K
KnowMBAAdvisory
ProductIntermediate7 min read

Sunset Strategy

A sunset strategy is the deliberate, communicated, and supported retirement of a product or feature. It is the most leveraged decision a PM can make and the one most reliably avoided. Killing a product frees engineering capacity, simplifies the surface area customers must learn, reduces support costs, and clarifies positioning. KnowMBA POV: sunsetting is the most leveraged decision PMs avoid. Every quarter you keep a zombie product alive, you pay the salary cost of the team supporting it AND the opportunity cost of what they could have built instead. Apple killed the iPod after 22 years even though it still generated revenue, because the cognitive and operational tax of maintaining it exceeded its strategic value.

Also known asProduct SunsetEnd-of-LifeEOL PlanDeprecation StrategyProduct Retirement

The Trap

The trap is sentimentality and sunk-cost reasoning. 'We've invested 4 years in this product' is not a reason to keep it — it's a reason it should already be paying back, and if it isn't, more years won't fix that. The deeper trap: PMs avoid sunset decisions because the political cost is immediate (angry customers, demoralized team, executive pushback) while the benefit is diffuse and delayed (capacity freed, focus restored). So products live on as zombies, slowly draining the org. Microsoft kept Clippy in Office for 8 years past widespread user hatred because no one wanted to be the PM who 'killed' a beloved-by-execs character.

What to Do

Run a sunset review every 6 months on every product/feature. For each, score: (1) Active user count and trend. (2) Revenue contribution and trend. (3) Engineering maintenance hours per quarter. (4) Support tickets per active user. (5) Strategic fit (does it advance the mission?). Anything in the bottom quartile across 3+ dimensions enters a sunset plan. The sunset plan has phases: announce (90+ days notice for paid products), migrate (provide a path to a successor or competitor), wind-down (stop new signups, freeze features), shut down. Communicate ruthlessly — sudden sunsets damage trust permanently.

Pro Tips

  • 01

    Frame sunsets externally as 'focus' not 'failure.' Basecamp's announcement of killing minor products was titled 'Less is more,' not 'We're cutting things.' Same decision, different reception.

  • 02

    Pre-commit a sunset trigger when you launch a new product. 'If we don't hit 10K active users in 12 months, we sunset.' Pre-commitment removes the political battle later — the criteria did the killing, not you.

  • 03

    Measure the cost of NOT sunsetting. Every product carries a 'maintenance tax' — engineering hours, support load, infra costs, cognitive load on roadmap meetings. Make this number visible quarterly.

Myth vs Reality

Myth

If a product still has revenue, you can't sunset it

Reality

Apple killed the iPod with $200M+ annual revenue still on the table because keeping it cost more in distraction than it earned. Revenue is necessary but not sufficient — strategic fit and opportunity cost matter more. Cash cows that don't compound the company strategy become anchors.

Myth

Customers will revolt if you kill their product

Reality

Customers complain loudly during the announcement window then mostly migrate or churn quietly. Google has shut down 200+ products and continues to launch new ones with healthy adoption. The reputation damage is real but localized; the strategic clarity gained usually exceeds it within a year.

Try it

Run the numbers.

Pressure-test the concept against your own knowledge — answer the challenge or try the live scenario.

🧪

Scenario Challenge

You inherit a product with 8,000 active users generating $1.2M ARR. It hasn't grown in 3 years. It consumes 4 engineers full-time in maintenance and 2 support reps. Those 6 people could be working on your highest-growth product line, which is capacity-constrained. What's the right call?

Industry benchmarks

Is your number good?

Calibrate against real-world tiers. Use these ranges as targets — not absolutes.

Sunset Notice Period (paid B2B SaaS)

B2B SaaS product sunsets

Customer-Respectful

12+ months notice

Acceptable

6-12 months

Short Notice (trust damage)

3-6 months

Trust-Destroying

< 3 months

Source: Hypothetical: industry observation; informed by Google Reader / Atlassian Stride sunsets

Real-world cases

Companies that lived this.

Verified narratives with the numbers that prove (or break) the concept.

🍎

Apple

2001-2022

success

Apple discontinued the iPod Touch in May 2022, ending the iPod product line after 21 years and an estimated 450M units sold. The iPod was still generating revenue at the time of sunset. Apple's reasoning, summarized by SVP Greg Joswiak, was that 'the spirit of iPod lives on' in iPhone, iPad, and Apple Watch — the strategic clarity of focusing on those three categories outweighed the residual iPod revenue. The sunset was textbook: long product life, dignified retirement, no public hand-wringing.

iPod Lifespan

2001-2022 (21 years)

Estimated Total Units Sold

~450M

Revenue at Sunset

Still meaningful (8-figure)

Sunset Communication

Single press release, dignified

Even iconic products with positive revenue should be sunset when they no longer compound the company strategy. Apple's willingness to retire the iPod is what makes Apple — discipline about what NOT to ship is the moat.

Source ↗
📎

Microsoft (Clippy)

1997-2007

failure

Clippy (officially the Office Assistant) launched in Office 97 and was widely loathed by users almost immediately. Internal Microsoft research as early as 1999 showed users actively disabled it. But Clippy persisted through Office 2000, XP, and 2003 — a full 8 years past widespread user rejection — partly because of executive attachment and partly because no PM wanted to be the one to kill an iconic Microsoft mascot. Office 2007 finally removed the assistant entirely.

User Hatred Documented Internally

By 1999 (within 2 years of launch)

Years Until Sunset

10 years (1997-2007)

Sunset Lag (data → action)

~8 years

Subsequent Brand Cost

Long-running comedy meme

Slow sunsets damage product brand AND signal to the team that data doesn't actually drive decisions. Clippy is the cautionary tale: when the data says kill it, killing it 8 years later does not undo the cost.

Source ↗
🟦

Meta (Home, Slingshot, etc.)

2013-present

mixed

Meta has sunset dozens of products: Facebook Home (2013), Slingshot (2015), Paper, Hello, Notify, Lifestage, Moments, tbh, Facebook Gaming app, and many more. Most got 12-18 months of runway before being killed. Meta's velocity at sunsetting is unusual and treated internally as a feature: the willingness to kill products fast frees the company to launch many more attempts. Critics call this disrespectful to users; Meta treats it as the cost of an experimentation portfolio.

Notable Product Sunsets

20+ since 2013

Typical Time to Kill

12-18 months from launch

Internal Framing

'Cost of experimentation'

External Cost

User trust erosion in new launches

Fast sunsets enable fast experimentation BUT erode trust in your future launches. Meta has the brand surface area to absorb this cost; smaller companies typically don't. Sunsetting velocity should match brand resilience.

Source ↗

Decision scenario

The Zombie Product

You run product at a 100-person SaaS company with 3 product lines. Product C has 1,200 customers ($1M ARR), hasn't grown in 18 months, eats 5 engineers and 2 designers, and has the lowest NPS of your portfolio. Your CEO is attached to it because it was the company's first product. Your CTO is begging you to sunset it.

Product C ARR

$1M

Product C Growth (18mo)

0%

Team Allocated

5 eng + 2 design

NPS

12 (lowest in portfolio)

Opportunity Cost

Product B is capacity-constrained

01

Decision 1

You have to make a recommendation to the exec team next week. The CEO has hinted strongly that he doesn't want to kill it. Product B (the growth bet) is capacity-constrained and could absorb the freed team immediately.

Recommend keeping Product C — the political cost of fighting the CEO isn't worth $1MReveal
Six months later, Product C is still flat, still consuming the team, still NPS 12. Product B has lost its window — a competitor caught up while your team waited. The CEO eventually agrees to sunset Product C 18 months later, but the strategic cost is now $5M+ in lost B growth. You realize the political cost was always smaller than you feared, and the strategic cost was always larger than the CEO believed.
Product C ARR: $1M → $0.9M (slow decline)Product B Velocity: StalledLost Strategic Value: $5M+
Recommend a structured 12-month sunset with migration support, redeploy the team to Product B in two wavesReveal
The CEO pushes back hard, then accepts when you walk through the math (the $320K/year that Product C costs above its revenue, plus the opportunity cost on Product B). Sunset announced. 40% of Product C customers migrate to a partner; 35% accept refund; 25% churn unhappy. Within 9 months Product B doubles its velocity and adds $4M ARR. The CEO publicly credits the call as 'the hardest right thing we did this year.'
Product C: $1M ARR → sunsetProduct B Velocity: DoubledNew ARR (B): +$4M in 9 months

Related concepts

Keep connecting.

The concepts that orbit this one — each one sharpens the others.

Beyond the concept

Turn Sunset Strategy into a live operating decision.

Use this concept as the framing layer, then move into a diagnostic if it maps directly to a current bottleneck.

Typical response time: 24h · No retainer required

Turn Sunset Strategy into a live operating decision.

Use Sunset Strategy as the framing layer, then move into diagnostics or advisory if this maps directly to a current business bottleneck.