I spent some time reviewing a large group of recently funded growth-stage SaaS companies. I was looking for industry patterns — where capital seemed to be concentrating across sectors like fintech, healthcare operations, logistics, or construction software.

But after scanning enough companies, the industry categories started to matter less.

What stood out instead were the underlying workflows these products were trying to organize: multi-role operational systems involving finance, operations, compliance, and field teams.

Different industries. Very similar coordination problems.

Across industries, healthcare coordination, last-mile delivery, field operations, financial workflows, investors are funding the same thing: companies that sit in the middle of a broken handoff.

Not broken because nobody tried to fix it. Broken because the handoff crosses too many stakeholders, too many systems, and too many ownership boundaries for any single party to hold the whole.

A healthcare coordination platform I looked at built its entire product around one insight: providers weren’t failing to refer patients. The referral was happening. The patient was just disappearing into a work queue on the other side — fax, email, portal, manual review — while everyone assumed someone else had it handled. The company tripled revenue in two quarters not by building smarter AI, but by designing around how the workflow actually operated, rather than how the software assumed it should.

The same structure shows up in last-mile delivery — drivers, dispatchers, recipients, and retailers each holding a piece of information the others need, with no single party able to see the whole handoff.

The pattern underneath the pattern

What kept showing up wasn't a technology gap in the products these companies were replacing. It was a coherence gap — and the more I looked at it, the more it pointed to something structural rather than situational.

There’s a principle in systems thinking worth naming here: there are no independent variables. In a complex system, optimizing one surface affects every other surface. The companies that struggle at growth stage aren’t struggling because they lack capability. They’re struggling because they’re treating their product surfaces, and often their customer segments, as if they were independent, when they’re not.

Every stakeholder optimized locally. Every team running its own definition of success. Every surface improved in isolation. Coordination debt is what accumulates in the gaps between them. It doesn’t announce itself. It compounds quietly — in slower decisions, in features that ship and don’t land, in customer segments that blur until nobody is quite sure who the product is actually for anymore.

The funding goes in. The engineering velocity increases. The product surfaces multiply. And somewhere in that growth, the whole-system view disappears — not because leadership stopped caring about it, but because the org structure makes it structurally impossible to hold.

I have seen this happen from the inside

Each vertical had its own roadmap, its own stakeholders, its own definition of what good looked like. That felt like focus. It felt like accountability. For a while, it worked.

What it didn’t have was anyone asking what the whole added up to.

By the time the question became urgent, the organizational muscle to answer it was gone. The people who might have held the whole-system view had been optimizing verticals for long enough that the wider lens no longer came naturally. The solution was talent: bring in fresh eyes, new leaders, people who hadn’t been shaped by the existing structure.

New people landed. And then, gradually, they got pulled into the same pattern. Not because they were wrong for the role. Because the structure reproduced the problem through whoever occupied it.

We were pruning branches on one side to feed new ones on the other. The tree kept standing. The fruit became fewer.

That’s not a people failure. It’s what happens when an org grows faster than its ability to ask the whole-system question — and has no structural place from which to ask it.

What AI does and doesn’t change

AI accelerates every surface it touches. That’s real and it matters.

What it doesn’t do is resolve the interdependency between surfaces. A faster workflow that’s optimized locally still degrades coherence at the system level. A model that processes referrals in two minutes still leaves patients in a work queue if nobody designed the coordination layer that acts on what the model found.

The companies getting funded aren’t winning because their AI is better. They’re winning because someone, somewhere, did the whole-system read first — understood which stakeholders needed what information, in which order, for the revenue-critical outcome to actually happen — and then built the AI into that architecture.

The clarity came before the capability. The capability just made it faster.

The question worth sitting with

If you’re running a product org at growth stage, the question isn’t whether you are embedding AI in your product fast enough.

It’s whether anyone in your org has the vantage point to know what it should be optimizing for.

Most orgs at this stage don’t. Not because the question isn’t valued. Because the person who could answer it is already accountable to a vertical, a roadmap, a quarterly plan. The whole-system read requires someone who isn’t inside any of those ownership structures — someone who can look across all of them without defending any of them.

That vantage point doesn’t emerge on its own. And it doesn’t get cheaper to establish the longer you wait.

Dima Todorova-Lilavois is a fractional product design leader and the operator of a grain-forward bakery in Long Beach, CA. Field Notes publishes when there’s something worth saying.

Since leaving my role at First American last year, I’ve been running two things simultaneously: a fractional design practice and a grain-forward bakery.

When I mention both in the same breath, people do a quick sort. One is the career, one is the hobby. One is strategic, one is operational. One is digital, one is physical. One is serious, one is charming.

That framing is wrong.

These two businesses are not unrelated. They are the same problem expressed in different contexts. And operating both at the same time — with real stakes, real constraints, and no safety net — has made me better at each of them in ways I didn’t expect.

That’s what this series is about.

Three things I keep learning in both places.

1. Knowing your customer is not a research exercise.

I’ve spent 15 years helping product teams answer the question of who they’re building for. I thought I understood what that meant.

Here’s a story that humbled me.

We built a new feature based on what a competitor was doing and anecdotal feedback from the field. We did our research. We formed hypotheses. We spent six months designing and building something genuinely better — cleaner, faster, more capable. When it launched, users went back to the old version. The clunkier one. The one we were replacing.

It took us a while to understand why. It wasn’t a usability problem. It wasn’t that users didn’t see the improvement. It was that they were operating in high-stakes situations with their clients, and the old version was the one they trusted. Not because it was better. Because it was familiar and had their trust. That’s not something a research study catches. It shows up in behavior, after the fact, when the cost of being wrong is already real.

Six months. Real opportunity cost. Delivered to users who needed to feel safe more than they needed a better interface.

I watched the same thing happen to me at the farmers market, in a single morning.

I sell two similar loaves — same format, different grains, different texture. I can explain the difference clearly: the flour, the fermentation, the crumb. I did explain it, every week. Customers looked confused. Not uninterested, but confused. Why are there two? Which one is right for me?

I was explaining features. They were making a trust decision. They wanted to know which one to bring to dinner, which one wouldn’t be a gamble, which one was the safe choice. The job wasn’t to understand the difference. The job was to not make a mistake.

My sales pitch was answering a question nobody was asking.

The feedback loop is different. One plays out over months and costs a team’s worth of runway. The other plays out in thirty seconds across a folding table. But the failure is identical — we assumed we knew what the customer needed to hear, and missed what they needed to feel.

2. You can’t design your way out of a broken operation

A bakery is a designed system. As a customer experience and as an operation. Even at micro scale, every decision about flow, station layout, product selection, timing, and margin per item compounds. You can’t fix what the customer experiences without fixing the system behind it.

The same is true in product. The experience a user has with a piece of software is not just a function of what the designers drew. It’s downstream of how the team is organized, how priorities get set, how decisions get made, and where attention goes. A beautiful interface sitting on top of a broken operating model is still a broken product.

In both cases, the visible output is just the surface. The real design work is invisible.

3. Focus is not a virtue. It’s a survival mechanism.

Both contexts punish diffusion. Too many menu items, too many features, too many customer segments — the result is the same: diluted effort, unclear value, slower growth.

I say this from lived experience. Every few weeks I want to add something new to the menu. A new loaf. A seasonal item. Something I saw that looked interesting. The instinct feels generative. In practice, it fragments attention, complicates production, and muddies what the bakery actually stands for.

I’ve watched the same pattern play out on product teams under growth pressure. The list of things that “should” get design attention expands faster than the team’s capacity to do any of them well. Everything becomes a priority. Nothing does.

You’ve probably seen a diner menu with 100+ options. You know the feeling. The slight overwhelm, the choice paralysis, the vague suspicion that nothing on it will be great. That’s what happens to a product when focus breaks down. The customer feels it before anyone on the team names it.

So this is Field Notes.

Observations from operating in both. What I learn in one shows up in the other. That’s the lens.

If you’re building a product team under real constraints, or running a small food business and tired of advice that doesn’t quite fit — this is written for you.

Dima Todorova-Lilavois is a fractional product design leader and the operator of a grain-forward bakery in Long Beach, CA. Field Notes publishes when there’s something worth saying.