People already loved their matches. We rebuilt the engine anyway. Here's why similarity isn't enough.
TL;DR: We've rebuilt our matching engine from the ground up. The old system matched on similarity, which worked but kept people in echo chambers. The new system calculates value for both sides independently, separates who you are from where you're going, and optimises across the whole cohort so one perfect match doesn't come at the cost of five good ones.
People love their matches. This means not just "it was good" level enthusiasm, but genuinely enthusiastic post-match responses. The conversations feel relevant, the awkwardness we feared hasn't materialised. One of our earliest business hypotheses has been checked off. Which makes this week's decision a bit counterintuitive: we've completely rebuilt the matching engine from scratch. So if it's working, why touch it? Because whilst our initial engine did a great job of delivering matches that look good on paper, it missed exceptional, less intuitive matches. Matching on similarity only means you'll continually get more of the same. The new system is reflexive, changing based on where you want to go next.
Our original engine was simple: find people who overlap. Similar careers, similar seniority, similar industries. The net result was safe, sensible matches every time (perfect for a V1!). The trouble is similarity misses the most important thing about a career: where you're trying to go. And where you're trying to go is something that's always changing. It's reflexive. For example we have a Data Scientist in the community who wants to become a Product Manager. The old system would pair them with another Data Scientist. They'd have a pleasant chat about data pipelines, but miss someone who could help with the transition. What they actually needed was a Product Leader open to giving the inside scoop on being a PM. Similarity creates echo chambers. Our aim is to build bridges.
The new engine changes the question. Instead of asking "How similar are Jane and John?" we now calculate two separate scores:
This handles real-world nuance. If George (Data Scientist wanting to become a PM) matches with Jane (VP of Product), he gets high value from her experience—but what about Jane? The system only creates the match if Jane has signalled interest in mentoring, or has a pattern of enjoying exploratory conversations. Both sides need to benefit, or it doesn't happen. Your preferences shape your matches too. If you're an explorer, we'll push you outside your arena. If you want depth, we'll keep you close to home.
We won't bore you with the math, but a few details that matter: We separate who you are from where you're going. Your job title says "Data Scientist." Your goals say "Product." Most systems would match you on the title. We match on both signals independently, which is how a Data Scientist ends up talking to a Product Leader instead of another Data Scientist. We optimise for the whole cohort, not just individual pairs. Greedy matching—always picking the "best" pair first—can leave others with weaker options. We solve for the best set of matches collectively, which sometimes means a 95% pair becomes 85% so that others jump from 60% to 85%. Reliability earns you better matches. Show up, respond promptly, respect people's time, and the system notices. New members start neutral, but over time, reliable people get matched with other reliable people. Humans review every cycle. The algorithm generates candidates. Our team applies the intuition test before introductions go out. Sometimes a pairing looks perfect mathematically but feels off for reasons the model can't see.
We're building for a world where your network grows in the direction you actually want to go. That takes more than similarity. It takes intention—yours and ours.
Ivan Franco – Co-founder & COO
Through my career, I've used data to solve complex problems. With Flynt, I'm applying that same analytical thinking to something more personal: helping you find the right people to connect with at exactly the right time in your career.
Connect with inspiring professionals once a month. Apply below and we'll be in touch when a spot opens up.
