Data & Analytics

Welcome to Data & Analytics at Kaern Schools. I'm Ada, your tutor. I've spent years turning raw event logs and messy spreadsheets into decisions that actually moved companies — and that's exactly what I want for you.

This course is for founders building Kaern startups. We'll use real data from cycleX throughout — device telemetry from the field, Shopify store analytics, and the KPI dashboards we live by — so nothing here is theoretical. By the end you will be able to:

• Choose the few metrics that actually matter for your stage and instrument them honestly.
• Collect, model, and store data so it can answer questions later.
• Write SQL to interrogate your own data without waiting on anyone.
• Build dashboards and visualizations people trust and act on.
• Design and read A/B experiments without fooling yourself.
• Make decisions under uncertainty — and know when not to.

Treat this like a studio, not a lecture hall. Bring your company's numbers. We'll get our hands dirty.

Module 1: Metrics That Matter

Learning objectives

• Distinguish vanity metrics from actionable metrics.
• Define a North Star Metric and supporting input metrics for your startup.
• Apply the AARRR (pirate metrics) funnel to your own product.
• Recognize when a metric is being gamed or misread.

Lesson 1.1 — The one number you'd save in a fire

Teaching script. Imagine your whole dashboard burns down and you can save a single number. Which one? Most founders freeze — they've been collecting dozens of metrics and never asked which one is the business. At cycleX our early dashboard proudly showed total app downloads. It climbed beautifully. Meanwhile, the number that paid rent — weekly active charging sessions per device — was flat. Downloads were a vanity metric: big, flattering, and disconnected from value delivered. Your North Star Metric is the one that, when it goes up, almost always means customers got real value and the business got healthier. It's a leading proxy for revenue, not revenue itself. A good North Star is hard to game by accident and impossible to game on purpose without actually helping customers. For cycleX we landed on paid sessions completed per active device per week — it bundles adoption, reliability, and willingness to pay into one honest signal. Everything else is an input feeding it. So: if your dashboard burned down tonight, which single number would tell you whether your company is alive?

Worked example. cycleX candidate metrics and verdicts:

Lesson 1.2 — The pirate funnel (AARRR)

Teaching script. Pirates say "AARRR," and so do growth people — Acquisition, Activation, Retention, Referral, Revenue. It's a map of a customer's whole life with you, and it's powerful because it forces you to find your weakest stage instead of pouring money into a leaky bucket. At cycleX we once spent heavily on Acquisition ads, proud of cheap installs, while our Activation — a user completing their first successful charge — sat at 28%. We were filling a bucket with a hole in the bottom. Fixing the onboarding flow lifted activation to 52% and did more for revenue than doubling the ad budget would have. The discipline is to measure each stage as a conversion rate, then attack the stage with the worst rate relative to its potential, not the one that feels exciting. Acquisition is seductive because it's visible; retention is where durable companies are actually built. Where do you suspect your own biggest leak is — and have you measured it, or are you guessing?

Worked example. cycleX funnel, last 30 days:

Weakest absolute drop: Activation→Retention. That's the fix target.

Hands-on exercise

Write down your product's AARRR funnel for the last 30 days. Fill in raw counts and step conversion rates. Then pick your North Star Metric and justify in two sentences why it resists gaming. Identify the single weakest conversion step.

Common mistakes

1. Tracking cumulative totals (total users ever) instead of active or rate metrics that can go down — cumulative numbers only ever rise and hide decay.
2. Picking a North Star that equals revenue — revenue lags too much to steer by; you need a leading proxy.
3. Optimizing the loudest stage (acquisition) instead of the weakest, wasting spend on a leaky funnel.

Check for understanding

1. Why is "total registered accounts" a vanity metric while "weekly active devices" is actionable?
2. Given the cycleX funnel above, which stage deserves attention first, and why?
3. Name one property a good North Star Metric must have to resist gaming.