Ask most people what makes a good relationship and you'll hear "chemistry," "communication," and "trust." Not wrong exactly — but not quite right either. These are descriptions of what good relationships have, not what produces them. Chemistry is the result of compatibility, not the cause. Communication is a skill that works when the underlying dynamic supports it. Trust is earned through behaviour patterns, not maintained by will.
Forty years of couples research — Gottman, Sternberg, Aron, Fisher, and others — has produced a surprisingly clear picture of what actually predicts relationship quality and longevity. It's not what most people expect.
The factors that matter (and how much)
Shared values — especially around what life is for
High predictorNot shared hobbies. Not shared Netflix tastes. The deep values: how you think about money and generosity, what you believe about family and work, how you handle difficulty, what kind of life you're trying to build and why. Research by Rosenfeld and Bulatao found that value alignment was among the strongest predictors of long-term satisfaction — more powerful than personality similarity. Two people with genuinely different values can sustain enormous attraction and affection for years before the friction creates insurmountable conflict.
The ratio of positive to negative interactions
High predictorGottman's research found that stable couples maintain approximately a 5:1 ratio of positive to negative interactions during everyday life — and a 20:1 ratio during conflict. This doesn't mean suppressing negativity. It means that the background level of affection, appreciation, humour, interest, and warmth is high enough to sustain the relationship through the inevitable difficult moments. The ratio is maintained through small acts, consistently — not through major gestures occasionally.
Compatible attachment styles
High predictorTwo securely attached people have a relationship advantage that is difficult to overstate. Their nervous systems don't amplify each other's insecurities. Conflict doesn't trigger existential fears. They can give each other space without interpreting it as withdrawal. Research consistently shows that anxious-avoidant pairings produce significantly higher distress than any other combination, and secure-secure pairings report the highest satisfaction. You can work across attachment styles — but it takes more effort for the same outcome.
How conflict is handled, not how much there is
High predictorConflict avoidance and conflict frequency are both less predictive of relationship failure than conflict style. Gottman's Four Horsemen — criticism, contempt, defensiveness, stonewalling — are the reliably destructive patterns. Their presence predicts relationship failure with around 90% accuracy. Their absence, combined with a willingness to repair after conflict, is one of the strongest markers of a genuinely healthy relationship. The research doesn't reward conflict-free couples; it rewards couples who fight well.
Active interest in each other's inner life
High predictorGottman's concept of "Love Maps" — how well you know your partner's current world: their fears, hopes, stressors, sources of meaning — predicts relationship resilience through major life transitions. Couples who fall into parallel lives, where each person becomes progressively less known to the other, report lower satisfaction over time regardless of the absence of explicit conflict. Being genuinely curious about who your partner is becoming — not just who they were when you met — is one of the most underrated relationship skills.
LoveCertain matches on the four highest predictors
Values 40%, Life Stage 25%, Attachment 20%, Communication 15%. Only matches above 70% shown. £49 once.
The myths worth discarding
"The right relationship shouldn't require work"
This one causes significant damage. Every long-term relationship requires active maintenance: attention, repair, investment of time and vulnerability. The research doesn't show that good relationships are effortless — it shows that the effort feels worthwhile and the baseline of goodwill is high enough that maintenance feels like care rather than obligation. "No work" is not a standard to aspire to; it's usually a sign the relationship is in its infatuation phase.
"Opposites attract — long-term"
Opposites do attract, initially. Research by Buston and Emlen found that on most dimensions people actually prefer similarity — similar values, education level, physical attractiveness ratings — despite often claiming otherwise. The "opposites attract" effect tends to describe surface-level novelty that diminishes over time, while the deep incompatibilities on values and life direction become increasingly visible. Short-term: novelty works. Long-term: similarity sustains.
"If you really love each other, you'll figure it out"
Love is necessary but not sufficient. Gottman's research found that love was not among the variables that predicted relationship stability — what predicted it was how couples treated each other day to day, independent of how much they professed to care. Two people who genuinely love each other can still produce a dynamic that is destructive if they have incompatible conflict styles, attachment patterns, or fundamental value misalignments.
"You'll know immediately if they're the one"
Instant chemistry tells you about physical and novelty-based attraction — which is real and worth noticing. It tells you very little about attachment compatibility, value alignment, conflict style, or long-term temperament match. Some of the most successful long-term relationships started with mild interest that grew as people actually knew each other. Some of the most intense immediate attractions produced short relationships that burned out completely.
What research consistently doesn't predict success
Several things that popular culture treats as important turn out to have little predictive power on their own: shared hobbies, physical attractiveness, educational level (beyond a moderate similarity effect), income beyond a baseline security threshold, and number of previous relationships. These factors matter in particular ways but are not reliable predictors of relationship quality.
Even shared values, the strongest predictor, isn't sufficient alone — it needs to be combined with mutual respect, compatible conflict styles, and the willingness to invest in the relationship actively.
The Certain Letter
Relationship science, no fluff. Weekly.
The Harvard Study on Adult Development
The longest-running study of adult happiness and health — 85 years, started in 1938, still ongoing — found that the quality of relationships is the single most powerful predictor of wellbeing in later life. Not wealth, not career success, not health status at midlife. Relationship quality. Specifically, the warmth and reliability of close relationships predicted cognitive health, physical health, and life satisfaction more reliably than any other measured variable.
The lead researcher, Robert Waldinger, summarised it plainly: "Good relationships keep us happier and healthier. Period."
What constitutes a "good relationship" in their data? Not conflict-free. Not particularly passionate in the conventional sense. Relationships characterised by mutual support, genuine understanding, and the sense that you could rely on your partner in difficult moments. The substance, not the performance.
What this means for how you date
If the research is right, the things worth looking for early in dating are not the things that dating apps optimise for. An ability to be curious about you without making it about them. Consistency between what they say and what they do. How they respond when something doesn't go their way. Whether they can acknowledge being wrong about something. How they speak about the people in their life.
The things that produce good relationships are mostly visible in ordinary behaviour, not in dramatic moments. The person who texts back consistently, shows up on time, asks how your day actually was, and apologises clearly when they've handled something badly — this person is demonstrating the raw material of a good relationship. The person with overwhelming chemistry who disappears for three days and reappears with an exciting story is demonstrating something else.
This is precisely the logic behind LoveCertain's compatibility model. Rather than relying on photo-based attraction and hoping the fundamentals work out, we measure the fundamentals first — values, attachment, life stage, communication style — and only show matches above 70% compatibility. The relationships that have formed through this approach start from a foundation that the research says actually predicts success.
Chemistry can grow from compatibility. It's harder for compatibility to grow from chemistry alone.
Meet someone who's genuinely compatible
LoveCertain uses the research above to match people who have what it actually takes to build something lasting. £49 once. Full refund if no relationship in 90 days. £99 bonus when you find one.
Join LoveCertain
