Strategy and planning
ICE, Kano and cost of delay
Faster scoring when reach is similar, categories of need, and how much waiting costs per week.
RICE is one conversation tool. Three others catch what it misses: ideas that hit very different audiences, basics that must not churn paying customers, and work whose delay has a dollar cost.
ICE
ICE = Impact × Confidence × Ease (ease is the inverse of effort, often 1–10). Faster than RICE; no explicit reach. Fine for early teams or growth experiments where every idea hits a similar audience. Weak when one idea hits 100 users and another hits 100,000.
Kano model
Noriaki Kano classified features by how they affect satisfaction:
| Type | If missing | If present | Example |
|---|---|---|---|
| Basic / must-be | Angry users | Neutral ("table stakes") | Login works; data is not lost |
| Performance | Linear dissatisfaction | Linear satisfaction | Faster search, more storage |
| Delighters | Nobody expected it | Disproportionate joy | Magical auto-organize |
Basics belong near the top of the queue even if RICE looks mediocre — customers do not "delight" at paying for a product that fails basics. Delighters are where differentiation lives, but only after basics are solid. Performance features are where continuous investment compounds.
Trade-off
RICE vs Kano. RICE ranks relative return per effort among optional bets. Kano tells you category of need. Use Kano first to protect basics and spot delighter opportunities; use RICE/ICE to sequence the optional work. Scoring a broken checkout below a shiny AI feature because "reach is low this month" is how you churn paying customers.
Cost of delay and WSJF
Cost of delay (CoD) asks: how much value do we lose per unit time by not having this? A feature that unlocks $50k/month of revenue delayed by two months costs ~$100k — before counting compounding or competitive loss.
WSJF (Weighted Shortest Job First), from SAFe / Don Reinertsen:
WSJF = Cost of Delay / Job DurationCoD often combines user/business value, time criticality and risk reduction. WSJF prefers short jobs with high delay cost — a bias toward finishing valuable thin slices instead of starting huge epics.
Walking through two projects
- Project Paywall: expected +$40k MRR once live; takes 4 weeks. Every week of delay costs about
$40k / 4.3 ≈ $9.3kof deferred MRR (rough; ignores compounding and churn). - Project Invite UX: expected +3 percentage points activation on 2,000 new workspaces/month; if 10% of activated workspaces convert to paid at $100/mo, that is
2000 × 0.03 × 0.10 × $100 = $600MRR per month of delay avoided — much smaller near-term CoD, but it may compound if activation feeds a growth loop.
WSJF would still prefer a short high-CoD job over a long one with similar CoD. If Paywall can ship a thinner version in 1 week that captures half the MRR, that thin slice often beats waiting for the "full" 4-week version — same intuition as MVP scoping.
You do not need perfect finance models. You need the habit of asking "what does waiting cost?" alongside "how hard is it?"
Common mistakes
- ICE when reach differs by 1000×. You hid the most important factor.
- RICE on a broken checkout. Kano basics jump the queue.
- Starting the huge epic when a one-week slice captures half the CoD.
Check yourself
Broken checkout vs a shiny AI feature. RICE likes the AI work this month. What do you ship? Checkout. Kano basic. Paying customers churn on missing must-bes.
Paywall is +$40k MRR in 4 weeks, or half of that in 1 week. Which does WSJF prefer? The 1-week thin slice. High delay cost, short duration.
Key takeaways
- ICE is faster when reach is similar; it fails when audiences differ a lot.
- Kano protects basics and spots delighters. Use it before you RICE optional bets.
- CoD / WSJF favors short, high-urgency jobs. Ask what waiting costs.
Further learning
Next: capacity allocation so debt and reliability are not guilt trips.