This is part 3 of a series on the Economics of Matchmaking. Read up on the previous entry here.
1. Qualitative Analysis
You found your dream job. The benefits are great. Work is engaging. Everyone’s so collaborative.
Six months in, your boss pings you at 2.30 am. Work piles up. Colleagues leave you on ‘read’.
Now picture this:
You met someone. They’re fine in photos (even better in person). But they’re kinda boring. Dry texts. Jokes hit and miss. But you gave it a shot.
Not a month into the relationship, you cannot imagine being with anyone else.
Finding a match was only the prologue. The real plot is whether the fit holds, whether the relationships thrive or dissolve. Economists call it match quality.
2. Goods come in Threes
Economists classify certain goods by how difficult it is to assess their quality:
Search Goods (or inspection goods):
Things that you can assess upfront. You can try out the shirt and inspect the seams. What you see is what you get.
Experience Goods:
You only learn about their quality after consuming them. Like trying out a new restaurant, or booking your next travel destination. You can’t knock it (or love it) until you try it.
Credence Goods:
These things are the trickiest to evaluate. Even after using them a couple of times, you’re still unsure what to make of it. Like hiring management consultants. How can you tell if that expensive procedure actually works?
Relationships sit uneasily between experience and credence goods: You only really know by living them. Even then, the attribution is messy.
There’s no good way of assessing the right fit upfront. First impressions are just that, impressions. You think you’re signing up for one thing. Three quarters into a year, you’re dealing with something else altogether. That algorithmically-optimized well-curated dating profile? Tells you nothing about how well that person would actually treat you.
3. What are we?
Even with experience, match quality is still hard to pin down and quantify. Moving from one relationship to another does not yield better insight on match quality. There’s seeming patterns, but each match is idiosyncratic.
A way to measure match quality is to take the direct approach. Companies evaluate if the worker is a good fit through operational KPIs. If there’s no complaints, and targets are being met on schedule, it means he’s doing well. You can’t quite do that with personal relationships though.
Another way is simply to ask people how they feel. Employers can survey their employees, get feedback on their engagement. Couples can check-in with each other. Think tanks also send out surveys to the population, asking how satisfied they are with their job, their partners, and even their life.
But surveys are flawed. There’s biases baked in. There’s ambiguity. Does an average score of 3 out of 5 mean they are indifferent about their relationships? Or borderline (dis)content?
Then there’s the indirect approach, by observing signals that proxy for match quality like tenure (duration of the relationship) and separation rates. The longer the tenure, or the lower the risk of separation, the better. Right??
4. Learning the Hard Way
One model treats match quality as hidden at the start and revealed by experience. This is the learning model, developed by Boyan Jovanovic.
Workers start their jobs being uncertain about the fit, or how productive this working relationship can be. They may have a belief about this inherent quality. They learn about and discover the match quality on-the-job. Successful outcomes like smooth product launches, or steadily growing margins indicate that you’ve made a good match. Missed targets and poor communication are negative signals. It’s a tug between these positive and negative signals. Workers quit their jobs when they believe the match’s quality drops below a certain threshold.
This risk of separation—the hazard rate—is higher early in the relationship, but declines over time. People don’t immediately quit because they still have to learn about the relationship, but overall, flight risks are higher in that early phase. Bad matches are assumed to be weeded out quickly. This turnover and churn isn't a dysfunction of the system. It’s just how the market filters out bad fits. Turnover is learning made visible.
But there’s a catch: you can’t learn about the match quality without committing to it. But that commitment is the very risk that makes a bad fit costly.
It’s why organizations put fresh hires under probation. It’s a way to de-risk the process, before hiring them on a permanent basis. To learn about how well the worker fits into the organization’s culture. It’s almost a win-win strategy. The employer gets to utilize workers’ skills and productive time, while the employee gets paid and some material security. It’s like a corporate situationship—benefits without total commitment.
The learning model also explains why people go on dates—to learn more about the person. To figure each other out before making a full commitment. Some couples may live together (cohabit) for a period of time before marrying.
5. The Times, They are a-Changin’
An alternative view comes from Ioana Marinescu’s changes model, which treats match quality as dynamic. It changes with time and shocks. External shocks—natural disasters, technological change, sudden layoffs—can push an otherwise good match toward the exit threshold, or pull a middling match back into the safe zone.
Both models generate declining hazard rates, but they differ on responses to shocks. In learning, a negative shock can scar, but fades with tenure (i.e. longer term relationships only feel a sting). In changes, shocks are temporary, but hurt longer-term mediocre matches.
A firm may get disrupted and be forced to course correct. Sudden deaths in the family might foster closeness in their shared grief or drive a wedge between couples.
Where the learning model implies adopting a strategy to manage risks on selection. The changes model suggest that long-run stability depends on ongoing adaptation as much as on initial selection.
6. Quality Investments
If match quality can change, then a good selection is only the starting point, the real work is on the continuing investment in the relationship.
Companies want to retain their best employee, so they fund training, redesign roles, and coach managers. Whereas workers who want to get better at their jobs can upskill, seek feedback proactively, and renegotiate scope. All of it improves the match quality and widens the ‘buffer’ that helps this working relationship absorb shocks.
For couples, they too can become more intimate with time and effort. They can improve individually because of their partner, and as a pair. They learn to communicate better and compromise, they are more accepting of each other’s quirks, perhaps they build projects together (like a nest egg or raise children).
Investment isn’t optional. It’s the price of quality relationships.
7. QA/QC
If match quality can improve (or deteriorate), then that initial spark is overrated. What matters more is the active, ongoing investment in the relationship.
Dismiss a poor match too early, we risk missing out if it was given a chance to flourish. Stay too long, you burn time in a relationship already in decline. Both errors are costly, and irreversible.
People change. Circumstances change. A beautiful partnership can drift apart, or simply run its natural course.
Unlike manufacturing, there’s no ISO that governs match quality. No manual on the perfect match. Great matches don’t just happen, they’re built. They last only with attention, intentional care, and an openness to change.