Shutdown Decision Framework (2026)

Summary

This framework provides structured criteria for deciding whether to shut down (sunset) a product, feature, or business unit versus attempting a turnaround, using a multi-dimensional scorecard that guards against sunk cost fallacy. The default bias should be toward shutting down earlier rather than later: most organizations delay shutdown by 6-18 months past the optimal point, consuming resources that could fund growth elsewhere. [src1] A structured scorecard evaluating financial performance, strategic alignment, user engagement, maintenance burden, and opportunity cost produces a kill/turnaround/pivot recommendation with stakeholder communication sequencing. [src3]

Constraints

• This framework applies to product lines, features, and business units within an ongoing company — not company-wide liquidation, which requires legal, fiduciary, and regulatory guidance
• Requires at least 3 months of operating data (revenue, usage, cost) to score meaningfully — do not apply to products launched less than 90 days ago
• Shutdown is a one-way door for customers, partners, and contracts; once communicated externally, reversal destroys credibility and trust [src2]
• Sunk cost fallacy is the dominant cognitive bias in shutdown decisions — teams consistently overvalue past investment and undervalue future opportunity cost [src5]
• Communication timing is critical: premature external announcement causes talent flight and customer churn; delayed announcement creates legal liability and erodes trust [src1]

Decision Inputs

Decision Tree

START — Should we shut down this product/business unit?
├── Is revenue declining >30% YoY with no recovery plan credible?
│   ├── YES
│   │   ├── Does it consume >15% of engineering capacity?
│   │   │   ├── YES → RECOMMEND: Shut down (6-month sunset)
│   │   │   │   Reason: Accelerating decline + high resource drag
│   │   │   │   Constraint: Review contracts for minimum notice periods
│   │   │   └── NO → RECOMMEND: Shut down (9-12 month sunset)
│   │   │       Reason: Terminal decline but low cost buys time for orderly exit
│   │   └── Are there contractual lock-ins >12 months?
│   │       ├── YES → Maintenance mode → sunset at contract end
│   │       └── NO → Proceed with standard shutdown timeline
│   └── NO — revenue is flat or declining <30%
│       ├── Does the product align with current strategy?
│       │   ├── YES (core or adjacent)
│       │   │   ├── Credible turnaround plan with <6 month payoff?
│       │   │   │   ├── YES → RECOMMEND: Turnaround (time-boxed 6 months)
│       │   │   │   │   Constraint: Set kill criteria upfront
│       │   │   │   └── NO → RECOMMEND: Pivot or divest
│       │   │   └── User engagement stable or growing?
│       │   │       ├── YES → Hold and optimize (reduce costs)
│       │   │       └── NO → Set 6-month kill criteria, decide then
│       │   └── NO (non-core or conflicts with strategy)
│       │       └── RECOMMEND: Divest or shut down within 12 months
│       └── Can the product be sold or divested?
│           ├── YES → Explore divestiture before shutdown
│           └── NO → Proceed with shutdown decision
├── OVERRIDE CONDITIONS:
│   ├── Legal/compliance risk → Accelerate shutdown regardless of financials
│   ├── Key talent leaving because of this product → Factor attrition cost
│   └── Customer concentration >50% in one account → Negotiate transition first
└── DEFAULT (if inputs are ambiguous):
    └── RECOMMEND: Maintenance mode + 6-month review with explicit kill criteria
        Reason: Reduces spend while gathering data for a definitive decision

Options Comparison

[src1, src4]

Decision Logic

If revenue declining >30% AND engineering cost >15% of capacity

→ Shut down with 6-month sunset. The combination of accelerating revenue decline and high resource consumption makes the case clear: every month of delay costs the organization twice — in direct spend on a dying product and in opportunity cost of resources not deployed to growth. [src1]

If revenue declining AND product is non-core to strategy

→ Shut down or divest within 12 months. Non-core products in decline will never receive the leadership attention needed for a turnaround. Divestiture is preferable if a buyer exists who values the product more. [src4]

If revenue flat AND strategically aligned AND credible turnaround plan exists

→ Time-boxed turnaround: 6 months with explicit kill criteria. Set specific revenue, usage, and retention targets. If targets are not met at 6 months, shut down — do not extend. The most common failure mode is repeated extensions of failing turnaround attempts. [src7]

If data is insufficient (product <6 months old, metrics unclear)

→ Maintenance mode with 6-month review. Reduce investment to minimum viable operation, set explicit metrics to evaluate at review point. This preserves optionality while stopping the resource bleed. [src6]

If contractual obligations prevent near-term shutdown

→ Maintenance mode until contract expiration, then shut down. Honor existing commitments but do not renew or extend. Communicate sunset timeline to customers aligned with contract end dates. [src2]

Default recommendation

→ Maintenance mode + 6-month review with explicit kill criteria. When inputs are ambiguous, the lowest-risk path is to stop investing while gathering data for a definitive decision. Set quantitative kill criteria and commit to a decision at the review date. [src1]

Anti-Patterns

Wrong: Delaying shutdown because of sunk costs

Teams justify continued investment by citing what has already been spent. Past investment is irretrievable and irrelevant to the forward-looking decision. Organizations that let sunk costs drive shutdown timing delay by 6-18 months on average, consuming an additional 30-50% of total project spend with no return. [src5]

Correct: Evaluating only forward-looking costs and returns

Frame the decision as: "If I had this team and budget available today with no history, would I invest in this product?" If the answer is no, shut down regardless of past spend. Use pre-commitment: set kill criteria at launch and honor them mechanically. [src5]

Wrong: Announcing shutdown without a customer migration plan

Telling customers the product is ending without providing alternatives, data export tools, or transition timelines destroys trust and damages the brand for future products. [src2]

Correct: Building migration path before external announcement

Before any external communication: identify alternative solutions, build data export capabilities, create migration documentation, assign customer success managers to high-value accounts, and establish a timeline with reminders at 90, 60, 30, and 7 days. [src3]

Wrong: Running indefinite maintenance mode as a substitute for a decision

Maintenance mode is a valid temporary state but becomes an anti-pattern when it persists for years. Products in permanent maintenance mode accumulate technical debt, demoralize the assigned team, and create security and compliance risk. [src6]

Correct: Time-boxing maintenance mode with mandatory review dates

Enter maintenance mode with a specific review date (3-6 months) and quantitative criteria for the decision at review. If the product does not meet revival criteria at review, shut down. Never extend maintenance mode more than once. [src7]

Cost Benchmarks

Hidden cost multipliers: Add 15-25% for legal and compliance review, 10-20% for customer migration support, 5-15% for internal change management, and severance costs per local employment law. Companies that invest in dedicated migration support during sunset retain 85-94% of customers on other products vs 60-67% without. [src2, src4]

When This Matters

Fetch when a user asks whether to kill, sunset, or shut down a product, feature, or business unit. Also relevant when someone asks about product EOL decision criteria, how to communicate a product shutdown to stakeholders, shutdown vs turnaround decisions, or how to avoid sunk cost fallacy in product investment decisions.