# The Cannibalization Decision Tree **A single-page worksheet for deciding sunset, refresh, or split.** By Falk Gottlob — falkster.com --- ## How to use this worksheet Print it. Fill it in by hand for the product line you are evaluating. Show the result to your CFO. Don't share with anyone else until the CFO has reviewed. Time required: 90 minutes. Most of the work is being honest with yourself. --- ## The four diagnostic questions ### Q1: Can the legacy architecture support the successor's quality bar? Imagine the agent-native version of your product you would ship in 12 months. Can the current architecture (data model, latency profile, security posture) support it without three foundational rewrites? - [ ] Yes, the architecture extends cleanly. - [ ] Partially — one or two significant rewrites needed. - [ ] No — fundamental rewrites required. **Test:** Have a senior engineer write a one-page sketch of how the successor would ship on the legacy architecture. If the sketch lists three or more foundational rewrites, the answer is "no." --- ### Q2: Is the legacy customer base the right ICP for the successor? The successor product has a target buyer. Maybe it's the same buyer as the legacy. Maybe it's a different role, a different department, a different size of company. - [ ] Same buyer, same use case. - [ ] Same buyer, different use case. - [ ] Different buyer entirely. **Test:** Pull your top 10 customer logos. For each one, ask: would they buy the successor? If 7+ would, same ICP. If 3 or fewer, different ICP. --- ### Q3: Can the company afford the gross margin trough? Gross margin will compress from 78-82% to 58-65% over 12 months, recovering toward 70-75% by month 24-30. - [ ] Yes — current cash position and revenue trajectory support the trough. - [ ] Maybe — the trough is borderline; we may need to fundraise. - [ ] No — without the legacy revenue, the company cannot survive the trough. **Test:** Model the trough at 12 points deeper and 6 months longer than expected. Does the cash position still hold? --- ### Q4: Is the buyer the same person? Same buyer means migration is feasible and the relationship handles both products. Different buyer means split. - [ ] Same buyer entirely (same role, same function). - [ ] Adjacent buyer (related role, same company). - [ ] Different buyer entirely (different role, different department). --- ## Routing to the operating mode Take the four answers and route. ### → Sunset **When:** Q1 = no, Q2 = yes, Q3 = yes. **What it means:** End-of-life the legacy on a published date 18-24 months out. Redirect all investment to the successor. Migrate customers in cohort waves. **The next step:** Run the seven-decision sequence in [The Cannibalization Playbook](https://falkster.com/blog/cannibalization-playbook). ### → Refresh **When:** Q1 = yes, Q2 = yes, Q4 = yes, AND AI is a feature improvement, not a business model replacement. **What it means:** Rebuild the legacy with AI capabilities under the same pricing and positioning. Customers experience the change as upgrades. **The next step:** Confirm by writing one roadmap that contains both products without lying. If you can't, the answer is actually sunset. ### → Split **When:** Q2 = no OR Q4 = no OR Q3 = no. **What it means:** Run legacy and successor as separate product lines under one company. Different teams, different sales motions, different P&Ls. Eventually one of them gets divested. **The next step:** Plan two operating models. The dual transformation model in [/handbook/dual-transformation](https://falkster.com/handbook/dual-transformation) covers the operational rhythm. --- ## The political map (preview) Once the operating mode is set, you have to move five seats. Quick reference: | Seat | Will support if | Will fight if | |---|---|---| | CEO | Option-value math is clear, board is pre-sold | Feels like CPO empire-building | | CFO | Trough is bounded and modeled | Revenue becomes variable without a forecast floor | | CRO | Comp plan is rewritten in their favor for the successor | Comp on legacy stays high | | CCO | CS gets redesigned around outcome quality | CS metrics stay tied to seat-based usage | | You | The legacy was right for its time, the successor is right for the next time | Loyalty to the product you built | The full coalition argument is in [The Cannibalization Decision Framework](https://falkster.com/handbook/cannibalization-decision-framework). --- ## The next-30-days plan If your answer is sunset: - Week 1: this worksheet, with CFO. - Week 2-3: unit economics model. Three scenarios. - Week 4-6: CFO sessions. Get the trough math committed. - Week 7-8: CRO sessions. Get the comp plan in draft. If your answer is refresh: - Week 1: this worksheet. - Week 2: write the single integrated roadmap and stress-test it. - Week 3-4: confirm the legacy unit economics still hold under the refresh. If your answer is split: - Week 1: this worksheet. - Week 2: define the two product lines and their separate P&Ls. - Week 3-4: identify the GMs or product leads for each line. --- ## What this worksheet doesn't do This worksheet doesn't tell you whether AI is going to disrupt your category. That's a separate analysis. This worksheet assumes you've already concluded it will. If you haven't, run that analysis first. This worksheet also doesn't tell you the answer to Q3 (can you afford the trough). That's a CFO conversation with real numbers from your specific company. Don't guess. --- *This worksheet pairs with the essay [The Cannibalization Playbook](https://falkster.com/blog/cannibalization-playbook) and the handbook chapter [The Cannibalization Decision Framework](https://falkster.com/handbook/cannibalization-decision-framework). Free to use, modify, and share. Attribution to falkster.com appreciated but not required.*