Why do internal processes fail when designed by committee?
Internal processes fail because committees design for organizational completeness rather than user experience, producing workflows that satisfy every stakeholder’s requirements but serve no individual user well.
I audited the expense approval process at a 160-person organization. The process had 11 steps, required 4 different tools, and took an average of 6.3 days from submission to reimbursement. It had been designed by a committee of 5 people from Finance, Legal, Operations, IT, and HR. Each had added requirements that made sense from their perspective. Finance needed receipt documentation. Legal needed policy compliance verification. Operations needed budget allocation tracking. IT needed system integration. HR needed manager approval. The result was a process that served every department and delighted no one.
The compliance rate was 44%. More than half of employees found workarounds: submitting expenses through a manager’s corporate card, deferring reimbursement until the frustration threshold was exceeded, or simply absorbing small expenses rather than enduring the process. The 56% non-compliance was not rebellion. It was rational behavior in response to a process with a cost (time, friction) that exceeded its value for the user.
How do you apply product research methods to internal workflows?
Apply the same 4 methods used for external products: user interviews, job-to-be-done analysis, usability testing, and success metrics.
Step 1: Identify the user
Every process has a primary user: the person who initiates and completes it. For the expense process, the primary user was the employee submitting the expense. Not Finance. Not Legal. Not the approving manager. The employee. This distinction matters because committee-designed processes typically optimize for the reviewing stakeholders rather than the submitting user. Reversing this priority is the fundamental shift.
Step 2: Define the job to be done
The employee’s job to be done is simple: “I spent money on behalf of the company and I want to be reimbursed with minimal effort.” The time target should match the user’s expectation, not the organization’s convenience. I benchmarked against consumer fintech: Venmo settles a payment in under 24 hours. An internal expense process that takes 6.3 days is competing with that expectation and losing.
Step 3: Usability test the current process
I observed 8 employees completing an expense report. Average completion time was 23 minutes. The longest step was finding and attaching receipts (7 minutes). The most confusing step was selecting the correct budget code from a dropdown with 147 options (4 minutes of scrolling and guessing). 3 of 8 employees selected the wrong budget code and were later asked to resubmit. None of this friction served the user.
Step 4: Redesign from the user’s perspective
I redesigned the process with 3 principles: minimize user input, automate categorization, and parallelize approvals.
- Receipt capture: Replaced manual upload with a mobile photo integration. Snap the receipt, the system extracts the data via OCR. User input reduced from 7 minutes to 30 seconds.
- Budget code: Replaced the 147-option dropdown with an auto-suggest based on the merchant and amount, pre-filling the correct code 84% of the time. User input reduced from 4 minutes to 5 seconds for the 84% that auto-filled correctly.
- Approval routing: Replaced sequential approval (manager, then Finance, then Legal review for amounts over $500) with parallel routing. Manager and Finance review simultaneously. Legal review is triggered only for flagged categories, which accounted for 12% of submissions.
Step 5: Define and measure success
Set 3 metrics:
- Completion time: Target under 5 minutes for submission (achieved 3.2 minutes, down from 23 minutes).
- Cycle time: Target under 48 hours from submission to reimbursement (achieved 1.4 days, down from 6.3 days).
- Compliance rate: Target above 85% (achieved 89%, up from 44%).
How do you sustain process quality over time?
Sustain process quality by treating internal processes with the same continuous improvement cadence used for products: quarterly user feedback, usage analytics, and iteration.
I established a quarterly “process review” for the top 5 highest-volume internal workflows (expense reporting, PTO requests, procurement, access provisioning, and incident escalation). Each review followed the same template: check the 3 success metrics, interview 3 recent users, and identify 1 improvement to implement before the next review. This cadence prevented the re-accumulation of complexity that had produced the 11-step original process.
The fundamental insight is that internal processes are products. They have users, jobs to be done, usability characteristics, and measurable outcomes. The only difference between an internal process and an external product is that the internal process’s users cannot switch to a competitor. This captive audience makes it easy to ignore their experience. It does not make it wise.