Skip to main content

Updated Sep 30, 2026Verified against the product Sep 30, 2026

Run a Campaign for a Cohort

When many accounts share the same gap — the same unused feature, the same pricing worry, the same missing setup step — writing to them one at a time takes weeks and reads inconsistently. A campaign handles the group in one motion: find the cohort, write the message once, send it through the right channel, and measure whether the cohort moves. This guide covers where cohorts come from in FunnelStory, how to choose a channel, and how to tell whether it worked.

Step 1: Find the cohort​

A good cohort shares one problem that one message can address. FunnelStory gives you three ways to find one:

SourceWhat it gives youBest when
Plays — Focus Areas → PlaysA recommended group of accounts that share a behavior, with the evidence behind it, what the data can't see, and a drafted email sequenceYou want FunnelStory to find the pattern and draft the copy
Outcome Agents — an activated cohortThe accounts renewing soon that sit below the adoption level that renews (or expands) better, with a strategy to close the gapThe goal is a renewal or expansion outcome
AudiencesA saved segment you define with filters, refreshed dailyYou already know who you want to reach

Plays and Outcome Agents are enabled per workspace by FunnelStory, and the Plays tab is visible to Admins and Super Admins.

Before you write anything, read the cohort's evidence. On a play, What we are not sure about tells you what to check before sending — for example, whether some accounts already solved the problem another way.

Step 2: Choose the channel​

ChannelHow it sendsUse it for
Play email sequenceFunnelStory drafts a multi-email sequence (Day 0, Day 3, …) with an exit rule; you copy it into your own email toolMulti-step outreach from your team's usual sender
CampaignOne email to every user in a saved audience, sent through your SendGrid, AWS SES, Mailgun, or Postmark connection, with opens and clicks tracked per recipientA single announcement or nudge to a whole segment
An AI Agent per accountRuns once for each account, reads its context, and drafts a personal email as an Action Card — or enrolls the contact in a HubSpot sequenceHigh-value accounts where one generic email won't land

Pick one channel per cohort. The same accounts getting a play sequence, a campaign, and a CSM email in the same week reads as noise.

Step 3: Write it once, for the whole group​

  • Lead with their problem, not your feature. The cohort's evidence tells you what they share — say it in their words.
  • One ask per email. Book a session, turn on a setting, reply with a question.
  • Personalize the basics. Play sequences use a {{FirstName}} placeholder; campaigns support personalization variables in the subject and body.
  • Test before sending. In Campaigns, use Test to send yourself a copy. A published campaign sends immediately and can't be edited or recalled.

Step 4: Send, and respect the exit rule​

  • Play sequence: paste each email into your tool on its day. Stop sending to an account once it meets the play's Exit rule — typically, the behavior the play targets has changed. Click Mark as Done on the play when the sequence is running, so the Plays tab reflects what's in flight.
  • Campaign: choose the audience and an email connection, then Publish.
  • Agent: publish the agent once its drafts look right; see Design an Agent: Detect, Investigate, Act.

Step 5: Measure whether the cohort moved​

Opens and clicks tell you the message landed; they don't tell you the problem got solved. Measure both:

MeasureWhere
Delivery and engagementCampaigns → Recipients: sent, delivered, opened, and clicked per recipient
Behavior changeThe metric the cohort was built on — feature usage on Adoption → Account Metrics, or needle movers of that type resolving
OutcomeFor an Outcome Agent cohort, the activated report: accounts that reached the target adoption, and ARR protected

Give it the length of the sequence plus a few weeks before judging. If engagement was high but behavior didn't change, the message was heard and the ask was wrong; if engagement was low, fix the subject line, sender, or timing first.