Deciding from your data

Getting out of gut-feel steering

A solo entrepreneur makes dozens of decisions a week. Without data, they make them on intuition — sometimes right, often biased. AI doesn't just produce and automate: it helps you see clearly in your numbers and choose with lucidity. This chapter covers two uses: reading your own data (analytics) and leaning on AI for research and decisions.

The principle: you only steer well what you measure. But measuring too much kills measurement. The challenge isn't to have a hundred indicators, it's to have the three or four that truly determine the survival and growth of your business.

Measuring your activity simply

Before analyzing, you must collect — without drowning yourself or betraying your users:

• Privacy-friendly web analytics: Plausible or Fathom (a few euros a month) give the essentials of traffic without invasive cookies; Google Analytics remains the free benchmark but heavier.
• Real behavior: Microsoft Clarity (free) records sessions and heatmaps — you see where visitors click, hesitate, and abandon. Valuable for understanding why it doesn't convert.
• Unified dashboards: lightweight tools aggregate your key numbers (sales, traffic, subscribers) into a single view, often fed automatically by the automations from the previous chapter.

Choose the minimum that answers your real questions. A one-page dashboard you look at every Monday beats ten reports no one opens.

AI as a data analyst

This is where AI changes the game for the non-specialist. You no longer need to know how to wrangle a complex spreadsheet: you converse with your data.

• Drop a CSV export of your sales into an assistant like ChatGPT (with data analysis) or Claude, and ask in plain language: "What's my most profitable product? What's the trend over three months? Where do I lose the most customers?"
• AI computes, spots trends, generates charts, and explains what it sees.
• You keep the judgment: AI describes the past, it doesn't know your context. A correlation isn't a cause, and an isolated figure can mislead.

This use democratizes analysis: an entrepreneur with no data training can query their numbers as they would an analyst — provided they verify the calculations on important decisions.

Researching and arbitrating with AI

Beyond your own numbers, AI speeds up the external research that feeds your decisions:

• Perplexity provides sourced answers with links, ideal for quick market research, competitive monitoring, or a factual question where traceability matters.
• Generalist assistants help structure a decision: list the options, weigh pros and cons, play devil's advocate, anticipate objections. They don't decide, but they broaden your field of vision and counter your blind spots.

Beware, though, of the false sense of certainty: a well-phrased answer isn't a true answer. For any committing decision, cross-check the sources.

The three indicators never to lose sight of

Whatever your project, keep these permanently in view:

1. Acquisition: how many new prospects/customers, and at what cost?
2. Conversion: what share of interest turns into revenue?
3. Retention: do your customers stay, return, recommend?

These three numbers tell the health of your business better than any cluttered dashboard. AI can compute and monitor them for you; it's up to you to act on the one that's off.

Summary

Deciding from your data no longer requires technical skill: lightweight tools like Plausible, Microsoft Clarity, or a unified dashboard collect the essentials, and AI plays the role of an analyst you talk to in plain language. Perplexity and assistants broaden research and structure arbitration. Keep few indicators but the right ones — acquisition, conversion, retention — and verify the AI on decisions that matter. You've now explored every territory of the stack. It's time to assemble it into a coherent system.