Finding the bottleneck in the deliverables workflow

Situation

The numbers didn’t add up. Working through the dashboards for monthly reporting, I could see individual analysts performing well, while the team’s output stayed flat and below what the historic data said was achievable. The team was following the standard workflow, the same one everyone in the company used. Nothing was obviously broken and nobody was underperforming.

But if the people were productive and the total wasn’t, the loss was happening somewhere between them, in how the work moved rather than how it was done.

Constraints

The workflow was a company-wide standard, so I could adjust how work moved through it but not replace it. No extra people. No overtime. And no trimming the quality checks - whatever I changed had to hold the same accuracy standard the SLA required.

What I did

The specifics of the workflow are covered by NDA, so this describes the method rather than the mechanics.

1. Found where the loss was happening. I worked back through the dashboards, reports and historical data, then mapped the workflow as it actually ran - every stage a deliverable passed through, and where it waited. The delay wasn’t in the analysts’ work. It was in how long finished work sat before it moved to the next stage.
2. Worked out what could change. The workflow was a company standard, so I looked for adjustments rather than a replacement - points where the sequence or the batching could shift without altering the steps themselves or the checks attached to them. I generated several options and chose the one that stayed inside those limits.
3. Tested it in two rounds. I ran the adjusted flow for a week and watched what happened. That surfaced a second improvement I hadn’t seen previously, so I added it and ran another week before committing.
4. Made it the way the team worked. Once the second week confirmed it, I rolled the change out properly - walked the team through the new sequence, updated the procedure, and set the point from which everyone worked to it.
5. Showed the team the difference. At the end of the month I took the numbers back to them and walked through what had changed and why - not just the new total, but which part of the flow had been holding things up.

Result

The month after the change was the team’s highest in twelve: 2.5x the 12-month average, against a total that had been flat for months.

Nothing else changed to get there. No extra people, no overtime, no shortcuts through the quality checks. The same analysts did the same work to the same standard - the difference was the workflow itself: how work moved through it and in what order. The gap between individual performance and team output closed, which was the thing that had been wrong in the first place.