How to Build a Cohort Analysis Dashboard

Blended metrics hide the truth.

Your average retention rate might be 35%, but that's masking the fact that users acquired in September had 45% retention while October users had 25% retention.

Cohort analysis reveals these patterns by grouping users based on when they installed (or which channel they came from, or which campaign converted them) and tracking their behavior over time.

The result is a clearer picture of what's actually working in your user acquisition strategy.

Here's how to build a cohort analysis dashboard that helps you make better decisions.

What is Cohort Analysis?

A cohort is a group of users who share a common characteristic.

The most common cohort type is install date cohort: all users who installed your app on the same day, week, or month.

But you can also create cohorts based on:

• Acquisition channel (Meta, Google, TikTok, ASA)
• Campaign (Summer Sale, Q4 Push, Brand Campaign)
• Geography (US, UK, Japan, Brazil)
• First action (completed onboarding, skipped tutorial)

Cohort analysis tracks how these groups perform over time on key metrics like retention, revenue, and engagement.

Why Cohort Analysis Matters

Blended metrics are useful for high-level trends, but they mask important details.

Example:

Your overall Day 7 retention is 30%. Looks stable.

But cohort analysis reveals:

• January cohort: 40% D7 retention
• February cohort: 35% D7 retention
• March cohort: 20% D7 retention

Retention is declining fast. You wouldn't see this in a blended number until it's too late.

Cohort analysis lets you spot problems early and identify what's driving success.

Key Metrics to Track by Cohort

1. Retention Curves

Track what percentage of each cohort remains active over time.

Key retention windows:

• Day 1, Day 7, Day 30
• Week 1, Week 4, Week 8
• Month 1, Month 3, Month 6

Why it matters:

Retention curves show you how sticky your app is and whether recent product changes improved or hurt engagement.

2. LTV Over Time

Track cumulative revenue per cohort as they age.

Example:

Cohort	Day 7 LTV	Day 30 LTV	Day 90 LTV
Jan 2025	$2.50	$8.00	$18.00
Feb 2025	$3.00	$9.50	$21.00
Mar 2025	$2.00	$6.50	—

This shows:

• February cohort is outperforming January
• March cohort is trending lower (investigate why)

3. ROAS by Cohort

Track return on ad spend for each cohort over time.

Why it matters:

Some campaigns take longer to become profitable. A cohort with 50% Day 7 ROAS might hit 200% by Day 30.

Cohort-level ROAS helps you avoid cutting campaigns prematurely.

4. Payback Period

Track when each cohort recovers its CAC.

Example:

• Jan cohort: 3.2-month payback
• Feb cohort: 4.5-month payback
• Mar cohort: 2.8-month payback

This helps you identify which channels or campaigns have the fastest payback.

5. Engagement Metrics

Track sessions per user, DAU/MAU ratio, or feature usage by cohort.

High engagement correlates with retention and LTV. Cohorts with strong engagement in Week 1 typically have 2-3x higher LTV.

How to Structure Your Cohort Dashboard

A good cohort dashboard has three layers:

Layer 1: High-Level Overview

Show blended metrics across all cohorts:

• Total users acquired (by week or month)
• Average retention (D1, D7, D30)
• Average LTV
• Blended ROAS

This gives you a quick health check.

Layer 2: Cohort Performance Table

Show metrics broken down by cohort:

Cohort	Installs	D1 Ret	D7 Ret	D30 Ret	D30 LTV	CAC	Payback
Week 1	10,000	42%	28%	18%	$12.00	$8	3.2 mo
Week 2	12,000	40%	25%	16%	$10.50	$9	4.1 mo
Week 3	15,000	45%	32%	22%	$14.00	$7	2.5 mo

This shows trends over time and identifies high-performing cohorts.

Layer 3: Channel/Campaign Drilldown

Break down cohorts by acquisition source:

Meta Cohorts:

Cohort	D7 Ret	D30 LTV	ROAS
Jan	35%	$15	180%
Feb	32%	$13	150%
Mar	38%	$17	210%

Apple Search Ads Cohorts:

Cohort	D7 Ret	D30 LTV	ROAS
Jan	48%	$22	320%
Feb	45%	$20	280%
Mar	50%	$25	380%

This reveals which channels are delivering the highest quality users.

Tools for Building Cohort Dashboards

Built-In Analytics Platforms

Amplitude, Mixpanel, Adjust:

These platforms have native cohort analysis features. You can build retention curves, LTV charts, and custom cohort reports without SQL.

Pros:

• Easy to set up
• Pre-built templates
• Real-time updates

Cons:

• Limited customization
• Can be expensive at scale

Custom Dashboards

Looker, Tableau, Mode Analytics, Google Data Studio:

These BI tools let you build fully custom dashboards using SQL queries.

Pros:

• Complete control over metrics and visualizations
• Can combine data from multiple sources
• More cost-effective for large teams

Cons:

• Requires SQL knowledge
• More setup time

Spreadsheet Approach (for startups)

Export cohort data to Google Sheets or Excel and build pivot tables.

Pros:

• No cost
• Fast to set up

Cons:

• Manual updates
• Doesn't scale

Step-by-Step: Building a Basic Cohort Dashboard

Step 1: Define Your Cohorts

Decide how to group users. Most teams start with weekly install cohorts.

Example cohort definition:

• Cohort = all users who installed between Monday 00:00 and Sunday 23:59

Step 2: Choose Metrics

Pick 3-5 key metrics to track:

• Retention (D1, D7, D30)
• LTV (D7, D30, D90)
• ROAS (D7, D30)

Don't overload your dashboard. Focus on what drives decisions.

Step 3: Set Up Data Queries

Use your analytics platform or SQL to aggregate data by cohort.

Example SQL query (simplified):

SELECT
  DATE_TRUNC('week', install_date) AS cohort_week,
  COUNT(DISTINCT user_id) AS total_users,
  COUNT(DISTINCT CASE WHEN day_1_active = TRUE THEN user_id END) / COUNT(DISTINCT user_id) AS d1_retention,
  AVG(revenue_d30) AS avg_ltv_d30
FROM user_cohorts
GROUP BY cohort_week
ORDER BY cohort_week DESC

Step 4: Build Visualizations

Create charts showing:

• Retention curves (line chart with each cohort as a separate line)
• LTV over time (stacked area chart or line chart)
• Cohort comparison table (heatmap or table with conditional formatting)

Step 5: Automate Updates

Schedule daily or weekly refreshes so your team doesn't need to manually pull data.

Most BI tools support automated refreshes. Set it up once, and your dashboard stays current.

Common Cohort Analysis Patterns

Pattern 1: Retention Decay

Early cohorts had high retention, recent cohorts have lower retention.

Diagnosis: Product changes, seasonality, or lower-quality traffic.

Action: Investigate what changed. Test reverting recent product updates.

Pattern 2: LTV Growth

Recent cohorts have higher LTV than older cohorts.

Diagnosis: Better monetization, improved onboarding, or higher-quality users.

Action: Double down on what's working. Allocate more budget to high-performing channels.

Pattern 3: Channel Divergence

One channel (e.g., Apple Search Ads) consistently outperforms others.

Diagnosis: Higher-intent users from that channel.

Action: Shift budget toward high-performing channels. Investigate why others underperform.

Pattern 4: Seasonal Spikes

Cohorts acquired in Q4 have higher retention and LTV.

Diagnosis: Seasonal behavior (holidays, New Year's resolutions).

Action: Plan for seasonality. Increase spend during high-value periods.

Mistakes to Avoid

1. Too Many Cohorts

Tracking 50 different cohorts creates noise. Start with install date, then layer in channel and campaign.

2. Not Waiting for Maturity

Day 30 LTV isn't accurate until the cohort is at least 30 days old. Don't make decisions on incomplete data.

3. Ignoring Sample Size

A cohort with 100 users isn't statistically significant. Focus on cohorts with 1,000+ users for reliable insights.

4. Comparing Incomparable Cohorts

Don't compare a 2-week-old cohort's Day 7 metrics to a 6-month-old cohort's Day 180 metrics. Compare like-for-like.

Key Takeaways

• Cohort analysis reveals patterns blended metrics hide
• Track retention, LTV, ROAS, and payback by cohort
• Start with install date cohorts, then layer in channel and campaign
• Use built-in tools (Amplitude, Mixpanel) or custom dashboards (Looker, Tableau)
• Focus on 3-5 key metrics; don't overload your dashboard
• Automate updates to keep data current

FAQs

What is cohort analysis?

Cohort analysis groups users by shared characteristics (like install date or acquisition channel) and tracks their behavior over time. It reveals patterns blended metrics hide, like whether September users retained better than October users.

What metrics should I track in a cohort dashboard?

Key metrics include retention (Day 1, 7, 30), LTV over time, ROAS by cohort, payback period, and revenue per cohort. Track these by install date, channel, campaign, and geography.

What tools can I use to build a cohort dashboard?

Popular options include Amplitude, Mixpanel, and Adjust for built-in cohort reports. For custom dashboards, use Looker, Tableau, Mode Analytics, or Google Data Studio with SQL queries.

How often should I review cohort data?

Weekly reviews for high-level trends. Monthly deep dives for strategic decisions. Daily monitoring during major campaigns or product launches.

What's the minimum cohort size for reliable analysis?

Aim for at least 1,000 users per cohort. Smaller cohorts can show misleading patterns due to statistical noise.

---

Cohort analysis turns user acquisition from guesswork into science. Build your dashboard once, and you'll have clarity on what's actually driving growth.