The Business Collective

Platform Architecture

How Match Scoring Works

Every match is computed by a multi-layer scoring engine that evaluates profile alignment, behavioral signals, AI-assessed compatibility, and real community deal data. Here's exactly what goes into every score.

Master Formula

baseScore = Layer1 + Layer2 + Layer3 + Layer4 + Layer5 + Layer6 + Layer7
multiplier = 1 + (avgReadiness − 2) / 4 × 2  →  range: 1.0× – 3.0×
finalScore = baseScore × multiplier

Scores are unbounded but practically range from 0 to ~60+. Top 5 matches per user are stored. Higher scores = stronger predicted value.

What it checks: Does the other person's offer categories include what you are actively looking for?
First overlap
+3 pts
Any single need of yours is covered by their offers
Each additional overlap
+1.5 pts (max +3)
Deeper overlap = stronger direct match, capped at 3 overlaps
Maximum from this layer
+6 pts
directOverlap = myNeedTypes ∩ otherOfferCategories if (directOverlap.length > 0) score += 3 + min(overlapCount - 1, 2) × 1.5

This is the most fundamental signal — if someone provides exactly what you need, that's the core of a valuable connection.

What it checks: Does the other person need something you offer? This identifies two-way value — both parties have something to gain.
First reverse overlap
+2 pts
Each additional overlap
+1 pt (max +2)
Capped at 3 reverse overlaps
Maximum from this layer
+4 pts
reverseOverlap = otherNeedTypes ∩ myOfferCategories if (reverseOverlap.length > 0) score += 2 + min(reverseOverlap.length - 1, 2) × 1

Reverse matches create mutual-value relationships — not just vendor/client pairings but true partnerships where both parties walk away with something.

What it checks: Are both businesses at a similar stage? Startups often align best with other startups; established businesses with established ones.
Stage difference ≤ 1
+2 pts
e.g., startup + growing, or growing + established
Stage difference > 1
+0 pts
1
Startup
2
Growing
3
Established

A stage-1 startup and a stage-3 established firm may have misaligned expectations around budget, pace, and scale — even if their categories align perfectly.

What it checks: Does each person's stated collaboration intent (Buying, Selling, Partnerships, All) complement the other's?
Buyer ↔ Seller
+3 pts
Perfect transactional complement
Partnerships ↔ Partnerships
+2 pts
Both explicitly seeking collaborative arrangements
Either party = 'All'
+1 pt
Flexible intent increases compatibility
No alignment
+0 pts
if (A='Buying' && B='Selling') score += 3 if (A='Selling' && B='Buying') score += 3 if (A='Partnerships' && B='Partnerships') score += 2 if (A='All' || B='All') score += 1
What it checks: Do both people target the same type of customer (Individuals, Small Business, Medium Business, Government, Non-Profit)?
Exact match on ideal customer type
+1 pt
Creates referral and co-marketing potential
No match
+0 pts

Two businesses serving the same customer segment can refer to each other, co-market, and build bundled offerings — even if their services are completely different.

What it checks: The AI reads both members' open-ended questionnaire answers and scores their qualitative compatibility — goals, challenges, ideal connections, and open opportunities.
Exceptional alignment
+5 pts
Strong alignment
+4 pts
Moderate alignment
+3 pts
Some alignment
+2 pts
Minimal alignment
+1 pt
No alignment
+0 pts

Fields evaluated by AI:

  • • Goal — #1 business goal right now
  • • Challenge — Biggest current challenge
  • • Ideal Connection — Who they most want to meet
  • • Open Opportunities — What they're open to exploring

The AI considers complementary goals, mutual problem-solving potential, whether their ideal connection description matches the other person, and shared opportunity spaces. This layer catches nuanced alignment that structured fields miss.

What it checks: Has the community already proven that businesses with similar offer/need profiles close deals together? Have these two specific people already worked together?

Part A: Community Deal Pattern Bonus

All Closed Won deals are aggregated across the entire community. The system builds a success pattern map:offerCategory::needCategory → # of closed deals. When scoring a new pair, if your offer categories have historically closed deals for their need types (or vice versa), you earn a bonus.

Per pattern hit (bidirectional)
+1.5 pts per hit
e.g., 'Marketing offers → Customers need' closed 3 times = +4.5 pts
Maximum deal pattern bonus
+4 pts (capped)
// Count how many times my offers matched their needs in closed deals patternHits = Σ closedDeals where (myOfferCat::theirNeedCat) OR (theirOfferCat::myNeedCat) dealPatternBonus = min(4, patternHits × 1.5)

Part B: Direct Collaboration Bonus

If these two specific members have already closed a deal together, they receive a strong bonus. This surfaces re-engagement — people who've already proven they can work together should stay connected and collaborate again.

These two users previously closed a deal
+5 pts
Strong signal — proven collaborators should be re-matched
No prior deal history
+0 pts
Why this matters: As your community grows and more deals are logged, the matching engine learns what actually works — not just what looks good on paper. The more deal data flows in, the smarter and more personalized every match becomes.
What it does: Amplifies the base score based on how ready both parties are to take action — factoring in urgency and budget readiness from both sides.
avgReadiness range
2.0 – 6.0
(myUrgency + myBudget + theirUrgency + theirBudget) / 2
Multiplier at minimum readiness (2.0)
1.0×
No amplification
Multiplier at maximum readiness (6.0)
3.0×
Score tripled — both parties are highly motivated
avgReadiness = (myUrgency + myBudget + theirUrgency + theirBudget) / 2 multiplier = 1 + (avgReadiness − 2) / 4 × 2 // Urgency: 1=low, 2=medium, 3=high // Budget: 1=not ready, 2=planning, 3=ready to spend

A perfect profile match between two people who aren't ready to act is less valuable than a good match between two motivated, budget-ready operators. The multiplier ensures match rankings reflect real opportunity, not just theoretical compatibility.

A special flag set when all conditions for an immediately actionable deal are met. Flagged matches are highlighted in conversation starters and match explanations.
highProbability = true if ALL of: ✓ directMatch (their offers cover your needs) ✓ myUrgency === 3 (you are highly urgent) ✓ myBudget === 3 (you are budget-ready) ✓ theirUrgency >= 2 (they are at least medium urgency) ✓ theirBudget >= 2 (they are at least planning budget)

High-probability matches trigger unique conversation starters ("How quickly can you take this on?") and are explicitly called out in the match explanation to help members prioritize their time at events.

Future Scoring Enhancements

Event Co-Attendance: Members who attended the same events and were matched score higher — proven shared context.
Feedback Loop Scoring: Post-event feedback ('Did you make a valuable connection?') feeds back to weight certain profile pairs higher.
Deal Velocity: How quickly deals move from Identified → Closed becomes a signal for predicting which connections convert fastest.
Referral Network Depth: Members who frequently refer each other get a community trust bonus in scoring.
Category Saturation Awareness: If a need category is over-supplied by existing matches, the engine can diversify recommendations.

Quick Reference: All Score Sources

Direct Match (1st): +3 Direct Match (extra): +1.5 each Reverse Match (1st): +2 Reverse Match (extra): +1 each Stage Match: +2 Buyer↔Seller: +3 Partnerships↔Partnerships: +2 'All' intent: +1 Ideal Customer Match: +1 AI Alignment (max): +5 Deal Pattern (max): +4 Direct Collaboration: +5 Readiness Multiplier: +1×–3×

Max theoretical base score (no multiplier): ~28 pts. With 3× multiplier: ~84 pts. Practical top scores range 15–50 depending on community data.