A financial services firm carried a thick middle layer of coordinators whose job was to move information between teams that could not talk directly. As shared systems replaced the handoffs, that layer became overhead. The firm compressed mid-layer coordination roles 35 percent in twelve months, from 260 to 169, while expanding the system engineer and quality steward roles the new model needed. It ran the change through attrition, redeployment, and reskilling rather than layoffs. Coordination cost fell 4.1 million dollars a year, onboarding dropped from 31 days to 12, and the firm rebuilt its shape around the work.
A layer that existed to pass messages
The firm was a mid-sized financial services business with about 3,400 employees. Between the frontline teams and senior management sat a coordination layer of 260 people whose function was almost entirely to move information: reconcile numbers between systems that did not connect, chase status across teams that could not see each other's queues, and assemble reports by hand from four sources. Nobody had designed this layer. It had accreted over a decade as the firm bolted on systems that never shared data, and each new gap was patched by adding a coordinator.
By the time Stratenity was engaged, the coordination layer cost 23.6 million dollars a year fully loaded and had become the firm's slowest link. A typical client onboarding touched nine coordinators and took 31 days, most of it waiting in queues rather than doing work. The firm had just finished rolling out a shared operating platform that gave teams direct visibility into each other's data. The platform made most of the message-passing unnecessary, but the roles built to do the message-passing were still fully staffed. Worse, the platform itself was undergoverned: it had gone live with only 34 system engineers to own its automated flows and 28 stewards to handle the exceptions it escalated, so the tooling that should have absorbed the coordination load was itself a source of new risk. The engagement was scoped to reshape the workforce around the platform: compress the coordination layer, and expand the system engineer and quality steward roles the new model depended on, treating the two moves as one program rather than a cut followed by unrelated hiring.
Reshape around the work, not the org chart
The engagement mapped the coordination layer's actual activities rather than its titles. Of the 260 roles, activity analysis showed 71 percent of hours went to work the shared platform now did automatically, and 29 percent went to genuine judgment work that needed to be kept and elevated. That split, not a headcount target, drove the redesign. The firm set a 35 percent compression over twelve months and paired it with deliberate expansion of the two families the new model was short on: system engineers who owned the platform's automated flows, and quality stewards who governed exceptions the automation escalated.
| Role group | Start of program | Twelve months later | Mechanism |
|---|---|---|---|
| Mid-layer coordinators | 260 | 169 | Attrition, redeployment, reskill; no like-for-like backfill |
| System engineers | 34 | 72 | Reskill from coordinator pool plus targeted external hire |
| Quality stewards (exception governance) | 28 | 61 | Promote experienced coordinators into elevated review roles |
| Frontline teams (net effect) | 1,180 | 1,205 | Absorbed simple checks the platform surfaced directly |
| Involuntary exits | Baseline | 0 in year one | Redeployment-first commitment held throughout |
The compression math worked because it was never a straight cut. Of the 91 coordinator roles removed, 39 people moved into system engineer and quality steward roles through a four-month reskill path, 28 were redeployed into frontline teams that were expanding, and 24 left through ordinary attrition and were not backfilled. The firm treated the coordinator pool as the primary talent source for the roles it was growing, which is why compression and expansion happened in the same population rather than through separate hiring and firing. That choice was not only humane, it was practical: coordinators already understood the firm's processes and clients, so a reskilled coordinator reached productivity in the engineer and steward roles faster than an external hire, who had to learn the domain from scratch. The savings from the removed roles were not swept to the bottom line either; roughly half were recycled to fund the reskill path and the expanded families, which is what let the program pay for its own transition.
The sequencing also mattered. The firm did not announce a 35 percent target and then work out who would go. It first stood up the expanded engineer and steward roles and opened the reskill path, so that people could see where they were moving before any coordinator role was retired. Only once 39 people had a destination did the compression begin in earnest, and the remaining reductions were timed to natural attrition and voluntary moves. Running expansion ahead of compression cost a few months of overlap, but it removed the fear that turns a workforce reshape into a retention crisis, and it meant the platform gained its engineers and stewards before, not after, the coordination layer thinned out beneath it.
What the reshape delivered
- The mid-layer coordination group compressed from 260 to 169 roles, a 35 percent reduction, achieved with zero involuntary exits in the first year.
- Annual coordination cost fell 4.1 million dollars, from 23.6 million to 19.5 million, with the savings partly redeployed to fund the expanded engineer and steward roles.
- System engineer headcount more than doubled from 34 to 72, and quality stewards rose from 28 to 61, giving the shared platform the ownership and exception governance it had lacked at launch.
- Client onboarding cycle time fell from 31 days to 12, because the nine sequential coordinator handoffs collapsed into direct platform visibility with steward oversight on exceptions only.
- 39 former coordinators moved into higher-skilled, higher-paid roles through the reskill path, turning a compression program into a visible internal mobility story that protected morale.
What the engagement taught
- A coordination layer is a symptom, not a role. When systems cannot talk, someone gets hired to talk for them. Fix the system and the role stops being necessary, but only if you also reshape the workforce deliberately.
- Map activities, not titles. The 71/29 split between automatable and judgment work was the number that made the compression defensible and told the firm exactly how much to keep.
- Compression and expansion should draw from the same pool. Sourcing the new engineer and steward roles from the coordinator population turned a cut into a redeployment and kept institutional knowledge inside the firm.
- The zero involuntary exit commitment slowed the program but bought the trust that made the reskill path credible. People reskill willingly when they believe the firm is moving them, not managing them out.
- Cycle time, not headcount, was the outcome leadership cared about. The 31-to-12-day drop is what made the board comfortable that the compression improved the business rather than just the cost line.
How to run this in your firm
- Identify any layer whose main job is moving information between systems or teams, and run an activity analysis to split automatable hours from judgment hours.
- Set the compression target from the automatable share of the work, not from a top-down cost number, so the cut lands on work the systems now absorb.
- Name the roles the new operating model is short on, and commit to sourcing them first from the population you are compressing.
- Fund a reskill path from the savings and make redeployment the default, keeping involuntary exits as a last resort to protect trust.
- Instrument a cycle-time metric that the reshape is accountable for, so the program is judged on how the business runs, not just on headcount removed.