Your DORA Metrics Look Great. So Why Does Shipping Still Feel Slow?

There is a strange moment that happens to a lot of engineering teams. The DORA dashboard is green. Deployment frequency is up, lead time looks respectable, change failure rate is low. And yet, if you ask the team how it feels to ship something, the honest answer is: slow. Sluggish. Like wading through treacle.

When the numbers and the felt experience disagree, the numbers are not lying. They are just not telling you the part that hurts.

DORA tells you the score, not the story

DORA metrics are deliberately high-level. They measure outcomes: how often you deploy, how long changes take overall, how often things break. That is exactly what makes them a good common language with the business — and exactly what makes them bad at locating a bottleneck. A healthy lead time averaged across a quarter can hide a review queue that adds two days to every single pull request.

Averages are comforting and a little dishonest. They smooth right over the waiting.

Most delay is waiting, not working

Here is the uncomfortable truth that appears the moment you break delivery into stages: most of a change's life is spent waiting, not being worked on. The industry term is flow efficiency — the share of total time that work is actually moving. Plenty of teams discover theirs is under fifteen percent. A pull request that takes five days end to end often had only a few hours of real work in it. The rest was queue.

Your developers feel that waiting acutely. The dashboard does not, because the deploy still happened and the metric still looks fine.

Where the time actually goes

When you split cycle time into stages — coding, waiting for review, review itself, merge — the culprit is almost never coding. It is pickup time: pull requests sitting unreviewed because everyone is heads-down on their own work. The fix is rarely "work harder". It is smaller pull requests, explicit review expectations, or simply lowering how much work is in progress at once.

Work in progress is the quiet killer here. The more things a team has half-finished, the longer every one of them takes — it is basic queueing theory, and it is why "stop starting, start finishing" is the most effective advice nobody follows.

From green dashboard to actual fix

This is the gap Process Intelligence is built to close. Instead of one averaged lead-time number, it maps where work actually slows down — which stage, which queue, which step is eating the days your team feels but your DORA chart hides. That turns "shipping feels slow" from a vibe into a specific, fixable thing.

The takeaway

Green DORA metrics are necessary but not sufficient. They prove your delivery is healthy at the level the business cares about; they do not tell you why the team feels stuck. When the dashboard and the developers disagree, trust the developers and go look at the flow. The bottleneck is almost always hiding inside an average.