Mar 2026 | Tangents Tangents is engawa’s shorter-form series: one object, one idea, examined closely.
I Still Remember Their Names
In 2021, my co-founder and I started a venture around one of the most taken-for-granted things on earth: water.
Not water scarcity in the abstract. Not a policy argument. Water in the specific, and specifically, the question of whether ordinary homeowners could actually be helped to use it more consciously, without revamping the plumbing systems they’d never thought about. The water flowing through your pipes, filling your glass, running your dishwasher, absorbed into your awareness so completely that you stop seeing it. That invisibility was both the problem we were trying to solve and the first obstacle we faced in almost every conversation.
I’ve spent the last six and a half years on the East Coast, and four of them building a venture. This is a piece about what those four years taught me, specifically the first 90 days of incubation, back when the tools were slower and the only way to understand people was to actually talk to them.
We had 90 days. We had a hypothesis, a Miro board, and no customers.
So we started talking to people.
One Hundred Strangers
This was 2021, full remote, full COVID. No coffee shop interviews, no chance encounters, no reading a room. Everything happened on screens. We cold-outreached relentlessly — homeowners, builders, restaurant operators, landscape architects, plumbers, municipalities, anyone who touched water in their daily work or life. I wrote more cold messages in those three months than in the previous three years combined.
What came back wasn’t data. It was people.
Dean, a small business owner, said it in four words: “Water is something you take for granted.” He didn’t say it as an insight. He said it as a fact, the way you’d say the sky is blue. That flatness was the data. We weren’t dealing with a knowledge gap. We were dealing with something closer to a structural blind spot, a resource so reliably present that the mind had simply stopped registering it.
Alice, a homeowner we spoke to early on, put it differently: “Frankly, we haven’t done anything because we don’t know what to do.” And then: “We have no idea how much water we are consuming.” Not embarrassed, exactly. More like someone who’d assumed a problem was being handled by someone else, and was only now being asked to look directly at it.
Vincent, who ran a small business in New York, said something I’ve thought about more than almost anything else from those months: “It’s not like electricity, where I could just turn it off and then I’ll save a few dollars. Water is just…it’s necessary.” That sentence explained an entire category of behavior. Electricity gives you feedback: the bill goes down, you feel it. Water doesn’t work that way. It’s there until it isn’t. The invisibility isn’t laziness. It’s a design problem that predates us by a century.
And then there was Brad, a sustainability-focused landscape architect in Texas, someone who had spent years trying to push his clients past the minimum. He said: “Worst what you can do is only to meet the codes. We can do better.” He believed it completely. And he was still losing the argument, project after project, because cheap water prices made the math not work and because the people who needed convincing were never in the room.
We started with open-ended conversations. No solutions, no product pitches, just listening. Once we moved into solution ideation, we tested thirty-plus concepts through hand-drawn storyboards, literally sketched on paper, because we’d learned that polished presentations invited polite responses and we needed honest ones. Over 100 conversations in the incubation phase alone. By the time we got to our Kickstarter launch, we had lost count.
What emerged from all of it was a water monitoring device, designed to be installed as simply as possible, because we had learned that complexity was the first thing that killed adoption. We spent more time on onboarding than almost anything else, obsessing over how to remove the plumbing barrier that had consistently stood between homeowners and water awareness. Making the invisible visible, without requiring a plumber to do it.
I’m not telling you this to be impressive. I’m telling you this because every one of those conversations left something behind.
Between the Answers
Here’s what I didn’t expect: how much of the real information arrived sideways.
Not in the answers people gave, but in the hesitations before them.
Our principle going in was simple: ask high-level questions, then dig. Why. Why. Why. Let people talk. Don’t pitch.
What we were really doing, though I wouldn’t have said it this way then, was accumulating context. Not data points — we quantified what we could, but the numbers were never the point. Context was. The difference between knowing that homeowners underestimate their water usage and understanding why, which turns out to be less about ignorance and more about a kind of learned helplessness. The system is opaque. The feedback is nonexistent. And the assumption, held almost universally, is that if something were really wrong, “someone” would have told them.
That insight didn’t come from a single interview. It sedimented. Conversation by conversation, until one day we looked at the Miro board and something had moved. We weren’t just mapping pain points anymore. We were starting to understand the relationship people have with an invisible resource. The particular texture of trust and neglect that defines how most people relate to water.
You cannot get that from a dataset. A dataset would have told us that water waste is a significant problem and consumers express interest in solutions. It would have been correct and useless.
The Weight of the Words
When we eventually presented to the board, not to raise money, but to make the case that this venture was worth continuing, I had a deck. Frameworks, segmentation, validation data. All the professional apparatus of a startup making its case.
But when I spoke, I wasn’t reading from the slides. I was thinking about Dean’s four words. Alice’s quiet admission. Brad saying we can do better like he’d been waiting years for someone to ask.
I’ve given a lot of presentations. I don’t think I’ve ever put that much weight on individual words. Not because I was performing — because I was accountable to specific people, not to a user persona. These were real people who had given us their time and their honest confusion and their quiet embarrassment about things they’d never been asked to think about. The direction we were recommending had come from their lives. Getting it wrong would have been a failure of a different kind, not to a business case, but to the people who had trusted us with something they didn’t know was valuable: their honest confusion.
What AI Would Have Given Us
I’ve thought about this a lot since, especially as the tools have gotten better.
If we’d run the same process today, we could have fed our transcripts into a language model and gotten a synthesis in seconds. Pattern recognition, theme clustering, and suggested positioning. All of it faster and cleaner than our scattered, chaotic Miro board covered in digital sticky notes at 1am.
And maybe, if we’d been trying to move faster, we’d have been tempted to skip the conversations entirely. Scrape existing research. Query a model trained on consumer behavior. Get to the answer without the friction of finding the people, scheduling the calls, sitting with the silences, writing down verbatim lines that felt important even when we didn’t yet know why.
Here’s what that version wouldn’t have contained: the moment someone went quiet on a Zoom call and didn’t fill the silence. The excitement of a person leaning forward at a hand-drawn storyboard. The specific texture of how people talk about something they’ve never been asked to think about before. The fact that my co-founders and I, after a long day of interviews, would sometimes just sit on the call together not talking, processing before one of us said, “Did you notice how he described that?”
The synthesis would have been statistically similar. The weight would have been absent.
I’m not making a case against the tools. I use them. This is not nostalgia for inefficiency, and not a fantasy about manual grind. But there’s a category of knowledge that you earn by being present for the confusion of other people that doesn’t survive the compression. What you get when you skip the conversations is the answer without the accountability. Correct, possibly. Yours? No.
What Stays
The East Coast chapter is closing.
Six and a half years. Four of them spent building something from zero — a Miro board, a hypothesis, and eventually a product that made it to market. I find myself thinking about what stays.
I still remember the names of the first people who tested the product with us. I can’t recall every word of every interview, but I remember enough: the tone of a voice on a Zoom call, the dark basement where I observed installation, the specific frustration someone expressed about their utility bill, the person who said “nobody’s ever asked me about this before” as if they’d been waiting.
They were the product, before the product existed. Their hesitations and their taken-for-granted assumptions and their quiet embarrassment about an invisible resource. All of it is in there somewhere, in the decisions we made, the language we chose, the problems we decided were worth solving.
An algorithm could have told us where the market was. It couldn’t have told us who we were building for. It couldn’t have put Dean’s four words in my head, or Alice’s quiet admission, or the moment Brad said we can do better like he’d been waiting years for someone to ask.
That’s what I’m carrying. Not the deck, not the data. The people. Always.
Droplet: https://www.kickstarter.com/projects/hydrific/droplet-smart-home-water-sensor
Taishi Okano writes about the intersection of technology, craft, and culture from New York and Tokyo. engawa is where he works things out. This piece is part of Tangents — shorter takes on specific objects, brands, and cultural moments.