Promising One Hundred. Finishing Fifty-Six.

A delivery organization was committing to roughly a hundred epics a quarter and finishing about fifty-six. Nobody knew, because nobody had ever measured what the team could actually carry.

05 FLOW
The condition

A delivery team keeps saying yes to the business and cannot say how much work it is able to finish.

Commitments made in rooms delivery never entered

The work began with a terse email. A delivery leader inside a global IT organization had just learned, from two project managers standing in her office, that more work had been promised to the business on her team's behalf. Nobody from delivery had been in the conversation where the promise was made. She wrote to the business in frustration, and the executive sponsor spent the following days repairing the damage.

The sponsor asked Agileon to work with that delivery team on how it forecast and committed work, so the relationship between the business and IT could be rebuilt. The frustration was justified. Commitments were being handed to the business with nobody on the delivery side in the loop. But underneath the frustration sat a simpler problem, and it surfaced with the first question we asked.

A question nobody could answer

We asked the delivery leader how much work her team could actually deliver. She did not have an answer. She began opening spreadsheets, clicking between them, looking for anything that would let her make an educated reply. Nothing in them could tell her.

We asked how the team decided what to say yes to. Her answer was that they did not really say no. Project managers collected commitments in meetings with the business. The delivery team had no sense of its own capacity, and nobody ever declined. From the business side, the conclusion was simple: this team could not deliver anything. There was no visibility into why, because there was no visibility into anything.

We asked whether anyone on her team used a board to track work. A couple of developers did, people who had come from elsewhere in IT, and they used boards only for their own tasks. There was no shared picture of what was in flight, how long anything took, or how much finished in a quarter. We were starting from scratch.

A small team working through printed documents and notes around a shared table.

Letting past work size the future

Asking a team to estimate its capacity when it has never measured anything produces a guess. So we did not ask for estimates. We asked the delivery leader to pull about a dozen examples of past work, deliberately mixed in size: something that had taken a year, something that had taken a quarter, something that had taken a few days.

Then we brought leadership into a room and asked them to sort those items into size buckets. What they did not know was that we already held the real start dates, the real end dates, and the time charged against each item. As they sorted, the true averages fell out of the data. The buckets were no longer opinions. They were durations the organization had already lived through.

Figure 1
What each size of work actually took
  1. InitiativeAbout 210 days
  2. EpicAbout 70 days
  3. FeatureAbout 30 days
  4. User storyAbout 12 days
  5. TaskAbout a day and a half
Leaders sorted a dozen past items into buckets; the real dates already on record gave each bucket an average duration. All figures are approximate averages from that sample.

Fifty-six against one hundred

Next we built the team a combined board, asked what steps work naturally moved through, and for a couple of weeks did nothing except have people move their items across it. No new metrics, no targets, no lectures. Only the habit of making work visible where everyone could see it.

Once the habit held, we showed them throughput. The team could finish about fifty-six epics a quarter. It was committing to an average of one hundred per quarter to the business. The delivery leader sat with that number for a moment and said the team was underwater.

That gap explains everything the business had been experiencing. The team was not failing to deliver. It was finishing what a team of its size could finish while the organization promised nearly twice that. Every missed date, every escalation, and the email that started the engagement were downstream of a commitment process that had no capacity number in it. The same missing measure also hid what happened when more work was poured in on top, which is a story of its own.

Figure 2
Epics committed versus completed, per quarter
Committed to the business About 100
Actually completed About 56
Once work moved across a shared board, throughput could be counted. The organization had been promising the business close to double what the team could finish. Both figures are approximate quarterly averages.
Figure 3
How capacity was measured without asking for estimates
  1. Pull a dozen past items Deliberately mixed: a year, a quarter, a few days
  2. Leaders sort them by size Real dates and time charges already in hand
  3. What changed Averages fall out Each size bucket gets a real duration
  4. One shared board Two weeks of only moving work across it
  5. The reveal Show throughput About 56 finished against about 100 promised
Each step replaced an opinion with something the organization had already done, so the result could not be argued away.

Commitments sized to what the team can finish

The tension that opened this story also produced a recurring forum where the people receiving, leading, and delivering work sat in the same room before commitments were made. But a forum cannot make an honest commitment without a capacity number, and that number had never existed. The strategy we set was plain: establish cycle times for each size of work, make throughput visible every quarter, and connect both to how much work the team was allowed to carry at once.

With those measures in place, the conversation with the business changed character. A request could be weighed against a known rate of completion instead of against goodwill. In the pilot, teams moved to committing smaller amounts of work sized to their real capacity, which reduced disruption in the middle of a cycle, and earlier sight of incoming work let them forecast and commit more realistically. That is not the same as closing the gap. What changed is that the organization could finally see the distance between what it promised and what it finished, and decide what to do about it in the open.

A commitment is only as honest as the capacity number behind it, and most teams have never been asked for one.

Share With Your Team

Choose who it is for. You send it yourself from your own email; Agileon never writes to your colleagues.

Not sure where to start?

Answer five questions about where work stalls. We send back a starting plan: where to begin and what to look at first.