What granularity should cohorts use: daily, weekly, or monthly?

Monthly for most B2B SaaS (sufficient volume per cohort, clear signal, consistent with billing cycles). Weekly if you have more than 10,000 signups/month (useful granularity for detecting rapid changes post-launch or after a marketing shift). Daily is rarely useful except for very-high-traffic products or server-side A/B tests.

Cohort analysis vs funnel analysis: what's the difference?

Funnel = conversion measurement across a sequence of steps (visit → signup → activation → paid), typically over a short window. Cohort = measurement of a group's behavior over the long term (retention M+1, M+2…). Funnel answers "where am I losing users in the journey"; cohort answers "how do my users evolve over time." The two are complementary.

Glossary

Cohort analysis

Cohort analysis is an analytical method that groups users by a common characteristic — typically the month or week of signup, sometimes acquisition channel or product version used — to observe how their behavior (retention, revenue, usage) evolves over time.

Also known as

cohort analysis
cohort analysis
cohort analysis

The fundamental value: without cohorts, an aggregated metric (e.g., "overall retention 40%") can mask a rapid product deterioration. If the 2024-Q1 cohort retains 60% at M3 but the 2025-Q1 cohort retains only 30%, your product has a real problem — even though the global average, pulled up by older cohorts, looks stable. Cohorts are the only way to observe **changes over time with rigor**.

Standard analysis types: (1) **cohort retention** (how many users from the January cohort are still active at M+1, M+2… visualized as a triangle), (2) **cumulative revenue by cohort** (LTV per cohort, to compare monetary health), (3) **cohorts by channel** (do SEO leads retain better than ad leads?), (4) **cohorts by feature** (do users who used feature X in their first month have a higher LTV?). 2026 tools: **PostHog Cohorts** (open-source, EU), **Amplitude Cohorts**, **Mixpanel**, **GA4** (basic cohorts). In B2B SaaS, cohort analysis is at minimum a monthly discipline — observing drift over 6–12 months is crucial for detecting PMF issues early.

In the getchatsocial.com product

getchatsocial.com uses PostHog Cohorts to track the retention and activation of its own signup cohorts, and to identify which acquisition channels (organic, paid, referral, AEO citation) produce the best cohorts in terms of LTV.

FAQ

What granularity should cohorts use: daily, weekly, or monthly?
Monthly for most B2B SaaS (sufficient volume per cohort, clear signal, consistent with billing cycles). Weekly if you have more than 10,000 signups/month (useful granularity for detecting rapid changes post-launch or after a marketing shift). Daily is rarely useful except for very-high-traffic products or server-side A/B tests.
Cohort analysis vs funnel analysis: what's the difference?
Funnel = conversion measurement across a sequence of steps (visit → signup → activation → paid), typically over a short window. Cohort = measurement of a group's behavior over the long term (retention M+1, M+2…). Funnel answers "where am I losing users in the journey"; cohort answers "how do my users evolve over time." The two are complementary.