On network architecture, social capital, and why platforms have to choose what kinds of ties they are optimizing for
Most social and community products would love to promise two things at once: intimacy and expansion. They want deep trust inside groups and broad movement across groups. They want retention from strong bonds and growth from new exposure.
That is understandable. It is also where a lot of strategic confusion starts. Bonding and bridging are not the same thing. They produce different kinds of value, different product behaviors, and often different trade-offs.
Bonding ties deepen closeness inside a group. Bridging ties connect people across groups. Put more simply: one is about cohesion, the other about movement.
Products optimized for bonding usually get better at trust, belonging, and local retention. People return because the environment feels familiar and socially dense. Group chats, small communities, and private collaboration products often benefit from this logic.
Products optimized for bridging usually get better at discovery, diffusion, opportunity, and reach. They help information or relationships cross boundaries that would otherwise remain intact. Professional networks, creator discovery systems, recommendation engines, and job platforms often benefit from this logic.
The strategic mistake is to talk about network effects as though all connection is equivalent. It is not. The same product decision can strengthen one kind of tie while weakening another.
You can see the difference in the contrast between products like WhatsApp and LinkedIn, or between private communities and recommendation-driven media platforms. WhatsApp gets much of its strength from already-meaningful relationships. LinkedIn gets much of its value from lighter ties that create mobility. One helps maintain the inside of the group. The other helps users move beyond it.
TikTok is interesting because it made bridging feel native. It did not depend primarily on who users already knew. It helped content cross clusters through behavioral similarity. LinkedIn's large-scale weak-ties research points in a related direction: opportunity often moves through people who are not especially close.
Neither logic is inherently superior. But teams do have to choose what kind of network they are primarily building, because that choice shapes discovery, trust, moderation, group design, and the very feel of the product.
Are we deepening the inside of clusters, or helping people move between them? Does our recommendation system mostly reinforce what users already know, or does it create adjacent exposure? Are group tools making belonging stronger, or making the product harder to escape from socially? Are discovery systems surfacing new voices, or only magnifying already-dense networks?
These are strategy questions, not just growth questions. A bonding-heavy product may feel warm, durable, and culturally coherent while struggling to grow outside existing circles. A bridging-heavy product may scale faster while feeling less intimate and more volatile. Both patterns are predictable if you name the tie structure honestly.
This is why I think of this essay as the architectural companion to Weak Ties Are a Growth Strategy. That essay makes the growth case for weak ties. This one makes the design case for being explicit about the type of social capital the product is producing.
Not all growth is the same, and not all connection is the same. Product strategy gets sharper when it stops treating network effects as a monolith and starts asking what kinds of ties the architecture is actually building.