Contract Pricing Module.
Collapsed a four-tool, 23-page pricing workflow into three decision surfaces — cutting pricing errors −24% and task time −40% across 9,000+ hotels.
I led the workflow and scope redesign: contextual inquiry in Dubai, 23 pages collapsed into three decision surfaces, and live-data prototypes that validated the model before production.
Key results
- -24% — Pricing errors reduced (system logs, pre vs post launch) (Measured)
- 23 → 3 — Pages collapsed into single-surface decision-making across the Yielding workflow (Measured)
- ≈€133K — ≈€75K protected margin + ≈€58K recovered pricing capacity (modeled, conservative) (Modeled)
What I measured
- User error reduction — -24% (System logs + user-reported corrections, pre/post redesign)
- Page count for core workflow — 23 → 3 (Information architecture restructure)
- Time on task — -40% (Usability testing, observed task duration)
- Perceived utility (UMUX-lite) — +22% (UMUX-lite, baseline vs redesign (three benchmark rounds))
- Learnability — No onboarding needed (Users completed the redesign test without explanation of changes)
Pricing across 9,000+ hotels, in a tool nobody trusted
Yielding set margins across 9,000+ hotels while copy-pasting through four tools. Small errors compounded into real margin loss.
Expanding scope
Literal requirements exceeded the budget and timeline.
Disconnected stack
Every markup crossed four tools.
9,000+ hotels
Small error rates scaled into margin leakage.
Distributed team
Users were in Dubai; delivery teams were in Europe.
If this doesn't improve, we're going to build our own Excel tool.
I don't trust what the system is showing me. I always double-check in another tool.
Yielding team member, Dubai
Contextual inquiry
From porting a prototype to collapsing a four-tool workflow
I reframed a tool migration as single-surface decision-making: compete on pricing intelligence, not navigation.
“Bring the Yielding tool inside the back-office and improve it.”
“Collapse a four-tool workflow into single-surface decision-making, so the team competes on pricing intelligence, not operations.”
Everything a markup decision needs now sits on one surface.
The requested scope was larger than the budget could support.
I visualized the full requirement cost with the PM; the business refined scope before implementation.
One model across four groups.
Business Owner
Reviewed the cost of the expanded scope.
Yielding team (Dubai)
Co-designed the three-tier markup model.
Dev team
Tested feasibility through live-data prototypes.
PM & Business Unit
Agreed phased, measurable delivery.
23 pages → 3 screens
Page count was a symptom. The target was a complete pricing decision without navigation.
Product, Sales, and Scheduled markups became three stacked levels. History and breakdown stayed beside every decision.
23 pages per pricing workflow
Copy-paste across four tools
No alert for new hotels
External double-checking of system data
What I measured, and what I couldn't
Directly observed metrics from system logs, UMUX-lite benchmarks, and user behaviour.
Real margin uplift
No booking integration connected pricing decisions to outcomes.
Shadow-tool displacement
Continued use of shadow Excel files was not instrumented.
Decision quality
The effect of faster work on markup quality was not isolated.
What worked · what I'd do differently
What worked
Testing with live data before writing production code.
Live database values let the team validate full workflows before production code.
What I'd do differently
Lobby for financial instrumentation from day one.
Connect pricing decisions to booking outcomes early enough to measure margin after launch.
FAQ
How did you measure the 24% error reduction, and what does that mean in business terms?
I tracked incorrect markup applications via system logs and user-reported corrections, pre and post redesign. Modeled conservatively from those measured proxies, that is ≈€75,000 a year in protected margin (~500 errors prevented × ~10 bookings affected per error × €15 margin-at-risk per booking), before counting competitive or reputational losses.
You reduced pages from 23 to 3. What did that actually achieve?
Page count was a symptom; the goal was less navigation and cognitive load. Everything needed for a decision now sits on one surface, ~40% faster in testing, worth €58,000+ in recovered pricing capacity that the team redirected to strategic pricing.