Intent Spikes vs Baseline: How to Read Them

Three pricing-page visits from one account look exciting until you learn that account visits your site weekly out of habit, in which case three is just Tuesday. Intent only becomes a signal when it deviates from that account's own baseline, which is why raw activity counts mislead so often. The funnel is dead and intent is public, but public intent is noisy, and the noise drowns the signal unless you normalize against what normal looks like for each account. Read the deviation, not the absolute number.

Building a baseline means storing each account's typical activity in your own warehouse or in HubSpot so you can compare this week to its usual rhythm. Koala and Warmly surface activity, but the judgment of whether it is elevated belongs to you, because only you know that this account always lurks and that one never did until yesterday. When marketing is observable, a spike is a measured departure from a stored baseline, not a vendor's opaque score you take on faith.

Not every deviation is worth a sales motion. A single anomalous day can be a curious analyst, a competitor, or a stray click, while a sustained lift across several days and multiple people from the same account is the shape of a buying committee waking up. Watching whether a spike persists separates a fluke from a trend, and it is the difference between burning a rep's time and catching a deal early. Many teams require two or three elevated days before they route an account to outbound.

Breadth matters as much as height. One person hitting your pricing page ten times is interest; four different people from one domain hitting pricing, security, and integration docs in a week is a committee evaluating you. RB2B, Leadfeeder, and Koala can attribute that breadth to the account, and the multi-person, multi-page pattern is a far stronger spike than raw volume from a single visitor. Score the spread of the signal across the buying committee, not just its size.

A spike has a half-life, so latency is everything. An account that surges today and gets a touch tomorrow is warm; the same account contacted in three weeks through a batch job is cold and indistinguishable from a list import. This is why reading signals in real time and acting while warm beats a perfectly scored quarterly pull. Wire spikes from Koala or Warmly to fire a workflow that alerts a rep and queues a Smartlead or Instantly sequence the same day.

Keep the action coordinated across allbound rather than letting one channel react alone. The same spike should warm up paid retargeting, prioritize the SDR queue, and tune content the account sees, all reading from the one shared signal layer. When inbound, outbound, paid, and content respond to a single observed deviation, the account feels a coherent motion instead of four disconnected ones. Owning that signal layer is what lets you orchestrate the response instead of renting fragments of it.