luvv to helpDiscover the Best Free Online Tools
Topic 2 of 8

Acquisition Cohort Definition

Learn Acquisition Cohort Definition for free with explanations, exercises, and a quick test (for Marketing Analyst).

Published: December 22, 2025 | Updated: December 22, 2025

Table of Contents

Why this matters

Defining acquisition cohorts correctly is the foundation of trustworthy retention, LTV, and payback analyses. Marketing Analysts use cohorts to answer real questions like:

  • Which channels bring users who retain after 3 months?
  • How long until a cohort pays back its CAC?
  • Did our onboarding change improve activation for new cohorts?

Get the definition wrong, and every downstream metric can mislead decisions and budgets.

Concept explained simply

An acquisition cohort is a group of users clustered by when they first became your users—based on a chosen “acquisition event” (for example: first signup or first purchase). The cohort label is typically a date bucket like 2025-01-15 (daily), 2025-W02 (weekly), or 2025-01 (monthly).

Mental model

Think of school classes: the “Class of 2025” is everyone who started in the same year. A cohort is the “class” of users who started using your product in the same time bucket.

Core elements of a cohort definition

  • Acquisition event: the first qualifying event that makes someone “yours” (e.g., first signup, first purchase).
  • Identity key: user_id or customer_id used to group events to the same person/entity.
  • Date grain: day, week, or month of acquisition.
  • Timestamp source and timezone: which timestamp field and which timezone standard to use.
  • Channel attribution snapshot: the channel/tag you freeze at the time of acquisition.
  • De-duplication rule: first qualifying event only; ignore later ones.
  • Exclusions: remove test, internal, bots, or incomplete signups.
  • Backfill policy: how to handle late-arriving data or merged identities.
Typical choices and when to use them
  • Use signup as acquisition if your product is free or a trial and value starts at account creation.
  • Use first purchase as acquisition if you care about paying customers and CAC payback.
  • Use monthly grain for small volumes; weekly or daily for larger volumes or fast experiments.
  • Use a single canonical timezone (often product timezone or UTC) and document it.

Worked examples

Example 1: Mobile app (freemium)

Goal: Understand retention from first signup.

  • Acquisition event: first_signup
  • Identity: user_id
  • Grain: weekly (YYYY-Www)
  • Timezone: product default (e.g., UTC)
  • Channel: install_source at signup
  • Exclusions: internal testers; device_limit > 10 in 1 day (likely bots)

Interpretation: Cohort 2025-W02 includes all users whose first signup happened during week 2 of 2025.

Example 2: E-commerce (conversion-first)

Goal: LTV and CAC payback on paying customers.

  • Acquisition event: first_purchase (exclude canceled/refunded purchases within 24h)
  • Identity: customer_id
  • Grain: monthly
  • Timezone: store timezone
  • Channel: last_non_direct_click at first purchase
  • Exclusions: employees, fraud-flagged orders

Interpretation: Cohort 2025-01 includes customers whose first valid purchase occurred in Jan 2025.

Example 3: B2B SaaS (lead to opportunity)

Goal: Activation and expansion from marketing-sourced signups.

  • Acquisition event: first_verified_signup (email verified + workspace created)
  • Identity: account_id (primary), user_id (secondary)
  • Grain: weekly
  • Timezone: UTC standardized
  • Channel: UTM_source captured at signup, frozen
  • Exclusions: free trial signups without verification

Interpretation: Cohort 2025-W03 = first verified signups that week; later multi-seat users still belong to the signup cohort.

Steps to define acquisition cohorts

  1. Clarify the business question. Is the focus on adoption (signup) or monetization (purchase)?
  2. Pick the acquisition event. Choose the earliest meaningful event; define filters (e.g., verified, non-refund).
  3. Choose identity. Decide user_id vs customer_id/account_id; specify merge rules.
  4. Set cohort grain and timezone. Monthly for low volume; weekly/daily for high volume or experiments.
  5. Freeze channel at acquisition. Decide the attribution model and exactly which fields to snapshot.
  6. Define exclusions. Internal/test users, bots, invalid events; document detection rules.
  7. Document backfill and late data handling. For late events and id merges, specify how cohorts are recomputed.
  8. QA with samples. Manually verify random users to ensure correct cohort assignment.

Data you need (checklist)

  • Identity: user_id/customer_id/account_id
  • Event table with timestamps (signup, purchase, etc.)
  • Attribution fields at acquisition (source, medium, campaign)
  • Time zone decision and reproducible conversion logic
  • Exclusion flags (internal, bot, fraud)
  • Merge history (id_links) if identities can change

Edge cases and rules

  • Merged identities: If two user_ids later merge, use the earliest qualifying event across both; keep a reproducible merge rule.
  • Refunded first purchase: If your acquisition event is purchase, exclude or re-evaluate when a first purchase is fully refunded.
  • Backfilled events: Document whether historical cohorts will be recomputed nightly.
  • Time zone drift: If source systems store mixed timezones, standardize before bucketing.
  • Internal/bot traffic: Maintain a list of domains, IPs, or device heuristics to exclude.

Common mistakes and how to self-check

  • Mistake: Using any purchase as acquisition, not the first. Self-check: Verify that each user appears in exactly one cohort.
  • Mistake: Channel changing after acquisition. Self-check: Confirm channel fields are frozen at the first qualifying event.
  • Mistake: Timezone inconsistencies. Self-check: For a sample of users near midnight, confirm cohort buckets match the documented timezone.
  • Mistake: Missing exclusions. Self-check: Confirm internal and bot flags are filtered out before cohort assignment.
  • Mistake: Identity collisions. Self-check: For merged accounts, ensure the earlier event is the cohort anchor.

Exercises

Mirror of the interactive exercises below. Do them now before the quick test.

Exercise 1 — Define cohorts for three scenarios

For each scenario, choose acquisition event, identity, grain, timezone, exclusions, and channel snapshot. Write a short config and one-sentence rationale.

  1. Streaming app trial: users start on a 7-day free trial; value starts at watching content.
  2. Marketplace: sellers matter, not buyers; measure seller retention and LTV.
  3. Consumer fintech: focus on funded accounts (first successful deposit), not just signups.
Hints
  • Pick the earliest event that represents real value for your business model.
  • Monthly grain often suffices at small scale; weekly for rapid experimentation.
  • Freeze attribution at the first qualifying event.
Expected output shape
{
  "scenario": "...",
  "acq_event": "...",
  "identity": "...",
  "grain": "daily|weekly|monthly",
  "timezone": "UTC|ProductTZ",
  "channel_snapshot": "...",
  "exclusions": ["..."] ,
  "rationale": "..."
}

Exercise 2 — Assign cohorts from raw events

Given raw events below, assign each user to a cohort (monthly) using first_verified_signup. Timezone: UTC. Exclude internal emails (@company.test).

Raw rows
user_id, event, ts, email, channel
u1, signup, 2025-01-31T23:50:00Z, a@x.com, ads
u1, email_verified, 2025-02-01T00:02:00Z, a@x.com, ads
u2, signup, 2025-02-03T10:00:00Z, b@company.test, organic
u3, signup, 2025-02-10T09:00:00Z, c@y.com, referral
u3, email_verified, 2025-02-12T11:00:00Z, c@y.com, referral

Rule: first_verified_signup = first signup followed by email_verified within 72h.

Produce: user_id, cohort_month (YYYY-MM), channel_at_acq.

Expected output
u1, 2025-02, ads
u3, 2025-02, referral
  • Checklist: Did you pick one clear acquisition event?
  • Checklist: Did you ensure each user appears in only one cohort?
  • Checklist: Did you freeze channel at acquisition?
  • Checklist: Did you apply timezone and exclusions consistently?

Practical projects

  • Build a cohort assignment query: Create a reproducible SQL (or notebook) that outputs one row per user with cohort label, channel, and acquisition timestamp.
  • Cohort QA dashboard: For the latest three cohorts, show user counts by channel and outlier detection (e.g., sudden spikes).
  • Compare definitions: Run retention curves for signup vs first purchase cohorts and summarize differences in a one-pager.

Who this is for

  • Marketing Analysts measuring channel performance, retention, and LTV.
  • Product Analysts validating onboarding and activation.
  • Data-savvy marketers running acquisition experiments.

Prerequisites

  • Basic SQL or spreadsheet skills.
  • Understanding of marketing channels and attribution concepts.
  • Comfort with timestamps and timezones.

Learning path

  • Start here: Define acquisition cohorts (this lesson).
  • Next: Retention metrics by cohort (D1/D7/D30, churn rate).
  • Then: Revenue/LTV by cohort and CAC payback.
  • Advanced: Channel-mix modeling and cohort forecasting.

Next steps

  • Finalize a written cohort definition for your team (one-pager with rules and examples).
  • Implement the cohort assignment pipeline and schedule QA checks.
  • Share early insights and get feedback from marketing and product stakeholders.

Mini challenge

Your signup cohorts show a sudden +40% spike in a single day. In one paragraph, list 3 possible causes and 3 checks you would run to confirm or rule out each cause.

Quick test note

The quick test is available to everyone. Only logged-in users will have their progress saved.

Practice Exercises

2 exercises to complete

Instructions

For each scenario, choose acquisition event, identity key, cohort grain, timezone, exclusions, and channel snapshot. Provide a short config and one-sentence rationale.

  1. Streaming app trial: users start on a 7-day free trial; value starts at watching content.
  2. Marketplace: sellers matter, not buyers; measure seller retention and LTV.
  3. Consumer fintech: focus on funded accounts (first successful deposit), not just signups.

Output your answers as small JSON-like blocks.

Expected Output
{ "scenario": "Streaming app trial", "acq_event": "...", "identity": "...", "grain": "...", "timezone": "...", "channel_snapshot": "...", "exclusions": ["..."], "rationale": "..." } ... (repeat for other scenarios)

Acquisition Cohort Definition — Quick Test

Test your knowledge with 6 questions. Pass with 70% or higher.

6 questions70% to pass
Repeat Purchase And Reorder RatesRetention By Acquisition Channel

Have questions about Acquisition Cohort Definition?

AI Assistant

Ask questions about this tool