Alternative Data

Preview

Non-financial signals — hiring, shipping, pricing, usage — and the discipline required to read them.

Alternative data is any observation about a business that does not come from its financial statements. It is attractive because it is timely and unattractive because it is almost always a proxy for the thing you actually want to know.

Surfaces#

SurfaceObservesProxies for
HiringPosting volume, role mix, seniority, geographyInvestment direction and internal priority
Product telemetryRelease cadence, changelog velocity, deprecationsEngineering capacity and strategic focus
PricingList price changes, tier restructuring, discount depthMargin pressure and competitive position
Web presenceTraffic mix, page changes, documentation growthDemand and go-to-market shifts
Supplier signalsProcurement notices, logistics filingsInput cost and capacity constraints

Why the proxy relationship matters#

A drop in job postings can mean a hiring freeze, a completed hiring round, a shift to contractors, a change in how postings are syndicated, or an artefact of the collector missing a source. The observation is the same in all five cases. Pimsy therefore reports the observation and the set of explanations consistent with it, and treats any single-explanation reading as a finding that requires corroboration.

  • Coverage changes look like signal — a collector adding or losing a source produces a step change that resembles a real one. Coverage metadata is reported alongside every series.
  • Survivorship — entities that stop being observable frequently disappear from the panel rather than being recorded as departed.
  • Reflexivity — once a signal is widely tracked, the observed party has an incentive to manage it.
  • Short history — most alternative datasets have too few cycles to establish whether a relationship is stable or coincidental.

Last updated 2026-09-08