True. But separating "influence" from "causation" is exactly where most attribution models fall apart - especially event-sourced pipelines.
Everyone parrots that line like it's profound, but in practice, pipeline attribution is a mess of overlapping touchpoints and arbitrary last-click or multi-touch models. Event-sourced data sounds clean until you realise most events are just vanity signals. A webinar registration doesn't cause a deal, it's correlation at best.
i've seen setups where six "influencing" events lead to zero closed revenue, yet the pipeline numbers get pumped into board decks as if they mean something. If you're not tying every event back to a concrete closed-won or closed-lost timestamp with a clear attribution window, you're just measuring noise.
Honestly, the more granular the event pipeline, the more it inflates perceived influence. It's like measuring how many times someone walked past your shop and calling it foot traffic that converts. Cause and effect require isolation - good luck getting that from a CRM that tracks a PDF download as a "pipeline driver."