Scale / Pivot / Kill
A Simple Framework for Project Decisions
When you’re running a pilot or testing a new idea, eventually you hit a moment of truth: Do you expand it, change it, or stop it?
This framework helps you make that decision based on data, not emotion.
The Three Paths
SCALE ✓
What it means: Roll out the project to full scope. It’s working. Go big.
When to choose: - You’re hitting most of your success criteria - Gaps are minor and don’t undermine the core value - Team is engaged and ready - You have confidence this works at scale
Example: “Accuracy hit 91%, satisfaction is 83%, team adoption is 80%. The small gaps are worth the value we’re delivering. Scale it.”
PIVOT 🔄
What it means: Keep going, but change direction. Refine the scope, fix gaps, or adjust the approach.
When to choose: - Core value is proven, but something needs improvement - Gaps are solvable (not fundamental failures) - You need more data before scaling - Time and resources are available
Example: “Accuracy is 88% vs. 90% target. That’s solvable: 4 more weeks of model retraining will get us there. PIVOT: Extend 4 weeks, then reassess for scale.”
KILL ✗
What it means: Stop the project. The approach isn’t working. Move on.
When to choose: - The core idea doesn’t work (not fixable gaps) - You’ve proven the assumption wrong - The cost to continue exceeds the value - Better alternatives exist
Example: “Customers rejected the approach. We built for speed but they want personalisation. A different approach is needed. KILL this version and start over.”
How to Decide: The 3-Step Process
Step 1: Score Your Criteria (5 minutes)
Look at your Go/No-Go criteria from planning. For each one, did you hit the target?
| Criteria | Target | Actual | Hit? |
|---|---|---|---|
| Accuracy | ≥90% | 88% | ❌ |
| Satisfaction | ≥80% | 82% | ✓ |
| Response Time | <4 hrs | 3.5 hrs | ✓ |
| Cost per Query | $12 | $14 | ❌ |
| Team Adoption | 80% | 75% | ⚠️ |
| Escalations | <2% | 3.5% | ❌ |
Count the checkmarks. If you hit 4+ of 6, you’re in SCALE/PIVOT territory. If you hit 2-3, you’re in PIVOT territory. If you hit <2, you’re in KILL territory.
Step 2: Solvable vs. Fundamental (5 minutes)
Look at the gaps. Are they solvable (need more time/resources) or fundamental (the approach doesn’t work)?
Solvable gaps: - “Accuracy is 88% vs. 90%, that’s a 2-point gap we can close with more training” - “Team adoption is low; they just need more exposure” - “Cost is higher than expected; we can optimise once we hit scale”
Fundamental gaps: - “Customers hate this approach” - “The core technology doesn’t work for this use case” - “We can’t get the data we need”
If all gaps are solvable: You can SCALE or PIVOT. If some gaps are fundamental: You PIVOT or KILL.
Step 3: Make Your Call (5 minutes)
Based on steps 1-2, decide:
- SCALE if: Most criteria met + no fundamental gaps + team ready
- PIVOT if: Some criteria met + gaps are solvable + you have time
- KILL if: Few criteria met + fundamental gaps + better alternatives exist
Common Mistakes to Avoid
❌ Sunk Cost Fallacy “We spent $150K, so we have to scale it.” Reality: That money is gone. Only future value matters.
❌ Cherry-Picking Data “Customers are happy, so ignore the accuracy gap.” Reality: All criteria matter. You can’t ignore metrics that fail.
❌ **Confusing “Not Perfect” with “Broken” ”We’re 2 points away from target, so we must kill it.”* Reality: Small gaps = PIVOT, not KILL.
❌ Making Emotional Decisions “I feel like this could work…” Reality: Use your criteria. Data beats gut feel.
The Decision Template
Use this to document your call:
Decision: [SCALE / PIVOT / KILL]
Criteria Met: [List how many of your success criteria you hit]
Why? [2-3 sentences explaining your logic]
Trade-offs: [What are you giving up?]
Next Steps: [If SCALE: rollout plan. If PIVOT: what changes? If KILL: what’s next?]
Remember
- Your Go/No-Go criteria from the planning phase are your anchor
- Data beats emotion
- “Not perfect yet” is not the same as “doesn’t work”
- You can decide with imperfect information
- Most projects PIVOT; that’s normal and healthy
The goal isn’t to make the perfect decision. It’s to make the best decision you can with the data you have.