Why Work Sits In Review For Days And What To Change First

Work is finished on Tuesday and ships the following Monday. Nobody was lazy. It sat in review, which is where most delivery time quietly goes.

Work sits in review because review is almost always one person's queue with no limit, no stated turnaround and no rotation. The fix is structural, not motivational: spread reviewing across more people, cap how big a single piece of work can be, and publish a turnaround everyone agrees to. Asking reviewers to be faster does not work, because they are usually already busy doing their own work.

Why it is a queue problem

Look at who reviews. On most teams, one or two people review nearly everything, because they know the most and because it happened that way.

A queue with one server and no limit behaves badly by nature. When that person has their own deadlines, the queue grows. When they take leave, it stops entirely. Everyone else sees delay and reads it as a person being slow, when it is a system with a single point of flow.

You can confirm this in an afternoon. Count the last thirty reviews and note who did them. If two names cover most of it, you have your answer, and no amount of encouragement will change it.

Three changes that work

Rotate the reviewing

Spread the work across more people, deliberately, even when it is slower at first.

Pair the newest reviewer with an experienced one for a few weeks. Accept that early reviews from new reviewers will be shallower. That is a training cost with an end, not a permanent tax, and it removes the single biggest cause of stalled work.

Be honest about the exception. Some work genuinely needs a specific person, for regulatory or safety reasons or because of deep knowledge nobody else has. Name that category explicitly and keep it small. Everything outside it is rotatable.

Cap the size of what goes in

The single strongest predictor of a slow review is a big change.

A large piece of work is hard to review in a gap between meetings. So the reviewer waits for a clear block of time, and clear blocks of time do not appear. A smaller piece gets reviewed in the gap it fits into.

Set a rule of thumb the team agrees on, and let people split work before submitting rather than after being asked. Splitting is annoying for the person who made the thing and cheap for everyone downstream.

State the turnaround out loud

Not a target. An agreement, with a stated consequence.

Something like: anything submitted before midday gets a first response by the end of the next working day. First response can be "I have looked and I need until Thursday", which is fine. What is not fine is silence, because silence is what makes people go and do something else and lose their context.

Then make breaches visible in the weekly review. Not to blame anyone, but because a repeatedly missed turnaround means the load is wrong and needs rebalancing. Give each breach a named owner, since an unowned warning has the same fate as amber nobody owns.

What to change first

If you only do one thing, do the turnaround agreement, because it makes the problem visible.

Once everyone can see which items breached and why, the other two changes stop being an argument. The data makes it obvious whether the queue is concentrated on one person or whether the pieces are simply too big.

Two things that look like fixes and are not

Adding a review column and calling it done. Seeing the queue is useful. It does not shorten it. Plenty of teams have a beautifully visible review column with twelve items in it, and a column that fills up but never empties is one of the signs a board has turned into fiction.

Hiring or assigning a dedicated reviewer. This concentrates the queue further and creates a bottleneck with a job title. It also separates reviewing from doing, which tends to make reviews more theoretical over time.

Watch the right number

Stop measuring how long things take from start to finish. Measure how long they wait in review specifically.

That one number tells you whether your changes are working. If it drops and stays down for a month, keep going. If it drops and rebounds, something outside review is pushing work in faster than the new arrangement can handle, and the conversation moves upstream to how much work you start at once.

Most teams are surprised by how much of their delivery time this single stage accounts for. It is worth measuring before you accept anyone's explanation, including your own, and it is a better thing to lead with when reporting progress to people who never open the board.