What it is
Signal-to-noise ratio is a borrowed engineering term: the proportion of a feed that carries real, actionable information against the proportion that is irrelevant, low-quality or simply repeats what is already known. Applied to BD, the 'signal' is a buying signal that genuinely points at a company about to hire and matches the desk's ICP. The 'noise' is everything else that gets flagged anyway: events on companies that will never be clients, duplicate mentions of the same funding round, or a signal so stale the hiring window has already closed.
A system can watch more sources and report more hits without raising its signal-to-noise ratio at all, and in fact the opposite usually happens. Coverage and precision are different variables: adding sources raises the volume, but only filtering and verification raise the proportion of that volume actually worth a consultant's attention. A feed that fires on every possible event has a poor ratio even if, somewhere inside it, every genuine signal is technically present.
Coverage raises the volume. Only filtering and corroboration raise the ratio that actually matters.
Why it matters
Signal-to-noise ratio is the metric every signal-led BD system is really tuned against, whether or not anyone calls it that. A consultant does not have unlimited attention, and a feed with a low ratio spends that attention on dead ends before it reaches the account that would actually convert. This is exactly the mechanism behind signal fatigue: once enough alerts turn out to be irrelevant or stale, the consultant stops trusting the whole feed, including the signals that were genuinely worth acting on.
The failure is easy to miss because it looks like success at first. More sources, more signal types, more alerts, all read as progress. But a desk measuring itself on volume alone has no way to tell a strong system from a noisy one until reply rates fall and mandates get missed. The ratio, not the raw count, is what tells you whether a signal-led approach is actually working.
How boilr handles it
boilr is built to keep that ratio high rather than chase volume for its own sake. Every one of the twelve buying-signal types it monitors across more than 10,000 sources is filtered against your ICP before it is even considered, so an event on a company outside your target shape never counts as a hit in the first place. On top of that filter, signal corroboration checks a signal against at least one independent source before it is trusted enough to become a task, which is what keeps a single ambiguous posting or a stale mention from diluting the feed.
The result is an inbox where nearly everything that reaches you is worth acting on, which is the entire point of signal-to-noise ratio as a design target rather than a vanity metric. Outcomes flow back into the Company Brain, which learns which signal types and accounts actually convert for your desk and sharpens the ratio further over time, instead of treating today's filtering as good enough forever.