Feature Prioritization When Everything Feels Urgent
Andres Max
The hardest part of product isn’t deciding what to build. It’s deciding what not to build. When your biggest customer threatens to churn without feature A, your sales team can’t close deals without feature B, and your engineers say the codebase will collapse without addressing technical debt C, you don’t have a prioritization problem. You have an emotional problem.
Everyone believes their thing is urgent. They’re often right. The problem is you can only do one thing at a time.
After 18 years of making these decisions, here’s how I think about prioritization when everything feels critical.
Why Standard Prioritization Fails
Most prioritization frameworks assume rational actors and complete information. In real startups, neither exists.
Problem 1: Everything Has a Champion
Feature A has your biggest customer demanding it. Feature B has your head of sales threatening to miss quota. Feature C has your CTO saying the system will fail without it.
Each champion genuinely believes their feature is most important. They have evidence. They have passion. They’re probably not wrong that their thing matters.
The failure: Frameworks that ignore politics don’t survive contact with stakeholders.
Problem 2: Effort Is Always Unknown
RICE scores, ICE scores, and similar frameworks require effort estimates. In software, effort estimates are notoriously unreliable. That “2-week” project takes 6 weeks. That “huge” project ships in a week.
The failure: Frameworks dependent on accurate effort estimates produce garbage outputs.
Problem 3: Impact Is Unknowable
How much will feature A increase retention? How much will feature B increase sales? You don’t know. You’re guessing. Even with data, you’re extrapolating from limited information.
The failure: Frameworks pretending to quantify unknowable impact create false precision.
Problem 4: Urgency Is Emotional
“Urgent” usually means “someone is anxious about it.” Your biggest customer’s urgency isn’t necessarily correlated with actual business impact. They just happened to email recently.
The failure: Frameworks that don’t account for recency bias prioritize whoever complains loudest.
The Prioritization Framework That Works
Here’s the framework I actually use. It’s not elegant. It works.
Step 1: List Everything (Without Judgment)
Dump every request, feature, bug, and technical task into a single list. No filtering. No categorizing. Just everything.
Include:
- Customer requests
- Internal requests
- Technical debt
- Bugs
- Ideas from anywhere
This list will be overwhelming. That’s intentional. You can’t prioritize what you can’t see.
Step 2: Apply the “What Happens If We Don’t?” Test
For each item, answer: “What happens if we never build this?”
| Answer | Category |
|---|---|
| We lose significant revenue | Critical |
| Customers are frustrated but stay | Important |
| It would be nice but life goes on | Nice-to-have |
| Nothing happens | Delete |
Be honest. Most features are nice-to-haves. Admitting this is painful but necessary.
Example applications:
“Customer X will churn” → How much revenue? Is it actually guaranteed? What’s the real probability?
“Sales can’t close deals” → How many deals? Can they close some deals without it? What’s the real constraint?
“System will fail” → What specifically fails? When? Is that timeframe actual or estimated?
Step 3: Identify the True Constraints
At any given time, your startup has 1-2 true constraints. Everything else is noise.
Common constraints:
- Acquisition: Can’t get enough users/customers
- Activation: Users sign up but don’t experience value
- Retention: Users leave after experiencing value
- Revenue: Users stay but don’t pay (enough)
- Referral: Users don’t tell others
Your constraint determines what matters. If you can’t acquire users, retention features are waste. If you can’t retain users, acquisition features are waste.
How to identify your constraint:
Look at your funnel. Where’s the biggest drop-off? That’s your constraint.
| Metric | Healthy | Concerning | Constraint |
|---|---|---|---|
| Visitors → Signup | >5% | 2-5% | <2% |
| Signup → Activation | >50% | 25-50% | <25% |
| Week 1 → Week 4 retention | >40% | 20-40% | <20% |
| Free → Paid conversion | >5% | 2-5% | <2% |
Step 4: Filter by Constraint
Now filter your Critical and Important items. Does this feature address your true constraint?
Example:
Constraint: Activation (only 20% of signups reach core value)
| Feature | Addresses Constraint? | Keep |
|---|---|---|
| Customer X’s request | No, they’re already activated | Deprioritize |
| Better onboarding | Yes, directly | Keep |
| Sales team feature | No, about acquisition | Deprioritize |
| Performance improvement | Maybe, affects activation | Keep |
This filtering is brutal. You’re explicitly saying “even though this matters, it doesn’t matter right now.”
Step 5: Apply the “One Thing” Rule
If you could only ship one thing this month, what would it be?
Not “what would you ship if you worked 80-hour weeks.” Not “what would you ship if everything went perfectly.” What would you ship with realistic resources and realistic timeline?
That’s your priority.
Everything else is backup.
Step 6: Create a “Not Now” List
The features you’re not doing need a home. Create an explicit “Not Now” list with clear reasoning.
| Feature | Why Not Now |
|---|---|
| Customer X feature | Doesn’t address activation constraint |
| Sales feature | Acquisition is not our constraint |
| Tech debt item | No imminent system failure |
This list serves two purposes:
- You can point to it when stakeholders ask
- You can revisit when constraints change
Handling the Hard Conversations
Prioritization isn’t a spreadsheet exercise. It’s a series of hard conversations.
Conversation 1: The Demanding Customer
Situation: Your biggest customer says they need feature X or they’ll leave.
What most founders do: Panic. Promise to build it. Disrupt the roadmap.
What to do instead:
-
Understand the actual need. What are they trying to accomplish? Often there’s a simpler solution.
-
Assess the real risk. Will they actually leave? Have they said this before? What’s their switching cost?
-
Make a business decision. Is this customer’s revenue worth disrupting your roadmap? Sometimes yes. Often no.
-
Communicate directly. “I hear you need X. Here’s our current focus and why. Here’s when we could realistically address X. Is there a workaround in the meantime?”
The key: Don’t let one customer’s urgency override your strategy. Big customers have leverage, but they’re not product managers.
Conversation 2: The Sales Team
Situation: Sales says they can’t close deals without feature Y.
What most founders do: Build everything sales asks for. End up with a feature graveyard.
What to do instead:
-
Get specific. Which deals? How many? What specifically was said?
-
Separate feature requests from sales skill issues. Sometimes “we need feature Y” means “I don’t know how to handle this objection.”
-
Look for patterns. One deal requesting Y is noise. Ten deals is signal.
-
Do the math. If feature Y takes 3 months and might close 5 deals at $10K each, is that worth $50K? Maybe. Maybe not.
The key: Sales teams optimize for closing the next deal. Product needs to optimize for all future deals. These sometimes conflict.
Conversation 3: The Technical Debt Advocate
Situation: Engineering says the codebase is collapsing and needs a rewrite.
What most founders do: Either ignore completely (disaster waiting to happen) or approve a 6-month rewrite (never-ending project).
What to do instead:
-
Get specific about risk. What specifically will fail? When? What’s the consequence?
-
Separate “should fix” from “must fix.” Tech debt is always accumulating. Not all of it is urgent.
-
Budget time, don’t approve projects. Instead of “we’ll rewrite the auth system,” try “we’ll spend 20% of engineering time on tech debt.”
-
Address the worst first. The scariest tech debt should be fixed first, even if it’s not the “right” refactor.
The key: Technical debt is real, but engineers often want to fix everything at once. Incremental improvement is usually better than big rewrites.
Conversation 4: Yourself
Situation: You’re attached to a feature idea you’ve had for months.
What most founders do: Build it anyway because it’s their idea.
What to do instead:
-
Apply the same tests. What happens if you don’t build it? Does it address your constraint?
-
Get outside perspective. Run it by someone without emotional investment.
-
Admit when you’re biased. “I really want to build this, which makes me suspicious of my reasoning.”
The key: Founder ideas aren’t automatically good ideas. Your judgment about your own ideas is compromised.
Prioritization Tools
I’m skeptical of frameworks, but some tools help.
Tool 1: The 2x2 Matrix
Simple and effective. Plot features on:
- Y-axis: Impact (high/low)
- X-axis: Effort (high/low)
| Low Effort | High Effort | |
|---|---|---|
| High Impact | Do First | Plan Carefully |
| Low Impact | Quick Wins | Don’t Do |
Limitation: Impact and effort are both guesses.
Tool 2: The “Who’s Asking” Analysis
Track where requests come from:
| Source | Count | Revenue Represented | Pattern |
|---|---|---|---|
| Customer A | 5 | $50K | Integration focus |
| Customer B | 3 | $30K | Reporting focus |
| Sales team | 12 | Unknown | Feature parity requests |
| Support tickets | 30 | N/A | UX confusion |
This reveals patterns. High volume from support might indicate UX issues worth fixing. Many requests from one customer might mean they’re wrong for your product.
Tool 3: The “Experiment First” Filter
Before building, ask: can we test this with an experiment?
| Feature | Experiment Alternative |
|---|---|
| New pricing tier | Run survey on willingness to pay |
| Dashboard redesign | Prototype and user test |
| AI feature | Manual version first |
| Integration | Landing page to gauge interest |
The benefit: Experiments are cheaper than features. If the experiment fails, you’ve saved months.
Tool 4: The “Kill One” Exercise
When you can’t decide between options, try: if you could only do one of these things in the next 3 months, which would it be?
The answer often clarifies thinking.
Common Prioritization Mistakes
Mistake 1: Prioritizing by Loudness
Whoever complains most gets built first.
The fix: Track all requests. Look at patterns, not volume. The quiet customer who cancels without complaint matters more than the loud customer who never leaves.
Mistake 2: Saying Yes to Everything
Afraid to disappoint anyone, you commit to everything. Nothing ships well.
The fix: Every yes is a no to something else. Make the trade-off explicit.
Mistake 3: Never Revisiting Priorities
You set priorities quarterly and treat them as sacred.
The fix: Revisit weekly. Constraints change. New information arrives. Priorities should adjust.
Mistake 4: Prioritizing by Seniority
The CEO’s ideas automatically take priority.
The fix: Apply the same tests to everyone’s ideas, including leadership. The best idea wins, not the most senior advocate.
Mistake 5: Ignoring Dependencies
You prioritize features independently without considering what enables what.
The fix: Sequence matters. Sometimes you need to build B before A makes sense, even if A has higher impact.
FAQ: Feature Prioritization
How do I balance customer requests with product vision?
Customer requests tell you what they want today. Product vision tells you what they’ll need tomorrow. The balance depends on stage: early-stage companies should weight customer requests heavily (finding PMF). Later-stage companies can weight vision more (building competitive moat).
How often should I re-prioritize?
Review weekly. Adjust when new information meaningfully changes the picture. Don’t change daily (chaos) or quarterly (too rigid).
How do I say no without damaging relationships?
Be direct, give reasoning, offer alternatives. “We can’t do X right now because we’re focused on Y. Here’s why Y matters. Here’s when we might revisit X. In the meantime, here’s a workaround.”
What if my team disagrees with my prioritization?
Listen carefully to disagreement. Are they seeing something you’re missing? If you still believe in your prioritization after understanding their view, explain your reasoning and make the call. You can’t consensus your way to good priorities.
Key Takeaways
- Standard frameworks fail because they assume rational actors, accurate estimates, and knowable impact. Real prioritization is messier.
- Apply the “What happens if we don’t?” test. Most features are nice-to-haves when you’re honest.
- Identify your true constraint. Everything that doesn’t address your constraint is lower priority, regardless of who’s asking.
- The hardest part is conversations, not spreadsheets. Telling your biggest customer “not now” is harder than scoring a backlog.
- Revisit weekly. Priorities aren’t permanent. New information should change your thinking.
What’s Next
Look at your current backlog. For each item, ask: “What happens if we never build this?” Be ruthless. Most items should be “nice-to-have” or “nothing happens.”
Then identify your constraint. What’s the biggest drop-off in your funnel? That’s what deserves your attention.
Prioritization isn’t about finding the perfect order. It’s about having the courage to focus on what matters and say no to what doesn’t, even when saying no is uncomfortable.
Related Reading:
- Why Product Roadmaps Fail - When planning becomes a trap
- Product Strategy for Early-Stage Startups - Strategy before tactics
- North Star Metrics - Focusing on what matters
- The Cost of Building Too Early - Why less is often more
Related Articles
From idea to traction
The stuff most founders learn too late. One email per week.