Numbers & Decisions

Upload a spreadsheet and ask "what does this tell me" and you'll get a description: totals, averages, an observation that March was higher than February. Technically true, entirely useless.

Analysis comes from questions with a decision attached. "Which of these activities produced revenue relative to the hours I put in, and what should I stop doing?" "Where in this funnel am I losing the most people, and is that where I'm spending my attention?" "What in this data contradicts what I believe about my business?"

That last one is the highest-value prompt in this module. You have a story about your business. The numbers usually disagree with part of it, and you will not spot the part yourself because it's the part you're not looking at.

Then verify. Ask it to show the arithmetic. Models make arithmetic errors, especially across many rows, and a confident wrong number is worse than no number.

Analyze this data. I want decisions, not a description.

[attach or paste data]

CONTEXT: [what the business is, what you're trying to grow]
WHAT I BELIEVE: [your current theory about what's working]

Answer:
1. What in this data CONTRADICTS what I believe? Be specific.
2. What's producing results relative to the effort it takes?
3. What should I stop doing, based on what's here?
4. What's the single number I should be watching weekly that I'm probably not?
5. What can this data NOT tell me? What would I need to collect?

Show your arithmetic for any figure you calculate. If the sample is too small to support a conclusion, say so instead of drawing one.

Nobody abandons a tracking system because it was too simple. They abandon it because it asked for twenty fields and they missed a week and never went back.

So: five numbers, entered weekly, in one row. For most people building something on the side, that's conversations started, follow-ups sent, content published, new customers and hours worked. Adjust to your business, but keep it to five and keep leading indicators in the list. You can control conversations started; you cannot directly control revenue.

The value shows up around week six, when patterns emerge that you'd never have felt. Almost everyone discovers their output is far spikier than they believed, and that the good weeks correlate with something unglamorous they'd stopped doing deliberately.

Design a weekly tracking sheet I'll actually maintain.

MY BUSINESS: [what you do]
MY GOAL: [specific, with a number and a date]

Give me EXACTLY 5 metrics. Rules:
- At least 3 must be leading indicators I directly control
- Each must take under 30 seconds to fill in
- No metric requiring a tool I don't already use

For each: the column header, exactly how I count it, and what a bad week looks like.

Then write the 5-question Friday review I should run against it.

If a metric I'd want doesn't make the cut, tell me what you cut and why. Five means five.

A review that produces feelings is journaling. A review that produces a changed decision is management. The difference is whether it ends with something specific you'll do differently, written down.

Paste your five numbers, your last few weeks and what you did. Then ask for the pattern and the argument against your explanation, because you will have an explanation ready for every bad week, and it will usually be about circumstances rather than choices.

Fifteen minutes, same time weekly. The consistency matters more than the depth. A shallow review every Friday beats a deep one every quarter, because the correction happens while the week is still steerable.

Keep the reviews. Reading six of them in a row shows you something no single review can: whether you keep making the same correction and not acting on it.

Run my weekly review. Be direct, not encouraging.

THIS WEEK'S NUMBERS: [paste]
LAST 3 WEEKS: [paste]
WHAT I ACTUALLY DID: [briefly]
MY GOAL: [with number and date]

Give me:
1. The pattern across these weeks: what's real, not noise
2. My likely explanation for this week, and the argument against it
3. Am I on pace for my goal? Show the math.
4. ONE thing to do differently next week, specific enough to schedule
5. One thing I should stop doing

Don't congratulate me on effort. Don't soften a bad week. If I'm off pace, say the number I'd need to hit and whether it's realistic.