
For receiving companies, this measurement keeps the average quality of circulating requests high, since a source that slips is spotted and downgraded before it can durably distribute weak contacts. It also lets them compare sources on a solid basis — a track record of behaviour — rather than on the impression left by a single delivery. The other dossiers in this series detail the neighbouring mechanisms: how the marketplace works overall, lead scoring, traceability under the FADP, and how to compare providers without relying on advertised volume alone.