The hardest part of data science is keeping track of and measuring the right things.

And you do that with metrics.

Most data scientists don't know what makes a good metric.

In this thread, I'll teach you what makes a good metric📈

A good metric has four qualities; it's...
1) Comparative
2) Understandable
3) A ratio or rate
4) Changes in behaviour

1) A good metric is comparative

Comparing a metric to other periods, groups of users or competitors helps you understand which way things are moving.

2) A good metric is understandable

If people can't remember and discuss it, it's much harder to turn a change in data into a shift in culture.

3) A good metric is a ratio or rate

Ratios are inherently comparative.

If you compare a daily metric to the same metric over a month, you'll see whether you're looking at a sudden spike or a long-term trend.

4) A good metric Changes Behaviour

This is by far the essential criterion for a metric.

What will you do differently based on changes in the metric?

Accounting metrics like daily sales revenue need to make your predictions more accurate when entered into your spreadsheet.

Experimental metrics like the results of a test help you optimize the product, pricing, or market.

Changes in these metrics will significantly change your behaviour.

Agree on what the change will be before you collect the data.

Metrics often come in pairs.

Conversion rate (the percentage of people who buy something) is tied to time-to-purchase (how long it takes someone to buy something). Together they tell you a lot about your cash flow.

The viral coefficient (the number of people a user successfully invites to your service) and viral cycle (how long it takes them to invite others) drive your adoption rate.

You'll notice these pairs as you explore the numbers that underpin your business.