# Years ago a leader told me “I measure performance by the number of lines of code shipped.” Today a chatbot would be his best performer...but — retained record

> Retained reddit record from reddit:Entrepreneur.

Canonical: https://highsignal.app/data/records/7fca7b96dd8ce9be

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:Entrepreneur)  published 2026-11-18 17:29 ingested 2026-11-19 02:36

[original source ↗](https://www.reddit.com/r/Entrepreneur/comments/1xzl6ku/years_ago_a_leader_told_me_i_measure_performance/)

## retained text



```
Funny, but I see a lot of teams making the same mistake right now. They just changed the metric: from code written to features shipped and token usage.

Here’s the problem:

1. They measure input: tokens, AI usage, adoption of tools. (Measuring input isn’t always bad, but here it creates a cobra effect. People optimize for usage, not results.)
2. Then they measure output, so teams produce more deliverables.
3. The dashboard is green, everyone is busy, leadership is happy.

A quarter later the board asks: did revenue go up? Did churn go down? Did adoption increase?

Nobody can answer.

MIT’s 2025 report on generative AI found that around 95% of company pilots showed no measurable P&L impact. The methodology has been debated, but it matches what I see with clients.

The root cause is usually simple: nobody showed the team how a deliverable connects to a business result.

What we use to fix this is impact mapping (Gojko Adzic’s framework).   
You build it backwards:

* **Why:** the business goal (e.g. reduce churn by 10% in Q2)
* **Who:** the people whose behavior affects that goal (users, admins, support, sales)
* **How:** the behavior change you need from them (admins finish onboarding in the first week)
* **What:** only then, the deliverables that could cause that change

If a feature doesn’t trace back up the map to a goal, it’s activity, not progress. It also gives you a real metric to track instead of tokens or tickets closed.

Curious how others handle this. If your team is using AI heavily, what are you actually measuring to know it’s paying off?
```



document fetched 2026-11-19 00:25 · canonical https://reddit.com/r/Entrepreneur/comments/1xzl6ku/years_ago_a_leader_told_me_i_measure_performance
