# EXOS Analytics vs. Commercial Analytics Tools

**Last updated**: 2026-03-22
**Author**: Joe Lanzone
**Purpose**: Comprehensive comparison of EXOS Analytics against 9 commercial tools across 12 dimensions. The argument: EXOS combines capabilities from all of these into one ClickHouse-native platform where ServiceLink owns the data.

---

## The Landscape

ServiceLink evaluated (or could evaluate) the following tools to solve different slices of the operational intelligence problem:

| Tool | Primary Category | Starting Price |
|------|-----------------|----------------|
| **PostHog** | Open-source product analytics + session replay | Free tier; Cloud at ~$450/mo for 10M events |
| **Amplitude** | Product analytics, cohorts, experimentation | $30K-$200K/yr enterprise |
| **Microsoft Clarity** | Free session replay + heatmaps | Free (data goes to Microsoft) |
| **Mixpanel** | Event analytics, funnels, retention | $28/mo (Growth); enterprise from ~$50K/yr |
| **Hotjar** | Heatmaps + session recording + surveys | $32/mo (Plus); Business from $80/mo |
| **LogRocket** | Session replay + error tracking + performance | $99/mo (Team); enterprise from ~$50K/yr |
| **Pendo** | In-app guidance + analytics + feedback | ~$50K-$150K/yr enterprise |
| **Celonis** | Process mining + execution management | $100K+/yr enterprise |
| **ClickHouse telco_marketing demo** | Reference implementation (CH + HyperDX + LibreChat + Langfuse) | Open source + ClickHouse Cloud costs |

EXOS Analytics replaces all of them with a single stack at ~$2K-$10K/yr depending on scale.

---

## Feature Comparison Matrix

### Legend

- **Y** = Feature exists and is production-quality
- **P** = Partial implementation (exists but has gaps)
- **N** = Not available
- **--** = Not applicable to this tool's category

| Capability | EXOS Analytics | PostHog | Amplitude | Clarity | Mixpanel | Hotjar | LogRocket | Pendo | Celonis | CH telco demo |
|-----------|---------------|---------|-----------|---------|----------|--------|-----------|-------|---------|--------------|
| **Session replay** | Y (rrweb + HyperDX) | Y | P (add-on) | Y | N | Y | Y | N | N | Y (HyperDX) |
| **Funnel analysis** | Y (`windowFunnel()`) | Y | Y | N | Y | P (basic) | P (basic) | Y | P | P |
| **Retention/cohort** | Y (matrix + dateDiff) | Y | Y | N | Y | N | N | Y | N | N |
| **Process mining** | Y (7 ClickHouse primitives) | N | N | N | N | N | N | N | Y | N |
| **Error tracking** | Y (grouped + timeline) | Y | N | N | N | N | Y | N | N | P |
| **Heatmaps** | N | P (toolbar) | N | Y | N | Y | N | N | N | N |
| **NL-to-SQL / AI query** | Y (Azure AI Foundry) | P (HogQL assistant) | N | N | N | N | N | N | P (Process AI) | Y (LibreChat + MCP) |
| **Raw SQL access** | Y (full ClickHouse SQL) | Y (HogQL) | N | N | N | N | N | N | P (PQL) | Y |
| **Real-time dashboard** | Y (10s refresh) | Y | P (minutes delay) | Y | P (minutes) | N | P | N | N | P |
| **Anomaly detection** | P (z-score SQL, no alerting) | P (basic) | Y | N | Y | N | Y | N | Y | N |
| **User paths** | Y (sliding window + transitions) | Y | Y | N | Y | N | P | Y | Y | N |
| **Operator/workforce analytics** | Y (native) | N | N | N | N | N | N | N | Y | N |
| **Conformance checking** | Y (`sequenceMatch()`) | N | N | N | N | N | N | N | Y | N |
| **Automation scoring** | Y (AI-assisted) | N | N | N | N | N | N | N | Y | N |
| **A/B testing** | N | Y | Y | N | N | N | N | N | N | N |
| **In-app guidance** | N | P (toolbar) | N | N | N | P (surveys) | N | Y | N | N |
| **Behavioral segmentation** | P (SQL-based) | Y | Y | N | Y | N | P | Y | N | N |
| **Saved dashboards** | N | Y | Y | Y | Y | Y | Y | Y | Y | N |
| **Alerting** | N | Y | Y | N | Y | N | Y | N | Y | N |

---

## Dimension-by-Dimension Analysis

### 1. Cost

| Tool | Annual Cost (10K visitors/day) | Per-Seat Fees | Data Retention |
|------|-------------------------------|---------------|----------------|
| **EXOS Analytics** | **~$3,120** ($260/mo) | **None** | 12 months hot, unlimited archive |
| PostHog Cloud | ~$5,400-$10,800 | None | 12 months |
| Amplitude Enterprise | $30,000-$200,000 | Per-seat | Varies by contract |
| Microsoft Clarity | Free | None | 30 days |
| Mixpanel Enterprise | $50,000+ | Per-seat | 12 months |
| Hotjar Business | ~$960-$3,840 | None | 12 months |
| LogRocket Enterprise | $50,000+ | Per-seat | Varies |
| Pendo | $50,000-$150,000 | Per-seat | 12 months |
| Celonis | $100,000+ | Per-seat | Varies |
| CH telco_marketing | ~$2,000-$5,000 (infra only) | None | Unlimited |

**EXOS advantage**: 10-50x cheaper than enterprise tools. No per-seat pricing means the entire ServiceLink organization can access analytics without budget negotiation per user. Cost scales with data volume (ClickHouse Cloud storage), not headcount.

### 2. Data Ownership

| Tool | Where Data Lives | Export Capability | Lock-in Risk |
|------|-----------------|-------------------|--------------|
| **EXOS Analytics** | **Your ClickHouse Cloud instance (Azure)** | **Full SQL access, any BI tool, Parquet export** | **None** |
| PostHog Cloud | PostHog's infra (or self-hosted) | API, batch export | Low (open source) |
| Amplitude | Amplitude's infra | Limited export API | High |
| Clarity | Microsoft's infra | None | High |
| Mixpanel | Mixpanel's infra | Data pipeline export | High |
| Hotjar | Hotjar's infra | None | High |
| LogRocket | LogRocket's infra | Limited API | High |
| Pendo | Pendo's infra | Data sync add-on | High |
| Celonis | Celonis Cloud or on-prem | Limited | High |
| CH telco_marketing | Your ClickHouse | Full SQL | None |

**EXOS advantage**: Data lives in ServiceLink's own ClickHouse Cloud instance. Query it from Grafana, Metabase, dbt, Python, or any JDBC/ODBC client. No vendor can cut off access or raise prices on your own data. If ClickHouse Cloud pricing changes, export to Parquet and migrate.

### 3. ClickHouse-Native

| Tool | Uses ClickHouse | Direct SQL Access | Leverages CH Analytics Functions |
|------|----------------|-------------------|--------------------------------|
| **EXOS Analytics** | **Yes (sole database)** | **Full ClickHouse SQL** | **Yes: windowFunnel, sequenceMatch, retention, quantile, etc.** |
| PostHog | Yes (underlying store, HogQL abstraction) | HogQL (ClickHouse dialect) | Partial |
| Amplitude | No (proprietary) | No | No |
| Clarity | No (proprietary) | No | No |
| Mixpanel | No (proprietary) | No | No |
| Hotjar | No (proprietary) | No | No |
| LogRocket | No (proprietary) | No | No |
| Pendo | No (proprietary) | No | No |
| Celonis | No (proprietary, PQL) | No | No |
| CH telco_marketing | Yes | Full ClickHouse SQL | Yes |

**EXOS advantage**: Direct access to ClickHouse's analytical arsenal. `windowFunnel()` for funnels, `sequenceMatch()` / `sequenceCount()` for conformance, `retention()` for cohorts, `lagInFrame()` for bottleneck detection, rolling z-scores for anomaly detection. No abstraction layer between you and the engine.

### 4. Process Mining

| Tool | Process Discovery | Conformance Checking | Bottleneck Detection | Variant Analysis | Automation Scoring |
|------|------------------|---------------------|---------------------|-----------------|-------------------|
| **EXOS Analytics** | **Y** | **Y** (`sequenceMatch`) | **Y** (`lagInFrame`) | **Y** (`groupArray`) | **Y** (AI-assisted) |
| PostHog | N | N | N | N | N |
| Amplitude | N | N | N | N | N |
| Clarity | N | N | N | N | N |
| Mixpanel | N | N | N | N | N |
| Hotjar | N | N | N | N | N |
| LogRocket | N | N | N | N | N |
| Pendo | N | N | N | N | N |
| **Celonis** | **Y** | **Y** | **Y** | **Y** | **Y** |
| CH telco_marketing | N | N | N | N | N |

**EXOS advantage over all except Celonis**: Process mining is EXOS's core differentiator against product analytics tools. None of them can answer "what is the happy path for order processing?" or "which operators deviate from the SOP?"

**EXOS advantage over Celonis**: Same analytical capabilities at 1/50th the cost. Celonis charges $100K+/yr. EXOS runs the same `sequenceMatch`, variant extraction, and bottleneck detection on a $170/mo ClickHouse instance. Plus, EXOS adds session replay and product analytics that Celonis does not offer.

### 5. Session Replay

| Tool | Replay Quality | Trace Correlation | Frustration Signals | PII Controls |
|------|---------------|-------------------|--------------------|--------------|
| **EXOS Analytics** | **Full rrweb DOM** | **Y (OTel trace overlay)** | **P** (dead/rage click tracking via custom JS) | **P** (configurable but currently disabled) |
| PostHog | Full rrweb DOM | Partial | Y (rage clicks) | Y |
| Amplitude | Basic (add-on) | N | N | Y |
| Clarity | Full DOM | N | Y (rage clicks, dead clicks, excessive scroll) | Y (auto-mask) |
| Mixpanel | N | N | N | N |
| Hotjar | Full DOM | N | P (basic) | Y |
| LogRocket | Full DOM | Y (error correlation) | Y (rage clicks) | Y |
| Pendo | N | N | N | N |
| Celonis | N | N | N | N |
| CH telco_marketing | Full rrweb DOM (HyperDX) | Partial | P | P |

**EXOS advantage**: The only tool that correlates session replay with full OTel backend traces. Click a slow network request in the replay and see the corresponding backend span with its database queries, service calls, and error details. LogRocket does error correlation but not full distributed trace overlay.

### 6. AI / Natural Language Query

| Tool | NL Query | Schema-Aware AI | Process Analysis AI | Anomaly Explanation |
|------|----------|-----------------|--------------------|--------------------|
| **EXOS Analytics** | **Y** (NL-to-ClickHouse SQL) | **Y** (full schema in prompt) | **Y** (auto-summarize process variants) | **Y** (dimensional RCA) |
| PostHog | P (HogQL assistant) | P | N | N |
| Amplitude | N | N | N | N |
| Clarity | P (Copilot integration) | N | N | N |
| Mixpanel | P (Spark AI) | P | N | N |
| Hotjar | P (AI surveys) | N | N | N |
| LogRocket | N | N | N | N |
| Pendo | P (AI guides) | N | N | N |
| Celonis | P (Process AI) | P | Y | P |
| CH telco_marketing | Y (LibreChat + MCP) | Y | N | N |

**EXOS advantage**: The AI assistant knows the full ClickHouse schema, the Operational Intelligence Playbook's SQL templates, and the domain context (ServiceLink's order lifecycle, LOBs, operator roles). It generates ClickHouse-native SQL, not a proprietary query language. The process analysis AI auto-generates executive summaries from variant and bottleneck data.

### 7. Customizability

| Tool | Custom Events | Custom Dashboards | Custom Queries | API Access | Source Available |
|------|--------------|-------------------|----------------|------------|-----------------|
| **EXOS Analytics** | **Y** (any OTel span) | **N** (planned) | **Y** (full SQL) | **Y** (18 REST endpoints) | **Y** (all source in repo) |
| PostHog | Y | Y | Y (HogQL) | Y | Y (open source) |
| Amplitude | Y | Y | P (limited SQL) | Y | N |
| Clarity | P | N | N | P | N |
| Mixpanel | Y | Y | P (JQL) | Y | N |
| Hotjar | P | N | N | P | N |
| LogRocket | Y | P | N | Y | N |
| Pendo | Y | Y | N | Y | N |
| Celonis | Y | Y | P (PQL) | Y | N |
| CH telco_marketing | Y | N | Y (full SQL) | Y | Y (open source) |

**EXOS advantage**: Full source code ownership. Every HTML page, every API handler, every SQL query is in the repo. Need a new analysis type? Add a file. Need to change the funnel algorithm? Edit `functions/api/funnels.js`. No vendor support tickets, no feature request backlogs, no "it's on the roadmap."

### 8. ServiceLink Fit

| Dimension | EXOS Analytics | Best Commercial Alternative | Why EXOS Wins |
|-----------|---------------|---------------------------|---------------|
| Order lifecycle tracking | Native (process mining on OTel traces) | Celonis ($100K+/yr) | Same SQL, 1/50th cost |
| Operator performance | Native (workforce analytics) | Celonis | Same capabilities, integrated with session replay |
| Multi-LOB segmentation | Native (`SpanAttributes['lob']`) | Amplitude ($30K+/yr) | No per-seat fees, SQL-level flexibility |
| Internal tool analytics | Full stack | PostHog (free tier) | Process mining + replay + AI query in one |
| Compliance/audit trail | Full SQL audit + ClickHouse retention | Celonis | Data in your own Azure subscription |
| Cost at ServiceLink scale | ~$10K/yr | $250K+/yr (Celonis + Amplitude + FullStory) | 25x cost reduction |
| Azure ecosystem | ClickHouse Cloud on Azure, Azure AI Foundry | Mixed | Single cloud, single billing |

---

## The Consolidation Argument

Today, achieving EXOS Analytics' full capability set requires purchasing and integrating multiple tools:

```
Process Mining:        Celonis                    $100,000+/yr
Product Analytics:     Amplitude                   $50,000+/yr
Session Replay:        FullStory or LogRocket       $50,000+/yr
In-App Analytics:      Pendo                        $75,000+/yr
Error Tracking:        Sentry or Datadog            $20,000+/yr
BI Dashboards:         Looker or Tableau            $30,000+/yr
                                                  ──────────────
Total:                                            $325,000+/yr
Plus: integration work, data silos, context-switching between tools
```

EXOS Analytics consolidates all of this into:

```
ClickHouse Cloud:      $1,800-$6,000/yr
AKS Compute:           $840-$2,400/yr
Cloudflare:            $0 (free plan)
Azure AI Foundry:      $200-$1,000/yr
                       ──────────────
Total:                 $2,840-$9,400/yr
Plus: all data in one place, one query language, zero integration tax
```

**Cost reduction: 35-115x.**

The data lives in one place. A single SQL query can join a session replay event with a backend trace with a process mining variant with an operator performance metric. No tool on the market can do this because no tool spans all of these categories.

---

## When NOT to Use EXOS Analytics

Be honest about limitations:

| Scenario | Better Tool | Why |
|----------|-------------|-----|
| You need a plug-and-play solution tomorrow | PostHog | EXOS requires customization; PostHog works out of the box |
| You need A/B testing with statistical rigor | Amplitude Experiment | EXOS has no experimentation framework |
| You need heatmaps for UX design | Clarity or Hotjar | EXOS has no heatmap capability |
| You need in-app tooltips and guides | Pendo | EXOS is analytics-only, no in-app guidance |
| Your team cannot write SQL | Amplitude or Mixpanel | EXOS's AI query helps, but power users need SQL literacy |
| You need SOC 2 / HIPAA compliance today | PostHog Cloud or Amplitude | EXOS needs PII masking hardening |
| You have no ClickHouse expertise | Any commercial tool | EXOS is built on ClickHouse; you need someone who can maintain it |

---

## Comparison with ClickHouse telco_marketing Demo

ClickHouse's own `telco_marketing` demo in the `ClickHouse_Demos` repository uses a similar stack:

| Component | telco_marketing | EXOS Analytics |
|-----------|----------------|----------------|
| Database | ClickHouse | ClickHouse |
| Session replay | HyperDX | HyperDX |
| AI chat | LibreChat + MCP | Custom AI assistant (Azure AI Foundry) |
| Observability | Langfuse | Direct ClickHouse queries |
| Frontend | React (Vite) | Vanilla HTML + Tailwind |
| Process mining | No | Yes (7 ClickHouse primitives) |
| Operator analytics | No | Yes |
| Conformance checking | No | Yes |
| Automation scoring | No | Yes |
| Product analytics (funnels, retention, paths) | Basic | Full implementation |

EXOS Analytics extends the telco_marketing pattern with the process mining and operational intelligence capabilities that ServiceLink needs. See `CLICKSTACK_ALIGNMENT.md` for details on how EXOS aligns with ClickHouse's broader ClickStack initiative.

---

## Related Documentation

| Document | Covers |
|----------|--------|
| `vs-amplitude.md` | Deep-dive PM evaluation: 33 features compared, code-level gap analysis |
| `vs-fullstory.md` | Deep-dive engineering evaluation: session replay, error tracking, 10 code recommendations |
| `ARCHITECTURE.md` | System architecture, data flow, component details |
| `PRODUCT_ANALYTICS.md` | How to build PostHog/Amplitude-class analytics on ClickHouse |
| `CLICKSTACK_ALIGNMENT.md` | Alignment with ClickHouse's ClickStack initiative |