---
title: "Engineering metrics: what to measure and why | QMates"
source: https://qmates.tech/en/services/engineering-transformation/software-delivery-assessment/engineering-metrics
description: "You can’t tell whether your team is improving. Which engineering metrics help you decide, and why story points and ticket counts tell you so little."
updated: 2026-11-21
---

Engineering metrics

# We can’t tell whether the team is improving, or where the time goes

You count what is easy to count, such as story points and closed tickets. Those numbers earn the trust of neither management nor the team, and without credible numbers engineering decisions stay opinions, so nobody invests in measuring better. Engineering metrics stay stuck for two reasons that add up. The data is scattered across different tools, and management and the team don’t share a yardstick.

## Does this sound familiar?

-   Ask ‘Is the team improving?’ and everyone answers in their own way, and nobody can prove it.
-   We count story points, velocity or closed tickets, and those numbers rise even when releases don’t speed up.
-   An investment in refactoring or tests struggles to get approved, because nobody can say what the previous one achieved.
-   Decisions about where to invest, whether to hire or whether to rebuild part of the system are made on instinct.
-   Anyone who proposes measuring meets suspicion from the team, who fear being judged individually.

## Where do engineering metrics get stuck?

You get there by degrees. The data stays scattered, you end up counting the activity a tool already shows, and those numbers give management and the team no common yardstick. At that point the investment in quality is left undefended.

1.  ### You count the team’s activity

    Story points, velocity and closed tickets say how busy the team is, and they inflate as soon as they become a target. The business needs to know how long a change takes to reach users and how often a release causes problems. That means DORA metrics and cycle time taken end to end, measured at the level of the flow of work and of teams, never of individuals.

    Feeds into

    -   

    [

    Handled by

    Software Delivery Assessment](https://qmates.tech/en/services/engineering-transformation/software-delivery-assessment)

2.  ### The data exists, scattered across different tools

    Commit history, pipelines, backlog and incidents each tell one piece of the flow, with different names and units. Until they are read together, the point where work waits stays invisible even to the people who work there every day.

    Feeds into

    -   

    [

    Handled by

    Software Delivery Assessment](https://qmates.tech/en/services/engineering-transformation/software-delivery-assessment)

3.  ### Management and the team have no common yardstick

    Management thinks in dates and costs, the team in complexity and risk, and every discussion about priorities becomes a negotiation. A shared measure of what a change costs gives business, product and engineering the same yardstick, and you build it by working together on the code.

    Feeds into

    -   

    [

    Handled by

    Embedded Teams](https://qmates.tech/en/services/engineering-transformation/embedded-teams)

4.  ### Refactoring and tests defended on instinct

    A refactoring or a piece of work on tests is decided on instinct and judged by feel, so at the next decision it loses to features. Knowing which areas of the code hold back the important changes, and measuring them before and after, lets you defend the investment with facts.

    [

    Handled by

    Technical Debt Assessment](https://qmates.tech/en/services/engineering-transformation/technical-debt-assessment)

- [The service that tackles it](https://qmates.tech/en/services/engineering-transformation/software-delivery-assessment)

## What changes when development can be measured

Management and the team look at the same few numbers and know where they start from. A discussion about where to invest begins from the point where work stops today, and an intervention is judged by comparing the measures from before and after.

The team uses the measures to choose for itself what to improve, because it knows they won’t be used against people. Requests for time on refactoring or tests arrive with the data to support them, and management accepts or postpones them with the cost in front of it.
