Stage-Gate Pulses: Matching Work Intake to Team Throughput
Most teams do not have a work problem so much as an intake problem. Requests arrive from sales, leadership, customers, and the people sitting next to you, and each one feels small enough to accept. Within a month the board holds forty items, half of them started, none of them clearly owned, and the team's sense of its own pace has quietly broken. This article takes a different angle from weekly priorities or meeting cadence. Instead of asking how to pick what to do, it asks how fast new work should be allowed to enter the system at all, and how to set that pace using evidence the team already has.
Start With Measured Throughput, Not Estimated Capacity
Capacity estimates are usually optimistic. People plan around the hours they wish they had, not the hours that actually produce finished work after meetings, support pings, and rework. A more reliable starting point is throughput: how many items the team has actually moved to done over the last four to six completed weeks. Look at the board's history and count finished items per week, then note the spread between your best and worst weeks.
Throughput gives you a range rather than a promise. If your team finished between four and seven items per week, that range is more useful for planning than any single number. It also surfaces a common pattern: the weeks that look most productive are often the ones where a deadline forced everything to close, which means they may not be sustainable.
Ask the team a few plain questions before trusting the numbers. Which finished items were genuinely new work, and which were carryover? Which weeks had holidays, outages, or unusual absences? Throughput built on distorted weeks will mislead you, so note the anomalies alongside the counts.
Set an Intake Limit Tied to Finishing, Not Starting
A useful intake rule is simple: new work enters the active lane only when a comparable item leaves it. This sounds rigid, but it changes behavior quickly. When a request arrives, the question is no longer whether it is important, which it usually is, but which current item it will displace or whether it waits in a visible queue.
Keep the queue distinct from the active lane. A queue holds accepted requests that have a rough description and an owner candidate. The active lane holds only what someone is working on now. Many teams collapse these into one column, and that collapse is where drift begins. Items in the active lane should have a named owner and a next action; items in the queue can be rougher but should still be dated for review.
Set the active-lane limit from your throughput range, not from ambition. If the team finishes four to seven items a week, an active lane of ten is a signal that work is starting faster than it can finish, regardless of how hard everyone is working.
Use Aging to Detect Intake Problems Early
Intake problems rarely announce themselves through missed deadlines. They appear first as age. An item that has sat in the active lane for three weeks with no movement is telling you something, whether the work is blocked, unclear, or quietly deprioritized. Tracking how long items have been in each state gives you an early warning that the rate of entry exceeds the rate of exit.
Review aging in a short recurring check rather than during crises. Ask which items have stayed in the same state longer than the team's typical cycle, why they are stuck, and whether they should be returned to the queue. Returning an item to the queue is not failure. It is an honest acknowledgment that the team's capacity has moved, and it keeps the active lane meaningful.
A Practical Intake Checklist
Before accepting a new request into the active lane, check the following. Does the request have a one-sentence description a stranger could understand? Is there a named owner or a clear person to assign it to? Which current item does it displace, or is there an open slot within the team's measured limit? Is there a definition of done that someone other than the requester would agree with?
If the answer to two or more of these is no, the item belongs in the queue. Keep the queue reviewed weekly so that items do not age there unnoticed. Compare the number of items accepted into the active lane each week against the number finished. If intake consistently outruns completion for three or four weeks, lower the limit or pause new acceptance and clear what is already open.
The goal is not to refuse work. It is to make the cost of each acceptance visible at the moment it is made, so that teams can say yes to the right things on purpose rather than discovering the trade-off after a launch slips.