Feature Scoring

Feature Scoring and Prioritization

Perform structured, data-driven feature prioritization for the Talosix EDC platform. Incorporates clinical trial regulatory considerations alongside standard product metrics.

Workflow

1. Gather Feature List: Either:

   • Accept a list of features from the user
   • Pull from Jira using searchJiraIssuesUsingJql (e.g., type = Story AND status = "To Do" AND project = [X])
   • Get details on each feature using getJiraIssue

2. Select Scoring Framework: Ask the user which framework to use, or recommend one:

   • RICE: Best for comparing many features with quantitative data
   • MoSCoW: Best for release scoping and stakeholder alignment
   • Weighted Scoring: Best when multiple custom criteria matter
   • Combined: Use RICE for initial ranking, MoSCoW for release slotting

3. Score Features using the selected framework with EDC-specific criteria.

4. Present Results with ranked output and recommendations.

RICE Framework (EDC-Adapted)

Reach (1-10)

How many users/customers are impacted per quarter?

• 10: All customers, all users
• 7: Most customers, specific user role (e.g., all CDMs)
• 5: Subset of customers (e.g., Phase III studies only)
• 3: Few customers, specific use case
• 1: Single customer request

Impact (0.25 - 3)

How much does this improve the user's workflow?

• 3: Massive -- eliminates a major pain point or enables new capability
• 2: High -- significant time savings or error reduction
• 1: Medium -- noticeable improvement
• 0.5: Low -- minor improvement
• 0.25: Minimal -- nice to have

Confidence (20% - 100%)

How certain are we about reach and impact estimates?

• 100%: Strong data (customer interviews, usage analytics, regulatory requirement)
• 80%: Good evidence (multiple customer requests, competitive gap)
• 50%: Some signal (a few requests, team intuition)
• 20%: Speculation (no direct evidence)

Effort (person-months)

Total engineering, QA, and validation effort:

• Include: development, testing, validation documentation, regulatory review
• Account for: 21 CFR Part 11 audit trail requirements, IQ/OQ/PQ test scripts
• Factor in: customer re-validation communication and support

RICE Score = (Reach x Impact x Confidence) / Effort

MoSCoW Framework (EDC-Adapted)

Must Have

• Regulatory requirements (FDA, EMA mandates)
• Critical defects affecting data integrity
• Contractual commitments to customers
• Security vulnerabilities
• Features required for validated system status

Should Have

• High-demand customer requests (5+ customers)
• Competitive table-stakes features
• Significant efficiency improvements for site users
• Features that reduce support burden

Could Have

• Nice-to-have UX improvements
• Features requested by 1-2 customers
• Internal tooling improvements
• Performance optimizations (non-critical)

Won't Have (This Release)

• Low demand, high effort features
• Features with unresolved regulatory questions
• Items that require infrastructure not yet in place

Weighted Scoring Matrix

Score each feature 1-5 on these criteria, with EDC-specific weights:

Weighted Score = Sum of (Score x Weight) for each criterion.

Output Format

Ranked Feature List

Scoring Detail (per feature)

• Feature name and Jira link
• Score breakdown by criterion
• Key justification notes
• Dependencies or blockers
• Regulatory impact flag (Yes/No + explanation)

Prioritization Recommendations

• Immediate (this sprint/PI): Top-priority items
• Next quarter: High-value items requiring planning
• Backlog: Lower priority, revisit next cycle
• Decline/Defer: Items not worth pursuing now, with rationale

Sensitivity Analysis

• Which features change rank if weights are adjusted?
• What new information would change the prioritization?
• Flag features where confidence is low and more research is needed

Tips

• Always flag regulatory-driven features separately -- these often override scoring
• Consider the "validation tax": features touching validated workflows require extra QA effort
• Group related features that share infrastructure investment
• Note when a feature is a prerequisite for higher-scored future features