The HAEVN matching model.

Every potential match moves through the same structured lifecycle: members define compatibility criteria, the system applies deterministic Gates, Weights measure relative compatibility, both sides must qualify independently, and qualified results are surfaced on a weekly schedule.

Matching lifecycle

Define → Gate → Weight → Qualify → Release → Connect

The rest of this page walks each stage in order.

Most modern matching systems rely on inferred behavioural signals — swipes, likes, click behaviour, dwell time, engagement loops. Those systems try to guess intent from reactive activity.

HAEVN takes a different approach.

Everything begins with explicit member input. The survey establishes the structured compatibility criteria the system will use, before behavioural interaction becomes relevant at all.

Compatibility dimensions may include

• Relationship goals and Shared Intent
• Relationship structure preferences
• Orientation and gender compatibility
• Geographic relevance
• Age parameters
• Values and boundaries
• Lifestyle and communication style
• Emotional and interpersonal indicators

The goal is not to predict chemistry. The goal is to establish a structured framework for deciding whether an introduction is worth the members’ time.

Once a member’s profile and qualification criteria are established, they become a standing query against the active local network.

Traditional browsing platforms require the member to search, filter, review, and repeatedly evaluate candidates. HAEVN reverses that model: the member defines the qualification logic once, and the system evaluates the network against it.

Evaluation is not a live browsing feed. It runs on a weekly cadence tied to the Match Monday release. There is no continuous scanning, no always-on scoring loop for members to check.

The matching itself is rules-driven. A structured system of Gates and Weights reads member answers and produces a Match Score. It is not a generative model, and it is not a black-box recommendation engine.

AI is used only to translate the structured result into a readable, human explanation of why a Match is a Match — the shared signals, the common ground, the tensions worth naming.

AI does not decide who someone will love. It does not score attraction. It does not rank chemistry. Those questions aren’t engineering problems, and HAEVN isn’t trying to answer them.

Some compatibility criteria function as absolute requirements rather than weighted preferences. Those are Gates.

A Gate is a deterministic binary qualification condition. If a candidate fails a required Gate, evaluation ends and the candidate is discarded from that member’s network — no further scoring occurs.

Gates may include

• Active local market eligibility
• Geographic proximity
• Age compatibility
• Orientation compatibility
• Gender compatibility
• Relationship structure compatibility
• Declared hard exclusions

Gates prevent compatibility inflation between members who may share broad similarities but fail a fundamental requirement.

Once a candidate clears every Gate, evaluation moves into Weights.

Not every compatibility signal carries equal importance. A system that treats all preferences equally tends to overstate compatibility by rewarding superficial overlap. Weights let HAEVN reflect the relative importance of different signals — how much each contributes to the overall picture of fit, in proportion to the others.

This page does not publish the specific weights the system uses. Those are part of the model, not the marketing. What matters here is the shape: identity, intent, and structural compatibility carry more weight than surface-level overlaps.

Weights allow the system to distinguish between broad similarity and meaningful compatibility. The output is a Match Score: a measurement of compatibility within the HAEVN system, not a probability of relationship success.

Compatibility must work in both directions.

A candidate is not qualified simply because they strongly fit one member’s stated preferences. The system evaluates fit from each member’s perspective independently, using that member’s Gates and Weights.

A candidate may clear the threshold for one member while failing to clear it for the other. In that case, no Match is surfaced. Both sides must qualify — separately — before an introduction becomes possible.

This is a qualification requirement, not the connection step. What happens after a Match is surfaced is covered in the Connection Model below.

Why matching begins at 80%.

The baseline threshold for a Match is 80% mutual compatibility. Below that, an introduction has not earned the word Match, and HAEVN does not surface it as one.

Lower thresholds produce larger volumes of introductions, but also noise, weaker fit, and member fatigue. Higher thresholds reduce volume while improving confidence that what does surface is worth attention.

Recommendations sit at 77–79%.

Some pairings fall just below the Match threshold — in the 77–79% range. These are surfaced as Recommendations rather than Matches: blind by default, and only revealed to both members if both accept.

Neither threshold is presented as a guarantee of chemistry or relationship success. No structured model can reliably predict human chemistry. The practical objective is narrower: improve introduction quality by requiring meaningful compatibility before connection becomes possible.

Structured matching is only as reliable as the dataset it operates against. HAEVN includes identity verification via Veriff because compatibility logic loses credibility inside a network containing duplicate accounts, fabricated profiles, bots, or anonymous participants with no accountability.

HAEVN retains the result of the verification check — pass or fail — and does not store the underlying identity documents.

Verification supports

• Stronger dataset integrity
• Reduced impersonation and fraud risk
• Greater trust in surfaced matches
• Lower spam risk
• Real accountability inside the network

Verification does not affect compatibility scoring. Its role is to strengthen trust in the participant layer the matching engine evaluates.

HAEVN operates as a collection of local networks rather than a single undifferentiated national pool. Every member sits inside one active local network at a time.

Why local, not national

• Geographic viability — matches you can actually meet
• Stronger local density over time
• Fewer mathematically-interesting-but-impractical results
• Real-world connection potential rather than a distant pen-pal

National-scale compatibility can produce fascinating pairings that are impractical in real life. HAEVN prioritises local network relevance because geographic viability materially affects connection outcomes.

Qualified introductions are not surfaced continuously. They are released on a weekly schedule — Match Monday.

Every Monday, each active member receives one of three possible statuses: a Match (80%+), a Recommendation (77–79%), or no new introduction that week. There is nothing to check between Mondays, and no benefit to opening the app more often.

This architecture intentionally separates compatibility evaluation from engagement mechanics. Traditional browsing platforms benefit when members return, swipe, and remain behaviourally active. HAEVN doesn’t.

HAEVN is not an open profile marketplace. Members are not universally visible to one another.

Visibility is governed by compatibility qualification. Exposure to another member happens only when structured relevance exists — either as a Match or, with mutual acceptance, as a revealed Recommendation.

Network Membership shows that a Match exists and provides the Match explanation. HAEVN+is what unlocks photos and mutual introductions once both members accept.

This visibility model reduces spam behaviour, low-fit outreach, and popularity-driven distortions common in public matching environments.

A qualified Match creates the opportunity for connection. It does not automatically create interaction.

Connection is mutual by design. Both members must independently choose to proceed before a conversation opens. Nothing is unlocked one-sidedly, and there is no cold outreach dynamic — no unsolicited messages, no browsing-and-picking a stranger.

This preserves member agency on both sides. Compatibility gives you a reason to consider someone. The decision to meet is still, and only, yours.