Amplitude Review: Product Analytics Platform (2026)

import ProsConsCard from ’../../components/ProsConsCard.astro’;

Amplitude is purpose-built for product teams — its core philosophy is behavioral analytics, understanding what users do and how those behaviors predict outcomes like retention, conversion, and revenue. Unlike Google Analytics (optimized for marketers and acquisition) or Mixpanel (similar positioning, different ergonomics), Amplitude specializes in helping product managers answer “what’s working and why?”

Core Analytics: Funnels, Retention, and Cohorts

Amplitude’s core charts cover the fundamental product analytics questions:

Funnel Analysis:

Example funnel: SaaS onboarding
Step 1: Account created
Step 2: Profile completed
Step 3: Integration connected
Step 4: First report generated
Step 5: Invited team member

Amplitude funnel capabilities:
- Overall conversion: 100% → 23% (across 30 days)
- Step-by-step dropout: Biggest drop is Step 3 (integration) at 45% loss
- Time-to-convert: Median 3.2 days from signup to first report
- Segment comparison: Free vs. paid users, US vs. international, 
  mobile vs. desktop — conversion rates side by side
- Funnel order: Strict vs. any order (did they complete steps 
  in sequence or in any order?)
- Conversion window: What's the right window? (7 days, 30 days?)

Retention Analysis:

N-day retention: "What % of users who did [event] on Day 0 
returned on Day N?"

Example: 
Day 0 (sign up): 100%
Day 1: 42%
Day 7: 28%
Day 14: 23%
Day 30: 19%

Bracket retention: "What % return in this time window?"
Week 1: 68%, Week 2: 54%, Week 4: 41%

Usage interval: "How often do users come back?"
Median return interval: 3.2 days

Retention by cohort:
- August 2025 cohort: D30 retention = 19%
- September 2025 cohort: D30 retention = 24% (after onboarding changes)
- October 2025 cohort: D30 retention = 27% (after product improvement)
→ Clear improvement trend visible across cohorts

Behavioral Cohorts:

This is Amplitude's differentiator from most analytics tools:

Cohort: "Users who completed onboarding within 24 hours"
Define: Did [Profile Completed] AND [Integration Connected] within 1440 minutes of signup

Cohort: "Power users" 
Define: Users who performed [Report Generated] 5+ times in past 30 days

Cohort: "Churned users"
Define: Users who were active 61-90 days ago, not active in past 30 days

What you can do with cohorts:
- Compare retention between cohorts (do onboarding completers retain better?)
- Export to Braze/Iterable/Intercom for targeted campaigns
- Compare product behavior (what else do power users do differently?)
- Size the impact before building (how many users would this affect?)

---

Amplitude AI

Amplitude’s natural language query feature lets non-technical users ask questions without building charts:

Natural language examples:

"What is the conversion rate from signup to first report in the last 30 days?"
→ Builds funnel chart automatically

"Show me retention for users who invited a team member vs. those who didn't"
→ Builds retention chart with comparison

"Which features do our highest-retention users use in their first week?"
→ Builds user property analysis

"What did churned users do differently than retained users in month 1?"
→ Behavioral comparison analysis

"What's the trend in DAU over the last 90 days broken down by plan type?"
→ Time series with segmentation

The AI doesn’t just answer questions — it builds the chart you’d have built manually, making the insights reproducible and shareable with the team.

---

Event Instrumentation

Getting full Amplitude value requires clean event tracking. This is where most teams either succeed or fail:

JavaScript SDK setup:

import * as amplitude from '@amplitude/analytics-browser';

// Initialize
amplitude.init('YOUR_API_KEY', {
  defaultTracking: {
    sessions: true,
    pageViews: true,
    formInteractions: true,
    fileDownloads: true,
  },
});

// Track events with user context
amplitude.setUserId(user.id);
amplitude.setUserProperties({
  plan: user.plan,        // 'free', 'growth', 'enterprise'
  company_size: user.companySize,
  signed_up_at: user.createdAt,
  is_admin: user.isAdmin,
});

// Custom events with properties
amplitude.track('Report Generated', {
  report_type: report.type,
  data_sources_count: report.sources.length,
  has_filters: report.filters.length > 0,
  export_enabled: report.hasExport,
  time_to_generate_ms: generateTime,
});

amplitude.track('Integration Connected', {
  integration_type: integration.provider,
  is_first_integration: user.integrations.length === 1,
  connected_via: source, // 'onboarding', 'settings', 'prompt'
});

Event taxonomy (the critical part):

Bad event taxonomy:
- "Button Clicked" (which button? everywhere?)
- "User Action" (meaningless)
- "page_view" + "pageView" + "Page View" (duplicates)
- Including PII in event properties (compliance risk)

Good event taxonomy:
Naming convention: [Noun] [Verb] (past tense)
  - "Report Generated"
  - "Integration Connected"  
  - "Team Member Invited"
  - "Subscription Upgraded"

Property naming: snake_case, consistent
  - report_type, not reportType or ReportType
  - is_first_time (boolean), not first_time_user (ambiguous)
  - Always include: created_at (timestamp), plan, user_role

What NOT to track:
  - Email addresses or names in event properties
  - PII that triggers GDPR/CCPA concerns
  - Every micro-interaction (noise overwhelms signal)

---

Session Replay

Amplitude’s session replay (included in Growth/Enterprise) connects behavioral data with actual user recordings:

How it works with analytics:

Workflow: funnel analysis → session replay

1. Build funnel: Signup → Profile → Integration → Report
2. Identify: 45% of users drop at "Integration Connected" step
3. Filter: Show me session replays of users who dropped at Integration step
4. Watch: See what users actually experience — where they get stuck
5. Fix: The error message during OAuth is misleading (now visible in replay)
6. Measure: Integration completion rate improves 18% after fix

Without session replay:
"45% drop at integration step" → guess at cause → A/B test → wait 2 weeks

With session replay:
"45% drop at integration step" → watch 10 sessions → root cause clear → fix today

---

Amplitude vs. Mixpanel vs. Google Analytics

Feature	Amplitude	Mixpanel	Google Analytics 4
Primary audience	Product teams	Product teams	Marketers
Behavioral cohorts	Best in class	Good	Limited
Funnel analysis	Excellent	Excellent	Basic
Retention analysis	Excellent	Good	Limited
Session replay	Yes (Growth+)	Yes (add-on)	Yes (built-in)
Natural language queries	Amplitude AI	No	No
Free tier	50K MTU	20M events	Full access
Data governance	Govern module	Strong	Limited
SQL access	Amplitude SQL	Yes	BigQuery export
Setup complexity	High	Medium	Low-Medium
Best for	Product analytics	Product analytics	Marketing analytics

---

Pricing

Amplitude’s pricing is event-volume-based and not publicly listed (custom quotes). Rough guidance:

Starter (Free):
- Up to 50,000 monthly tracked users
- Core charts: Funnel, retention, segmentation
- 2 seats

Growth (Custom):
- 100K-1M+ monthly tracked users
- Behavioral cohorts, predictive analytics
- Amplitude AI
- Session replay
- ~$500-5,000/month depending on volume

Enterprise (Custom):
- Unlimited scale
- Advanced governance
- SSO, RBAC
- Dedicated support
- $2,000-20,000+/month

The free tier is genuinely useful for early-stage products. The jump to paid can be significant — get a quote before assuming it fits your budget.

---

Who Amplitude Is For

Best for:

• Product managers who make data-driven roadmap decisions
• B2B SaaS and consumer apps with meaningful user behavior to track
• Teams with engineering resources to implement properly
• Growth teams running experiments and retention campaigns
• Companies willing to invest in a clean event taxonomy

Not ideal for:

• E-commerce focused on acquisition and attribution (use GA4 + Northbeam)
• Teams without engineering to implement SDK properly
• Startups too early to have meaningful behavioral data
• Teams who need simple dashboards without complexity

---

Bottom Line

Amplitude earns its 4.4/5 for being the best tool at what it does: understanding how user behavior connects to outcomes. The funnel analysis is the most intuitive in the market, behavioral cohorts unlock insights that generic analytics cannot, and Amplitude AI genuinely helps non-technical stakeholders get answers without building charts. The caveats are real — pricing scales quickly, implementation requires engineering effort, and getting full value requires committing to clean instrumentation. Teams that make that investment report Amplitude is one of the highest-ROI tools in their stack.