I've been watching a portfolio company hold three shapes at once. Same technology — a sync primitive most of the industry doesn't fully understand — packaged three ways for three worlds. One shape has traction. Another is emerging around agentic workloads, pulling the product somewhere new. Underneath both, the original science that predates either container.
These shapes pull in different directions. Different customers, different urgency, different energy in the room when each one comes up. You can feel it — a kind of productive discomfort, the conversation shifting register depending on which shape is on the table. The founders haven't collapsed it. They're sitting with it, letting it teach them something that neither shape could teach alone. Every board meeting I leave thinking: they're getting smarter faster than companies that picked a lane.
When shapes meet, the instinct is to resolve. To pick. To collapse the ambiguity into something that fits on a slide. I keep watching what happens when founders resist that instinct.
Sometimes shapes don't press against each other on their own. Everyone agrees. The shape settles. That quiet consensus is more dangerous than any friction. A product without competing shapes pulling on it isn't learning anything new about itself.
If the friction isn't there, create it.
Another portfolio company built something beautiful for data practitioners. Composability was the point. People who used it could feel the difference in their hands. A felt shape: precise, opinionated, loved.
Then the market moved. A different buyer started pulling on the product. Natural language made the technical surface less relevant. Agents became the primary consumers. What mattered wasn't composability anymore. Reliability. Auditability. "Just works."
The instinct was to resolve. Pick a lane. The board wanted clarity. Investors wanted a narrative.
They didn't pick. And I made it harder for them.
I built a prototype — rough, provocative — that showed a different made shape for their technology. Another shape thrown into a room already full of competing shapes, to see what happened when the friction intensified.
Months of late-night Slack messages followed. Co-iterations. Many hands. What emerged wasn't a compromise. It was a third thing — carrying what the market demanded while reshaping what the original tool could become. What survived wasn't unchanged. The friction reshaped it. It crystallized because the friction was held long enough — and sometimes fed. Nobody saw it at the start. That was the point.
One I got wrong.
A portfolio company with a strong founder shape and a natural fit with early customers. Top-down enterprise. High-value contracts. Long cycles. Every meeting confirmed the thesis. Everyone in the room in the same shape, nodding.
By the time shape work started, everything had set. I could feel the absence of any competing pull. Conversations about the product's future circled back to the same assumptions. Nothing soft left to push on.
When the ruled shapes arrived — what follow-on investors needed to see, what the runway demanded — the product couldn't move. They sold to a larger organization that could hold that shape better than a startup could.
Shape work can't be bolted on when things get stuck. It needs to run from the beginning — when conviction is strong but the form is still loose, when there's still room for friction to enter and generate information. After the shapes get dense enough that changing them costs more than most startups can afford, you're not doing shape work. You're doing renovation.
Friction dies in predictable ways. I've watched it happen enough times to recognize the patterns.
The board wants clarity. "Are we X or Y?" The room exhales. The binary feels like progress. But the answer was something without a name yet — a third shape the tension was still producing. The binary killed it. I call this the narrative collapse: the moment when the need for a clean story destroys the messy signal.
Tension gets uncomfortable. Someone reaches for the most measurable shape. Pipeline numbers, activation rates, feature completion. It takes over — not because it matters most, but because it's most legible. Felt shape and enacted shape can't defend themselves in a spreadsheet. That's the metric grab. It looks like rigor. It's a retreat.
A design partner asks for something. The team builds it. Then another partner, another request. Each one reasonable. The accumulation reshapes the product around the loudest voice rather than the deepest signal. Customer capitulation — not any single decision, but the slow drift of building for whoever's in the room rather than what the friction is trying to teach you.
A round is coming. The team picks the shape easiest to explain to investors. This is when most founders first encounter something like shape work — they call it "storytelling." But storytelling is picking the most fundable narrative. Shape work is understanding the system of tensions you're inside. They look similar from the outside. They produce very different products. The fundraise snap is the most common, and maybe the most forgiven, way to kill friction. Everyone understands why it happens. The product still pays the price.
Every one of these is the same reflex — making friction go away by collapsing into one shape. Every one destroys information.
The diagnostic: is this friction generating new information, or generating paralysis? If it's generating information — stay in it. If paralysis, the friction needs to be reshaped. Not resolved. Reshaped.
The sync company held their friction long enough to produce real results — three shapes from the same science, each one a different conversation with the world. The data tool founders held theirs long enough to produce real insight — the kind you only get from months inside competing pulls. The company that settled early never had friction to hold. They shipped a product that looked right from the outside. It wasn't earned.
I think this is the thing that's hardest to see about shape work. Every act of collapsing tension — picking a lane, choosing a narrative, resolving the ambiguity — destroys information. The friction between shapes is where a company learns things about itself that no single shape could reveal. The companies that hold competing shapes long enough don't just endure the discomfort. They accumulate a depth of understanding — about their technology, their market, the space between what's possible and what's adoptable — that companies who resolved early never had access to.
The shallow version of a product is the visible surface — what competitors can see and replicate. The deep version is everything the tension kept alive. The options you can see because you let the friction keep generating signal instead of shutting it down. That information doesn't show up on a dashboard. It lives in the company — in the conversations that kept going, the shapes that kept pulling, the questions that wouldn't resolve.