Why I Built Metrone

The way people interact with products has changed. A potential customer might land on your website, start a chat conversation, and later talk to a voice AI agent — all as part of the same journey. Each of those touchpoints has its own analytics tool, its own dashboard, its own metrics. But the funnel is one funnel. That gap is why I built Metrone.

Great tools, separate worlds

There are excellent tools for each piece of the picture. Your voice AI provider gives you call duration, intent resolution, sentiment analysis. Platforms like LangFuse offer deep observability into LLM chains and agent behavior. Your web analytics shows traffic, conversions, and referral sources. Each one is good at what it does.

The challenge comes when your customer's journey crosses all of them. A lead visits your pricing page, asks a question through your chatbot, and later calls your AI voice agent to complete a purchase. That's one funnel — but today it lives in three separate dashboards with no shared context. You can see each step, but not the path.

Metrone exists to be that shared layer. Not to replace the specialized tools — they're valuable — but to give you one place where web traffic, chat interactions, and voice AI calls all feed into the same pipeline, the same funnel, the same dashboard.

Traffic is the right abstraction

When I started thinking about what Metrone should actually measure, I kept coming back to a simple idea: traffic.

Not "product usage" in the sense of tracking what each individual does inside your app. Not user-level behavioral profiles. Just traffic — the flow of events across every channel where people interact with your product. A page view is traffic. A chat message is traffic. A voice call is traffic. They're all signals in the same stream.

The questions you want answered are traffic questions: How many people visited this week? Where did they come from? Which channels are converting? Did the campaign we launched drive more calls, more chats, or more page views? How does the voice AI funnel compare to the web funnel?

None of these require knowing who each person is. They require knowing what happened — which events occurred, through which channels, and how they connect. The event is the unit of insight, not the identity.

Why the multi-channel funnel matters now

A few years ago, your traffic funnel was mostly web. Someone found you through search, visited a few pages, maybe submitted a form. The entire journey happened in a browser.

That's no longer the case for a growing number of products. AI has added new communication channels — voice agents that answer phone calls, chatbots that handle first-touch conversations, copilots embedded in workflows. These aren't replacing the web; they're extending the funnel. A lead might discover you through a blog post, ask a question via chat, and close through a voice conversation. The journey spans channels.

Measuring each channel in isolation gives you a fragmented view. Your web analytics says traffic is up. Your voice platform says call volume is steady. Your chat tool says conversations are increasing. But are those the same people moving through different stages, or different people in parallel? Without a unified view, you can't tell — and that makes it hard to know which channels are actually driving outcomes.

Metrone was built specifically for this. Web events, chat events, and voice events all land in one data model. A conversion funnel can start with a page view and end with a voice call. You see the whole path, not just the segments.

Privacy as architecture, not policy

There's one more thing I cared about deeply when building Metrone: measuring all of this traffic without tracking identities.

Metrone has no cookies — not "cookies are optional," but the system has no cookie mechanism at all. IP addresses are hashed with a monthly-rotating salt at the network edge before anything reaches a database. The hash is one-way; when the salt rotates, continuity breaks. Sessions are random IDs in sessionStorage that disappear when the tab closes. There's no cross-session linking, no fingerprinting, no user profiles.

This isn't just a privacy stance — it's a practical benefit. No cookies means no consent banners. No personal data means GDPR, CCPA, and ePrivacy don't apply to the analytics data. No third-party scripts means ad blockers don't interfere. You get complete, accurate traffic data across every channel, and your legal team doesn't have to worry about it.

The tracking script loads from your own domain via a CNAME record, so you see 100% of your traffic — not the 60-70% that survives an ad blocker. For teams making decisions based on data, that completeness matters.

What you see in Metrone

In practice, here's what this looks like: your voice agent handled 340 calls this week, resolved 78% of intents, averaged 2 minutes 15 seconds per call, and the "billing inquiry" intent has a 45% escalation rate. That's displayed alongside your web data — 12,000 page views, 3.2% conversion rate, organic search driving 60% of traffic. One dashboard. One funnel view.

You can build conversion funnels that cross channels — a page visit leading to a chat session leading to a voice call leading to a signup. You can compare which entry channel produces the highest conversion rate. You can see whether a marketing campaign drove more web visits or more voice calls.

And everything in the dashboard is also available through a REST API, a real-time SSE stream, and an MCP server — so your AI agents and monitoring tools can query the same data programmatically. Humans read the dashboard; machines call the API. Same data, two interfaces.

The business model

Metrone charges a straightforward price based on event volume. There's no free tier — and that's intentional. When analytics is free, the business model usually involves monetizing the data itself. We didn't want that dynamic. You pay for the service, your data stays yours, and the incentive is aligned: we succeed when you get accurate traffic analytics, not when we collect more information.

There's a 14-day trial so you can see it working before you commit. After that, pricing scales with how much traffic you're measuring. Simple.

Who Metrone is a good fit for

Metrone works best for teams whose traffic funnel spans more than one channel. Specifically:

• You have a web presence plus voice AI, chatbots, or AI copilots — and you want one dashboard that shows how all those channels contribute to the same funnel.
• You work in a regulated industry — healthcare, finance, education, government — and you need analytics that doesn't collect personal data, so compliance is simpler from the start.
• You want accurate traffic numbers without losing 30-40% of your data to ad blockers and consent banners.
• You're building AI-native products and want your agents to have programmatic access to traffic data through APIs, SSE, or MCP.
• You'd rather have one unified analytics layer than stitch together separate dashboards for web, chat, and voice.

Every new AI channel you add — a voice agent, a chatbot, a copilot — extends your traffic funnel into territory that traditional analytics can't see. That gap doesn't shrink over time. It grows with every integration. Metrone was built to close it: one pipeline for every channel, complete traffic visibility, zero identity tracking. The products that understand their full funnel will outperform the ones that only see part of it.