Softeq Development — enterprise software studio, Fortune 500 clients
Turning a Design Department into a Profit Engine
How I rebuilt a loss-making design unit into a profitable, high-retention team.
The challenge
The design unit ran at a −27% annual result. The root cause was bench time: designers sitting between projects with no billing, while hiring ran ahead of confirmed demand. Design was treated as a cost center, and designers had no clear path to grow — a churn risk in a competitive market.
What I did
1. Killed the bench. Moved the team onto long-term dedicated assignments instead of short project-to-project work, so billing became predictable. The bench was the whole problem: designers between projects generate cost and no revenue, and the schedule had been built as if that gap didn't exist.
Fig. 01 — Where the bench went
Mechanism, not a slogan
Before
Deal is sold
Engineers guess “design hours”
→↓
Designer is hired
Ahead of confirmed demand
→↓
Project ends
Short engagements, gaps between them
→↓
Bench
Cost with no billing
Result: every design contract sold below its real cost.
After
Designer joins pre-sales
Whoever does the work estimates it
→↓
Estimate lands on plan
Or under — never over
→↓
Hire only against a contract
Under three months — I take it myself
→↓
Long dedicated assignments
2–3 years at premium rates
Result: bench close to zero — the whole of the swing in Fig. 04.
2. Made hiring a last resort. Headcount only against a confirmed project, never in anticipation of one. When a project could be covered without a new hire, I took the delivery work myself.
3. Moved estimation to the people doing the work. Designers estimated and owned their own projects rather than receiving a number decided above them. Estimates got more accurate and delivery came down to 2–3 sprints.
4. Built a team people stayed in. Ten designers, zero attrition over the period — in a market where designers change jobs every 18 months.
Fig. 02 — Team and delivery
Source: Softeq staffing and delivery data
2 → 10 designers, 0 left
● When I took the unit
● Hired against confirmed projects
I hired seniors only. The two mid-level designers were hired by those seniors, under my supervision — so they had people to mentor on their way to lead. Both grew to senior. Every hire, grade and growth plan ran on one competency matrix.
Design delivery: 5–7 sprints → 2–3
Sprints per design delivery
Designers estimate their own future projects instead of leads doing it for them — which is also why estimates became trustworthy enough to price against (close to 100% accuracy).
5. Won the budget, then the tools. A design department inside an engineering company gets nothing by default. I secured a training budget — Nielsen Norman Group courses, with certification for those who passed — and company-wide licences for design tooling: Zeplin and InVision first, later Figma. Handoff moved from pictures in Jira and assets on Drive to designs engineers inspect themselves.
Fig. 03 — Handoff and the operating layers
What the budget bought
Handoff · before
Pictures in Jira, assets on Drive
Engineers rebuilt screens from images and asked the designer for every measurement.
Handoff · after
Inspectable designs for the whole company
Company-wide licences — Zeplin and InVision first, later Figma. Engineers pull specs and assets themselves.
Training
Nielsen Norman Group courses
A training budget I had to win. Designers who passed the exam got certified.
Layer 1
Standards
One library · a one-page brief · an explicit “design done”
Layer 2
Rituals
Weekly written review · escalation path for mid-contract scope changes
Layer 3
Estimation
The designer who does the work estimates it
Beyond design
Process navigator
Role-by-role delivery map in Figma over Confluence — used by all 450 employees
The solution
Company BI recorded the annual department result moving from −27% to +32%. The team grew from 2 to 10 with 0% attrition, and design shifted from a back-office function to a recognized driver of enterprise wins across SaaS, healthcare, fintech, and IoT. Led design teams across enterprise and Fortune 500 engagements — Lenovo, Halo Collar, jDisc, G.J. Gardner Homes. On several accounts I staffed the team myself: scoped the engagement, hired the designer against that specific project, then led delivery. Design also carried presale at that scale — interactive prototypes of the prospect's future product for pitches to companies like Deloitte and ExxonMobil.
Fig. 04 — Annual department result
Source: Company BI · Softeq
2022
2023
A 59-point swing in one year.
The loss was not a pricing problem. It was bench time: designers between projects with nothing to bill, and hiring that ran ahead of confirmed demand. Remove the bench and the same team, at the same rates, is profitable.
What I’d do differently
I fixed the economics before I could explain them. The department was already profitable by the time leadership had a clear picture of why — I was reporting outcomes rather than the mechanism, and that made the result look like luck instead of a system. It also meant the argument had to be re-made each time budgets were reviewed. I'd instrument bench time and utilisation from month one now, so the numbers make the case before I have to.