| name | ph-algorithm-guide |
| description | Understand how the Product Hunt ranking algorithm works. Use this skill to optimize your launch strategy based on known algorithm factors and behaviors. |
Product Hunt Algorithm Guide
This skill explains how Product Hunt's ranking algorithm works, helping you optimize your launch strategy based on publicly known factors.
When to Use This Skill
- Planning your launch strategy
- Understanding why rankings change
- Optimizing for algorithm factors
- Diagnosing ranking issues
- Setting realistic expectations
Algorithm Fundamentals
Key Insight
Upvotes ≠ Points
Product Hunt CTO Mike Kerzhner confirmed: "There is not a 1:1 correspondence between upvotes and points."
What This Means
- Not all votes count equally
- Account quality matters
- Engagement quality matters
- Timing patterns matter
Known Ranking Factors
Factor 1: Vote Weight
Higher Weight Votes:
- Older accounts (months/years old)
- Active accounts (regular engagement)
- Diverse activity (not just voting)
- Organic voting pattern
Lower Weight Votes:
- New accounts (recently created)
- Inactive accounts (created but unused)
- Single-purpose accounts
- Suspicious patterns
Potentially Discounted:
- Brand new accounts
- Accounts created same day
- Bulk votes from same source
- Coordinated voting patterns
Factor 2: Engagement Depth
Positive Signals:
- Thoughtful comments
- Discussion threads
- Maker responses
- Question-answer exchanges
Why It Matters:
- Comments indicate genuine interest
- Discussions show community value
- Engagement harder to fake than votes
Factor 3: Velocity Pattern
What Algorithm Watches:
- Rate of upvote accumulation
- Time distribution of votes
- Spikes vs steady growth
- Natural vs artificial patterns
Healthy Pattern:
Hour 1: [████████░░] 40 votes
Hour 2: [██████░░░░] 35 votes
Hour 3: [███████░░░] 38 votes
Hour 4: [█████████░] 45 votes
Suspicious Pattern:
Hour 1: [██████████] 150 votes (spike!)
Hour 2: [█░░░░░░░░░] 5 votes
Hour 3: [█░░░░░░░░░] 3 votes
Hour 4: [█░░░░░░░░░] 2 votes
Factor 4: First 4 Hours
Special Period:
- Rankings randomized initially
- Vote counts hidden publicly
- Algorithm observing patterns
- Critical for initial position
After 4 Hours:
- Rankings become vote-based
- Position reflects accumulated strength
- Top positions attract organic traffic
- Momentum becomes visible
Factor 5: Account Relationships
Flagged Patterns:
- Votes from connected accounts
- Same IP address votes
- Same device votes
- Employee/team votes (weighted less)
Clean Patterns:
- Diverse geographic sources
- Independent account histories
- Organic discovery paths
How Rankings Are Determined
The Daily Cycle
12:01 AM PST → Day begins
↓
Hours 0-4: Randomized ranking
↓
Hour 4+: Algorithm-sorted ranking
↓
Throughout day: Continuous re-ranking
↓
11:59 PM PST → Final rankings locked
↓
Awards: POTD, Top 5, etc.
Ranking Formula (Approximate)
Score = (Weighted Votes × Quality Multiplier)
+ (Engagement Depth Bonus)
- (Spam/Manipulation Penalty)
Where:
- Weighted Votes = Sum of all votes adjusted by account quality
- Quality Multiplier = Based on product profile completeness
- Engagement Depth = Comments, discussions, maker activity
- Penalty = Deductions for suspicious patterns
Optimizing for the Algorithm
Do: Quality Over Quantity
Instead of:
Getting 200 votes from low-quality accounts
Aim for:
Getting 100 votes from active, established accounts
Do: Stagger Engagement
Instead of:
All supporters voting at 12:01 AM
Aim for:
Supporters spread across 5-6 waves over 24 hours
Do: Encourage Real Comments
Instead of:
"Please upvote!"
Aim for:
"Would love your honest thoughts in the comments!"
Do: Respond to Everything
Why:
- Shows you're present
- Creates discussion threads
- Signals genuine launch
- Builds engagement depth
Algorithm Behaviors
What Triggers Scrutiny
-
Vote Velocity Spikes
- Sudden burst of votes
- Then dramatic dropoff
- Unnatural acceleration
-
Account Patterns
- Multiple new accounts
- Same creation timeframe
- Similar activity patterns
-
Geographic Clustering
- All votes from one location
- No geographic diversity
- Pattern doesn't match product
-
Timing Uniformity
- Votes in exact intervals
- Automated-looking patterns
- Unnatural consistency
What the Algorithm Rewards
-
Organic Growth
- Steady accumulation
- Natural peaks and valleys
- Timezone-appropriate waves
-
Diverse Sources
- Various account ages
- Different activity levels
- Geographic spread
-
Deep Engagement
- Multiple comments
- Discussion threads
- Question-answer pairs
-
Maker Presence
- Quick responses
- Genuine conversation
- Helpful attitude
Featured vs Unfeatured
Getting Featured
Requirements (Unofficial):
- Product is clearly explained
- Meets category standards
- No obvious manipulation
- Complete profile
Helps Your Chances:
- Quality visuals
- Clear value proposition
- Active maker engagement
- Previous PH presence
Getting Unfeatured
Common Causes:
- Vote manipulation detected
- Spam reports received
- Policy violations
- Low-quality product
Recovery:
- Usually not possible same day
- Contact support (respectfully)
- Learn for next time
Realistic Expectations
What You Can Control
- Quality of your product
- Quality of your assets
- Your community engagement
- Your response rate
- Your outreach authenticity
What You Can't Control
- Competitor strength
- Algorithm behavior
- Vote weighting details
- Featuring decisions
- Final ranking
Healthy Mindset
Focus on: Building something people love
Not on: Gaming the system
Focus on: Genuine community
Not on: Vote numbers
Focus on: Long-term reputation
Not on: One-day ranking
Algorithm Myths Debunked
Myth: "Having a famous hunter guarantees success"
Reality: 79% of featured products are self-hunted. Hunter followers help awareness but don't guarantee votes.
Myth: "More votes always means higher rank"
Reality: Vote quality matters more than quantity. 50 high-weight votes can beat 100 low-weight votes.
Myth: "The first hour determines everything"
Reality: First 4 hours matter, but the entire 24 hours count. Late momentum can overcome slow starts.
Myth: "Weekend launches are easy wins"
Reality: Lower competition, but also lower traffic. Easier badge, fewer users.
Myth: "The algorithm is random/unfair"
Reality: It's designed to surface genuinely interesting products. Work with it, not against it.
Output Format
ALGORITHM OPTIMIZATION CHECK
VOTE QUALITY:
- Expected high-weight votes: [Number]
- Expected low-weight votes: [Number]
- Risk of discounted votes: [Low/Medium/High]
ENGAGEMENT PLAN:
- Comment depth strategy: [Description]
- Maker response plan: [Description]
- Discussion seeding: [Description]
VELOCITY PATTERN:
- Wave 1 timing: [Time]
- Wave 2 timing: [Time]
- Expected distribution: [Natural/Concerning]
RISK FACTORS:
- [Risk 1]: [Mitigation]
- [Risk 2]: [Mitigation]
REALISTIC TARGETS:
- Conservative estimate: [Rank range]
- Optimistic estimate: [Rank range]
