There is a comforting myth in startups: the breakthrough arrives like lightning.

One day the right investor calls back. One customer says yes. One distribution partner opens the floodgates. One product moment clicks, and the company’s story suddenly looks inevitable.

Here is the reframe, informed by science:

Parasites do not become organs. They get rejected, or they kill the system.

Breakout companies are not the ones that got lucky. They are the ones that became so useful, so embedded, and so genuinely mutual inside a customer’s life that biology’s oldest rule took over — two systems started functioning like one.

That is the move. Everything else in this essay is the operating manual.

When a relationship becomes machinery

In 2024, scientists at UC Santa Cruz confirmed something researchers had suspected for years, and gave it a name: the nitroplast.

The paper earned science’s oldest prize. The metaphor it contains is worth more than the headline.

The subject is a marine alga, Braarudosphaera bigelowii, and a bacterium, UCYN-A, that lives inside it. For decades, UCYN-A looked like an unusually intimate partner — a guest, but still a separate organism. Then the evidence crossed a threshold. UCYN-A scales with the host’s cell size the way organelles do. It imports proteins made by the host. Its genome is streamlined in a way that signals long-term dependence, not casual cohabitation.

That was the moment the story changed from relationship to integration.

As Tyler Coale, the paper’s first author, put it:

“It’s very rare that organelles arise from these types of things. The first time we think it happened, it gave rise to all complex life. Everything more complicated than a bacterial cell owes its existence to that event.”

The origin of organelles through primary endosymbiosis has only happened a handful of times in the record of life. It previously gave us mitochondria and chloroplasts. We now have a third confirmed case.

The biology took an eon. Your company has roughly 24 months. The compression is the point.

Over a long enough timeline — or a short enough one — the interface between two organisms becomes tight enough, reliable enough, and useful enough that the system starts treating them as one. Proximity becomes dependence. Dependence becomes coordination. Coordination unlocks a new capability.

That is not lightning. That is construction.

In biology it is slow construction. In startups it is urgent construction. Same rule.

The founder’s job is not to chase miracles

In startups, people talk about luck as if it lives outside the building.

But the most valuable kind of luck is engineered. It shows up when a company is designed to make rare outcomes more likely, then more repeatable.

A simple model:

Rare outcome = Alignment × Time × Interface

• Alignment — are you solving a real, repeated, painful problem?

• Time — do you have enough runway, belief, and stamina for compounding to work?

• Interface — do you have a tight loop with the market that allows reality to shape you fast?

Most founders spend time on alignment. Many obsess over time. Fewer treat interface as a first-class operating system.

They should.

Interface is where the company stops presenting itself to the market and starts absorbing it. Customer feedback becomes product instinct. A use case becomes a workflow. A distribution relationship becomes embedded motion. What used to be external becomes something the company can count on.

Ambition is ancient. The tools are new. The companies that win the next decade are the ones that use the new tools to compress what used to take generations into something that fits inside a seed round.

Why this matters more now

Q1 2026 broke every record in venture history — $300 billion poured into global startups in a single quarter. But the distribution is the story: fewer than 3% of deals captured more than 79% of all capital. Deal volume globally is near its lowest point in a decade. Seed funding is up 31% year-over-year — while the number of seed deals fell 30%.

Mike Volpi at Index Ventures described it plainly:

“You have a handful of companies raising rounds that look more like sovereign debt issuances, and then you have everyone else competing for a shrinking pool of capital. The middle has been hollowed out.”

The data underneath is unambiguous. Only 15.5% of companies that raised seed in early 2023 made it to Series A within two years. The median seed-to-Series A interval has stretched past 616 days.

The winners are not always the teams that raised the most. Often they are the teams that built internal systems strong enough to keep compounding while the market took longer to decide.

In a tighter environment, time is not just something you survive. It is something you convert.

The money follows momentum. Momentum follows integration.

What integration actually requires

One distinction is worth holding onto:

Parasites do not become organs. They get rejected, or they kill the system.

The nitroplast worked because the exchange was genuinely mutual. UCYN-A delivered fixed nitrogen. The alga delivered everything else UCYN-A needed to survive. Integration only lasts when both sides are getting something they cannot replace.

The same rule holds in startups.

You do not earn deep integration by trapping customers, manufacturing friction, or dressing up weak value with better branding. You earn it by becoming meaningfully useful inside a workflow that matters.

That means shifting from:

• pitching → proof loops

• networking → compounding relationships

• shipping features → shipping and absorbing

• growth spurts → growth systems

• hoping customers understand → making customers feel the difference

That last one matters more than most founders admit.

Relief is data. Confusion is data. Obsession is data. Trust is data.

If your customer cannot feel the difference between using your product and not using it, you are still adjacent to their life. You are not integrated into it. You are a hitchhiker, and hitchhikers get expelled.

If It Works…

Picture your company 12 months from now.

Not the vanity version. Not the press hit. Not the logo slide. Not the round announcement.

Picture this: your company has built one internal organ — a capability that runs even when you are tired.

Customer discovery is no longer a phase. It is a rhythm. Distribution is no longer a hope. It is a repeatable motion. Product learning is no longer occasional. It is built into the way the company works.

Then the economics begin to change. Retention improves because the product is anchored in real workflow. CAC falls as referrals and embedded channels start to do more of the work. Insight turns into shipped improvement faster. The same effort produces more value per unit of time.

That is what rare looks like when it becomes repeatable.

Not a spike. A capability.

Belief becomes capital when the capability becomes visible.

Serving the Vision

Five moves. In sequence. Skip one and the architecture stays permeable — you stay a hitchhiker.

1. Build the interface, not just the product

Most founders treat feedback like a suggestion box. Treat it like life support. Build a consistent loop where the same kinds of users show you their reality and your team turns that reality into shipped change.

Action: Run five structured customer sessions a week for six weeks.
Signal: Time from insight to shipped change starts compressing.

2. Earn dependence the right way

Pick one core job your user does repeatedly and make your product unmistakably better at that job. Not more impressive. More indispensable.

Action: Redesign onboarding around one weekly use case and get users to value faster.
Signal: Retention rises and users start saying some version of “I can’t go back.”

3. Write a runway thesis

Time only compounds if it is directed. With the median seed-to-Series A interval now past 600 days, the founders who survive are not the ones who waited — they are the ones who treated every quarter as a compounding cycle.

Action: Set one proof milestone each for product, distribution, and customer value in the next 90 days.
Signal: You can explain your next raise or your path to sustainability in three sentences with numbers.

4. Turn partnerships into shared motion

Most partnerships stay polite. Integration requires cadence, incentives, and operational overlap. The goal is not access. The goal is embedded movement.

Action: Run a 30-day pilot with one partner around one shared customer profile and one shared offer.
Signal: Partner-sourced pipeline becomes consistent instead of occasional.

5. Instrument the emotional arc

Founders measure clicks and ignore feelings until churn forces the lesson. But confusion, relief, confidence, and delight are early indicators of whether your product is becoming part of a real workflow.

Action: After first value, ask two questions: “How did this feel?” and “What almost stopped you?”
Signal: Confusion trends down. Relief and confidence trend up. Support burden drops with them.

Closing Shift

Rare things happen.

But integration has a cost most founders do not see until they are inside it. The alga cannot expel the nitroplast now. The capability is real. The dependency is permanent. You do not get to un-integrate.

The company that becomes an organelle in its customer’s workflow has built something real — and has also accepted that its future is bound to that workflow’s survival. The partner that becomes an embedded channel is also a single point of failure. The capability that runs even when you are tired is also the capability you are now obligated to sustain.

This is not a reason to avoid it. It is a reason to build it deliberately — to earn the integration rather than engineer it, to make the dependency mutual rather than extractive, to choose the workflow you are willing to be inside for the long run.

Because the market does not reward hope. It does not reward the appearance of integration either.

It rewards the real thing.

The company that lasts is not the one that hoped hardest. It is the one that became real inside someone else’s life.

Not next quarter. Not after the raise. Not after the rebrand.

Five customer conversations. One workflow. One week.

If you cannot do that, no amount of capital, runway, or market timing will make you rare.

If you can — the rare thing is already starting.

The surface read of this week is straightforward: record capital, record capex, record valuations. Amazon, Alphabet, and Meta committed nearly $700 billion in 2026 AI infrastructure spending. OpenAI closed $110 billion at a $730 billion valuation. Andreessen Horowitz raised $15 billion across five funds. By every measure, the capital environment for AI has never been larger.

But I don’t think the capital story is the story. I think the cost story is.

The number I keep bringing into founder conversations this week is 4 to 15x. That’s the cost multiplier for agentic AI workflows compared to standard inference, according to new research from Contrary. It means the most exciting product category in AI — autonomous agents that plan, execute, and iterate — is also the most expensive to scale. And in the same week, OpenAI cut token prices by 50%, which sounds like relief until you realize the agentic workflows eating those tokens cost multiples more per completed task.

I don’t think these are separate stories. I think they’re the same story told from two sides of the table: the infrastructure side and the application side. The infrastructure is getting funded. The applications are getting taxed.

The number I keep bringing into founder conversations this week is 4–15x — because it changes what “unit economics” means for anyone building agentic features.

If I were building this week, pull up my inference cost dashboard and run a cost-per-completed-task analysis on every agentic workflow; not cost-per-token, cost-per-task. The compound effect of multi-step reasoning, retries, and tool calls is where margin disappears.

I’d audit whether a smaller, fine-tuned model could handle 80% of my workload. The cost difference between a frontier model and a well-tuned 7B parameter model can be 10–50x! Microsoft’s $38B quarterly capex confirms that even the biggest builders are moving toward smaller models where they can.

I’d separate my agentic pricing tier from my standard tier before my customers discover the cost difference for me. Price the value, not the tokens.

Issue #002 of Signals Served — all 13 signals, Founder Signal for each — is live on The VC Concierge. thevcconcierge.com/insights/signals-served-002

The room is dark except for the stage lights catching smoke in the air.

Four musicians. No sheet music. The bassist nods, counts off with his head, and they’re in.

The first few bars are tight — walking bass, brushed snare, piano comping behind. Then the saxophonist steps forward and starts to solo.

And you can feel the room lean in.

Not because the notes are fast or flashy. Because they land. Every phrase answers the one before it. The rhythm section stays in the pocket; giving the soloist space, but never losing time. When the sax player takes a risk with a bent note, or unexpected rest, it doesn’t break the song. It deepens it.

The solo builds. Tension, then release. A motif appears, disappears, returns transformed. The pianist responds. The drummer accents without overpowering.

Four bars before the end, the sax returns to the head (the main melody) and the whole band comes back in, perfectly together.

This is good jazz.

You don’t need to be a musician to feel it. The conversation is happening in real time, and everyone in the room is inside it.

Bad Jazz

Now imagine the same room, different night.

Same setup. Bass, drums, piano, sax.

But something is off from the first measure.

The drummer is rushing. The pianist is comping like they’re in a different tune. The sax player launches into a solo before anyone establishes the form. Fast runs, big gestures, everything forte… nothing connects.

The bassist tries to anchor it, but no one is listening.

The notes are technically “fine.” The musicians look serious. Everyone is “expressing themselves.”

But nothing lands.

Nothing resolves.

The band isn’t together — they’re just playing at the same time.

This is bad jazz.

And after two minutes, you can feel the room quietly give up.

I hear the same sound in startups all the time.

Founders will call it flexibility. Investors will politely call it exploration.

Inside the company, it feels like motion.

But from the outside, it sounds like bad jazz.

And here’s the uncomfortable truth: most startups don’t fail from rigidity. They fail from thrash.

Thrash sounds like bad jazz. And the room can always tell.

The Myth of Flexibility

Early stage startups worship flexibility. Pivot culture has trained founders to believe that speed plus openness equals intelligence.

“Nothing is sacred.”

“We’re just experimenting.”

“We’ll follow the signal.”

In theory, this is healthy. In practice, it often becomes a cover for something else:

a lack of constraints, cadence, and commitment.

Jazz improvisation is not “play anything.”

It’s conversation with structure.

Every jazz standard has “changes,” the chord progression everyone agrees to follow. The musicians know the form. They’ve internalized the time signature. They’re counting the same way.

That’s what makes improvisation possible.

Startups without form aren’t flexible. They’re just noisy.

What Bad Jazz Sounds Like in Startups

Among serious musicians, “bad jazz” doesn’t mean dissonance or complexity. Jazz can be harmonically wild and still be great.

Bad jazz shows up in patterns, one startups can learn from:

When Startups Start Soloing Over the Changes

Here’s where founders mistake thrash for flexibility.

A deal doesn’t close. | A metric stalls. | A loud customer complains. | A competitor launches something shiny.

So the startup “improvises.”

New ICP. | New feature. | New positioning. | Same confusion.

But here’s what’s actually happening: the team is reacting to noise, not responding to signal.

In jazz, this is the equivalent of soloing over the changes — ignoring the chord progression and playing what feels good in the moment.

It can sound exciting for eight bars.

But by the sixteenth, the band is lost.

And the root cause is almost always the same:

founders confuse motion with momentum.

Motion is activity. Momentum is direction.

Bad jazz has plenty of motion. It has zero swing.

The Nuance I Don’t Want to Miss

The counterargument is obvious: some of the best startups did pivot radically.

Instagram started as Burbn. Slack started inside a game company. Twitter came out of a different product attempt.

So isn’t rigidity the real enemy?

Here’s the nuance: those pivots weren’t random. They came after real customer exposure, and they kept invariants (the parts of the “song” that stayed true even as the arrangement changed.)

In jazz terms: they changed the tune, but they kept the pocket.

Bad jazz is when you lose both the song and the pocket.

Then, the room gives up.

The Good Jazz Pivot Loop

Great jazz musicians don’t improvise randomly. They improvise inside constraints.

The same is true for great startups.

Here’s a simple operating model founders can borrow:

Constraints Create Freedom

Founders often fear constraints because they feel like doors closing.

In reality, constraints are what open creative space.

A clear ICP enables better ideas.

A fixed cadence enables faster learning.

A narrow promise enables stronger pull.

You earn the right to improvise by proving you can keep time.

What to Track So Pivots Aren’t Vibes

If improvisation is earned, it should leave evidence.

Healthy pivoting produces artifacts:

• hypotheses written before experiments

• decisions tied to metrics, not dashboards

• fewer initiatives, finished more often

• a team that can explain why something changed in one paragraph

If the only justification for a pivot is “it felt right,” you’re not adapting — you’re guessing.

And the room can tell.

If It Works…

If this discipline holds, something subtle but powerful happens:

Teams move faster because they change less.

Customers feel coherence instead of confusion.

Execution compounds.

The economic case isn’t mystical. It’s cognitive.

Research summarized by the American Psychological Association points out that task switching can cost a meaningful share of productive time — the mental “gear shifting” is real overhead.

And research on interruptions in knowledge work has found it can take around 20+ minutes to fully resume a task after an interruption.

That’s what thrash really is: repeated unnecessary interruption, at the organizational level.

Strategy changes mid-sprint. Priorities swap mid-quarter. And the reason for the change is never explicit. Context resets mid-conversation.

Even if you don’t attach a number to it, you can feel the burn.

But if you do want a grounded reference point: the U.S. Bureau of Labor Statistics reports a median annual wage for software developers around $133k.

So every hour of focused engineering time is expensive. At a $133k median salary, that’s roughly $64/hour in direct cost. But the real burn isn’t the hourly rate — it’s the compounding cost of delay.

One abandoned initiative per quarter = ~240 wasted founder-hours annually (assuming 20hr/week per pivot). That’s $15k in direct cost, but 6+ weeks of momentum lost — momentum that would have compounded if the team had stayed in the pocket.

And the payoff isn’t only internal. Delivery discipline correlates with outcomes. The DORA/Accelerate research program has repeatedly connected strong software delivery performance (fast lead times, frequent deploys, low failure rates) with broader organizational performance.

When the pocket is tight, you don’t just “feel” momentum.

You can measure it.

Serving the Vision: How Founders Like This Win

1. Install a Pocket

   A pocket is a shared rhythm. It doesn’t mean you never pivot — it means you don’t pivot mid-sprint.

   Action: Commit to a cadence you won’t break for four weeks: one weekly review, one decision log, one clear owner per initiative.

   Signal: Fewer “open threads” week over week — priorities converge instead of multiplying.

1. Name the Changes

   Constraints aren’t cages. They’re the chord chart that keeps the band together.

   Action: Write one sentence for ICP + problem + promise, plus a short “not doing” list for the quarter.

   Signal: Ask three people “what are we building and for whom?” — answers match within ~10 words.

1. Define Signal Before You Hear It

   If you decide what “signal” means after the fact, you’re not listening — you’re rationalizing.

   Action: Pre-define 3 behaviors that count as real pull (e.g., repeats, invites, paid usage) and 3 that don’t (compliments, hypotheticals, one-off edge requests).

   Signal: Weekly reviews cite customer behavior first, opinions second.

1. Protect the Motif

   Most thrash comes from changing too many variables at once. Keep one or two invariants sacred.

   Action: Choose 1–2 invariants for the quarter (usually: who you serve, and the core promise). Put them in every deck and doc.

   Signal: Even when tactics change, the motif remains obvious to customers and the team.

1. Force Resolution

   A solo that never ends isn’t brave. It’s selfish.

   Action: For every experiment, log one of three outcomes in writing: kill / commit / rerun — within two weeks of completion.

   Signal: No zombie projects older than 30 days without a clear decision.

Closing Shift

Improvisation isn’t freedom from structure.

It’s mastery inside it.

Bad jazz sounds like motion without meaning.

Good jazz sounds inevitable.

So does a great startup.

Belief becomes capital* but only when the song holds.

I would’ve wired the money into Sonder.

And I would’ve been wrong.

Here’s the $2.2 billion lesson that became a $6.8 million warning.

Sonder went from a SPAC valuation of $2.2B in 2022 to a market cap of $6.8M in late 2025, 99.7% value destruction. Even a strategic partnership with the world’s largest hotel chain, couldn’t save it.

Last month, guests were finding eviction notices on their Sonder doors. In the same day, thousands of guests scrambled for new accommodations while Sonder’s 1,400 employees were laid off without severance.

Sonder collapsed under a structure that, in hindsight, looks eerily familiar.

This isn’t just one startup’s story.

Story scaled faster than the system.

Diving deep on that below.

The Deal I Probably Would’ve Done

I started my career in hospitality management - from a 91-room select service hotel in Lawrence, KS to the prestige of hospitality with the Ritz Carlton. I saw what Sonder could be in an intimate way.

The name alone is irresistible: sonder—the realization that every passerby has a life as vivid and complex as your own. For a hospitality brand, it’s perfect. Empathy as a product thesis.

On paper, Sonder had everything:

• A massive, global market.

• Clear analogies (Airbnb, boutique hotels).

• A crisp narrative: software + design + operations → better stay, better margins.

• Traction, brand, investor validation.

If you’re a VC, this is the kind of company that fits neatly into an investment committee memo. If you’re a founder, this is the kind of story you’re told you should build.

And still, it broke.

Sonder purportedly raised a $6.5M Seed round in 2015. If given the chance, I would’ve been an early backer. And I would’ve been wrong.

Sonder broke for a structural reason: it was a capital intensive, operations heavy business priced and funded as if it were software.

What a startup’s business is, how it’s treated, and what it becomes is the needle founders are threading everyday.

The mismatch in the in-between, that gap, that’s the Story–System Gap.

The Pattern: When Story Outruns System

Strip away the branding and you see a pattern:

1. A compelling narrative: “We’re reimagining X with tech.”

2. Large checks written on that narrative.

3. Rapid scale: leases, headcount, expansion, marketing.

4. A system (ops, unit economics, financial controls) that never fully catches up.

5. A shock: market turn, demand shift, or just the end of cheap capital.

6. A collapse that, in retrospect, looks obvious.

I’ll walk you through these examples so you can see the pattern clearly and introduce a counterexample of a startup that got it right.

Asset-Heavy “Tech”: Sonder, Stay Alfred, Lyric

Stay Alfred leased and furnished apartments to rent like hotel suites. Great growth, strong revenue, but everything hinged on occupancy covering fixed leases. When COVID hit, the model snapped.

Lyric raised ~$180M, had Airbnb as a strategic investor, and branded itself as the premium, design-forward short-term rental experience. Under the hood? A labor- and lease-heavy operation with the same exposure to cycles and downtown demand.

Sonder took this thesis and ran it at scale: more cities, more units, more leases, more burn.

The common mistake: treating lease obligations and housekeeping like a tech variable instead of a financial anchor.

Tech can optimize hospitality. It cannot turn leases into software.

Narrative vs. Numbers: WeWork’s Gravity Problem

WeWork is the canonical example.

• $47B in lease obligations.

• ~$4B in revenue at IPO filing.

• A culture and story that insisted it was a tech company, despite fundamentally being levered real estate.

Sonder wasn’t as large, but the gravity was the same. Both companies:

• Took long-term obligations (leases).

• Monetized them with short-term, cyclical demand.

• Priced themselves and raised capital as if they were compounding software businesses.

The story got the term sheets. Gravity got the last word.

Scaling Before Fit: Fast and Quibi

Fast burned ~$10M/month on one-click checkout while doing low six figures in revenue.

Quibi raised $1.75B to solve a “mobile video” problem consumers didn’t actually have.

Sonder did a hospitality version of the same thing: multi-city expansion before hard proof that a single city’s economics were robust across cycles.

Capital accelerates truth. It doesn’t change it.

Tech vs. Basics: Zume’s Pizza Paradox

Zume built robotic pizza trucks with AI ovens and predictive routing. It raised ~$445M. The tech demo was compelling; the pizza wasn’t.

The company optimized an incredible amount of complexity before nailing the simplest requirement: is this pizza so good and so reliably delivered that people reorder it?

That’s Sonder in a different costume.

Sonder invested in apps, locks, and yield management. But the guest still judges:

• Was it clean?

• Was it safe?

• Did it work?

• Was it worth the price?

Your tech is invisible if the basics fail.

Toast: The Exception That Proves the Rule

Now look at Toast, a company that could have blown up under this framework.

• Hardware terminals (expensive to deploy).

• Payments (thin margins, intense competition).

• Restaurants (high churn, low margins).

Instead, Toast is a profitable public company, processing well over $100B in annual payment volume and worth tens of billions in the public markets.

Why did Toast survive while Sonder didn’t?

1. They proved one micro-market to death.
Toast spent years in Boston, sweating the details with a few hundred restaurants before scaling nationally. They didn’t scale the story; they scaled proof.

2. Their costs flexed with usage.
Restaurants didn’t require Toast to sign hotel style leases. Toast shipped hardware, but it got paid through ongoing payment processing and usage. When volumes dip, a meaningful share of the cost base flexes down. Sonder, by contrast, locked itself into multi-year leases and staffing commitments that did not shrink when occupancy did.

3. Switching costs compounded.
Once a restaurant runs orders, inventory, payroll, and payments through Toast, ripping it out is operational surgery. That’s a moat. Sonder competed on convenience and design a valuable, but fragile moat. A moat Marriott or Airbnb could realistically replicate within their own supply.

Asset heavy models work when:

• Micro-markets are proven deeply before scale.

• Obligations flex with demand and revenue.

• Switching costs compound over time.

Toast threaded that needle in a way that Sonder’s collapse demonstrated that they could not.

Recap of Sonder’s Collapse

When you rewind Sonder’s last three years, the story looks tragic. But when you consider the operational and financial realities, it starts to look preventable.

The Valuation Mirage

Sonder went public via SPAC in 2022 at a roughly $2.2B valuation. By late 2025, it was filing for Chapter 7 liquidation with a market cap of $6.8M. That value didn’t just disappear, it was never really there.

The Operational Reality

Sonder took on long term commitments like rents, build-outs, housekeeping, local staff, while monetizing with the familiar hospitality business model — short term, cyclical demand.
But some dynamics should have stood out:

The Operational Failures

• A Sonder managed property, the Jung Hotel in New Orleans sued, alleging it had become “a magnet for violent crime” and a “warzone” due to inadequate security.

• Guest complaints included toenails in bedsheets, blood on linens, and properties that weren’t properly cleaned.

• Online reviews: “SCAM—NOT AN OPERATING HOTEL. Just another short term rental place poorly taken over by Sonder. You’ve been warned.”

• A San Francisco hotel sued for $1.2M in unpaid rent under a lease termination agreement.

• Multiple ongoing lawsuits for lease defaults and mismanagement.

• CEO & CFO both stepped down in summer of 2025.

The Financial Warnings

• In early 2024, Sonder disclosed in an SEC filing that its financial records since its public debut “could not be trusted” after being reviewed by an auditor.

• The stock fell 38%. Sonder asked lenders not to call in loans over the accounting errors.

• Stock trading under $1 for most of 2024, facing Nasdaq delisting.

• 17% staff layoffs in 2024.

• November 2024: Sonder filed with the SEC stating there was “substantial doubt about the Company’s ability to continue as a going concern for at least one year.”

When Strategic Partnerships Can’t Save You

By August 2024, Sonder had landed what looked like salvation: a licensing deal with Marriott International.

The partnership brought:

• $146M in liquidity improvement

• $15M in “key money” from Marriott (paid in tranches tied to milestones)

• 9,000+ rooms added to Marriott Bonvoy

• Distribution through the world’s largest hotel chain

The Integration Never Worked

Despite the “tech-enabled” pitch, Sonder struggled with basic integration. The company blamed “technical problems integrating with Marriott’s booking systems.”

Marriott’s booking system may not be cutting edge, but it has worked for decades. If you can’t integrate with your strategic partner’s core system, you can’t deliver the partnership’s entire value proposition.

Instead of the promised revenue boost from Marriott’s distribution, bookings from Marriott Bonvoy dropped sharply.

This does not demonstrate the tech advantage of Sonder’s story, this became a tech problem.

The Death Spiral

By early November 2025, Sonder was drowning in debt, nearly out of cash, and out of options. The company had been negotiating for emergency financing and a potential buyer to take over its assets in bankruptcy, but the bidder abruptly pulled out on November 2.

Here’s how it unraveled:

• 11/05: Marriott agrees to provide Sonder with $1.5 million in funding to cover one week of US payroll, a generous short-term move to help keep thousands of guests housed.

• 11/06: Sonder sends Marriott a proposal: $50 million to cover the costs of winding down Sonder’s operations. Marriott declines.

• Sonder revises the proposal: $28 million. Marriott declines again.

• Sonder comes back with a third proposal: $14 million. Marriott declines for a third time.

• 11/07: Marriott terminates its 20-year license agreement after Sonder warns of “imminent free-fall liquidation.”

• Sonder lays off all 1,400+ employees without severance.

• Signs appear on doors at Sonder properties: guests are told to vacate immediately.

• November 10: Sonder files for Chapter 7 bankruptcy.

What the Industry Saw

Alan Reay, president of Atlas Hospitality Group, captured what many were thinking: “I don’t know how anyone with any iota of business sense could have thought this was a good idea... Whoever did the due diligence, whoever did the underwriting on this, if they’re still at Marriott, I’d be surprised.”

Even with Marriott’s resources, Sonder still needed to:

• Keep properties clean and safe

• Pay its rent

• Integrate its technology

• Maintain accurate financial records

• Generate positive unit economics

It failed at all five.

The Customer Cost

In bankruptcy filings, Marriott alleged that “Sonder collected tens of millions of dollars in advance payments for reservations it now admits it will never honor, spent weeks on a failed restructuring without any contingency plan, and failed to reserve sufficient liquidity to support an orderly wind down.”

The company, Marriott claimed, used guests’ advance payments and deposits to bankroll its own operating expenses.

This couldn’t be further from what I believe the founders’ vision, or the story that investors wanted to back.

The Story–System Gap: A Framework

Here’s the mental model that ties all of this together:

The Story–System Gap = (What the pitch promises) – (What the business can reliably deliver at scale).

When that gap is small, you have integrity: the story matches the system.

When that gap is wide, you have fragility disguised as growth.

Sonder, Stay Alfred, Lyric, WeWork, Fast, and Zume all died in that gap:

• The story promised “tech-enabled” margins and compounding growth.

• The system ran on leases, headcount, fragile demand, and optimistic assumptions.

• The financials quietly told the truth long before the headlines did.

Toast, by contrast, narrowed the gap relentlessly: one city, one use case, one painful iteration at a time.

So how do you know which side you’re on before you scale?

A Founder Self-Diagnostic

If you’re building in an asset-heavy or operations heavy category, ask yourself and your team these seven questions:

1. Can we show one unit or market that works without heroic assumptions?

2. If capital dried up for 24 months, what exactly breaks, and when?

3. What is our biggest fixed-cost exposure, and how fast can it flex down?

4. Which metrics would make us stop expanding, not just start?

5. What’s our Plan B if our best-case pricing power never materializes?

6. Where are we using narrative to justify decisions we don’t yet have data for?

7. If an investor stripped away the adjectives in our deck, would the economics still be compelling?

8. Bonus question (the Marriott test): If our biggest strategic partner walked away tomorrow, are we still a viable business?

If you can’t answer at least five clearly, you’re probably widening your Story–System Gap instead of closing it.

If It Works…

If founders and investors internalize this, we don’t end up with fewer big visions. We end up with better built ones.

• Asset-heavy and ops-heavy companies raise capital on the strength of proven micro-markets, clean financials, and stress tested obligations, not just a great deck.

• VCs reallocate billions away from narrative only bets into structurally sound businesses that show real uplift in per-unit revenue, healthier contribution margins, and far lower failure rates.

• Stakeholders don’t discover the truth in a 72-hour panic; they see it in steady, compounding execution over years.

• The next generation of “tech-enabled hospitality” winners look less like WeWork with nicer sheets and more like Toast applied thoughtfully to physical space.

Instead of avoiding hard categories, we’ll build fewer Sonder style blowups and more Toast style compounding machines.

Serving the Vision: How to Help A Founder Win

1. Prove One Market to Boring

The fastest way to close your Story–System Gap is to make one city, property type, or segment boringly predictable. Run it through a full seasonal and macro cycle, and don’t let yourself be seduced by “early signals” in ten other places.

Action: Freeze expansion beyond your core market until you have 12–18 months of full-cycle unit economics.
Signal: You can show, on one page, how that market behaves in upturns and downturns—and investors stop asking, “But what happens in a recession?”

2. Turn Fixed Costs into Knobs

Every long-term lease, headcount plan, and fixed vendor contract is a bet on future demand. Your job is to negotiate as many of those as possible into variable obligations: rev-share deals, management agreements, flex staffing.

Action: Renegotiate or redesign your top three fixed-cost categories into variable or cancellable structures.
Signal: At least 40% of your cost base can be dialed down within 90 days without destroying the business.

3. Run the 30% Down Scenario Now

Don’t wait for the shock. Ask your finance and ops teams to model demand down 30% for a full year, costs up 20%, and your strategic partner (if you have one) pulling their support. Then decide what breaks—and whether you’re okay with that.

Action: Schedule a quarterly “gravity review” where you only look at downside scenarios and contingency plans.
Signal: Leadership can answer, in real time, “What breaks at −30% demand?” without needing a week to rebuild the model.

4. Write and Honor Your Stop-Scaling Line

Growth is addictive, especially when investors cheer it. Decide in advance what metrics will force you to pause expansion, even if the headlines look great. That’s how you protect the system from your own storytelling.

Action: Codify a simple stop-scaling rule (If contribution margin < X% for Y quarters, we stop new markets) and put it in board minutes.
Signal: You can point to at least one moment where you slowed or halted expansion because the system said “no,” not because the market did.

5. Keep the Books Boring

When the story starts to wobble, it’s tempting to lean on adjusted metrics, creative add-backs, and optimistic forecasts. Don’t. The credibility you gain with clean, conservative accounting is worth more than any short-term narrative bump.

Action: Treat your primary accounting standard as the default language and use non-GAAP metrics sparingly and consistently.
Signal: Investors describe your reporting as “boring in the best way,” and your audit process surfaces friction early rather than in crisis.

Closing Shift

I still love the word sonder.

It captures a deep truth: every “user” is a person with invisible constraints, trade-offs, and pressures.

The same is true of businesses. Behind every pitch deck is a system of contracts, people, and physics that doesn’t care about narrative.

Sonder’s collapse will be remembered as a headline. It should also be remembered as a syllabus.

The danger comes when the story gets so far ahead of the system that nobody notices the physics until the floor gives out.

If you’re building in an asset-heavy or ops-heavy category, you’re certainly not playing an easier game. The numbers will always tell on you, eventually. The lesson is not “don’t build ambitious, asset-heavy companies.”

It’s this:

Big visions deserve systems strong enough to carry them into the future.

Data & Context Sources

Sonder went public via SPAC in 2022 at an implied valuation of around $2.2B before losing more than 99% of that value by its bankruptcy filing. Source: Public SPAC merger filings and market data, 2022–2025. SEC EDGAR: Sonder SPAC Registration

WeWork’s 2019 IPO filing disclosed roughly $47B in minimum future lease obligations against under $4B in revenue and multi-billion dollar annual losses. Source: WeWork Form S-1 Registration Statement, 2019. SEC EDGAR: WeWork S-1

Toast processes over $100B in annual payment volume and has traded at a public market valuation in the tens of billions of dollars. Source: Toast quarterly earnings and investor presentations, 2023–2024. Toast Investor Relations

Sonder disclosed unreliable financial statements for two years, leading to a 38% stock drop. Marriott signed a partnership agreement six months later providing $146M in liquidity and $15M in key money. Source: Sonder SEC filings, 2024; Business Insider bankruptcy coverage, November 2025.

November 2025: Sonder attempted to secure $50M, then $28M, then $14M from Marriott to fund wind-down operations. Marriott declined all proposals and terminated the 20-year licensing agreement on November 7. Sonder laid off all 1,400+ employees the same day and filed Chapter 7 bankruptcy on November 10. Source: Marriott bankruptcy filings; Business Insider, “Marriott’s Bet on Sonder Backfired,” November 2025.

Major hotel groups have shifted toward asset-light models, using management and franchise contracts rather than owning real estate directly. Source: McKinsey & Company, “The Asset-Light Revolution in Hospitality,” 2019. McKinsey Insights

Rising interest rates in the 2020s have disproportionately impacted capital-intensive and levered business models reliant on cheap financing. Source: International Monetary Fund, “The Return of High Interest Rates,” 2023. IMF Blog

I was reminded of this during a panel at Venture135:

The moderator asked: “What tactical advice do you have for founders who believe they’re seeing early product-market fit and want to scale distribution?”

I couldn’t help myself: “Before we talk tactics… how are you defining early product-market fit?”

Because in my view, no founder is building toward a final perfect image that ends up being true, at no part during the journey.

New pieces appear through customer conversations. Old pieces stop fitting as constraints emerge. Edges you thought were fixed dissolve as the market shifts.

And yet, the job is to stay committed anyway. This isn’t a flaw of the founder journey.

It IS the founder journey. The ability to stay loyal to something that refuses to sit still is the prize.

What No One Warns You About the Puzzle

Most people think startups are hard because execution is hard.

That’s only the surface.

The deeper difficulty is the emotional strain of holding a vision that evolves faster than your capacity to articulate it.

Each week brings: new data | new contradictions | new truths | new missing pieces

CB Insights reports that 35% of startup failures happen because founders build something the market didn’t want, not because of poor execution, bad timing, or weak talent.

They wasted their time assembling a picture that didn’t match reality.

The evolving puzzle isn’t noise. It’s the strongest signal you have.

The Vision Almost Always Changes

Your vision is a hypothesis, not a blueprint. And hypotheses are supposed to break.

Founders often confuse consistency with conviction.

But conviction isn’t staying loyal to the first picture in your head.

Conviction is the willingness to rebuild the picture as the world reveals what’s true.

Great founders are indeed obsessed with the picture, but the exceptional ones are obsessed with the process of assembling it.

Their identity isn’t tied to the first idea. It’s tied to the arc of discovery.

The Emotional Math of Commitment

You must continue building something you can’t fully see.

That’s the emotional tax of early stage.

Each evolution requires a small grief cycle: letting go of a feature, a plan, a theory, a dream. And while you’re grieving the old picture, you’re simultaneously persuading your team, investors, and customers to believe in the new one.

You’re pitching a moving target. You’re building a moving target. You ARE a moving target.

And yet, great founders can’t quit, because somewhere inside the ambiguity, they see momentum others can’t yet name.

The Canva Case Study: When Strategic Patience Reveals the Bigger Picture

Before Canva became a $40B+ global design platform, it was barely even a startup.

It was a drag-and-drop yearbook builder called Fusion Books, bootstrapped in Western Australia.

A tiny puzzle.

A tiny market.

A tiny first step toward a massive vision.

Strategic Piece Placing Changed Everything

Melanie Perkins always had the big vision: democratize design for everyone.

But in 2007, she and co-founder Cliff Obrecht faced a familiar founder reality:
No technical background | No Silicon Valley connections | No funding

So they made the smartest possible move: Start with what’s achievable.

Fusion Books became profitable from day one, (and more importantly) it validated the core hypothesis of Canva: people needed simplicity, collaboration, and templates, not complexity.

Within five years, Fusion Books was the largest yearbook provider in Australia, expanding to New Zealand and France.

The Micro Moments of the Puzzle Coming Together

Then something interesting happened: A school emailed asking if they could use the editor to create flyers for an event.

Then another asked for posters.

Another for social graphics.

A teacher requested access for her small business.

None of these pieces fit the “yearbook” puzzle.

But they perfectly fit a much bigger one.

What Made the Difference

The insight wasn’t that yearbooks were the business.

The insight was that non-designers desperately needed simple, collaborative, cloud based visual creation tools, and yearbooks were just the wedge that proved it.

When Bill Tai agreed to invest (after 100+ rejections), it wasn’t because of Fusion Books alone. It was because Perkins showed both vision and validated execution:

• Profitable demand

• A working product

• A clear emergent pattern

• A founder willing to assemble the puzzle piece by piece

In 2013, Canva launched.

Today: over 170M monthly users and billions in annual revenue.

This is what great founders do: They don’t predict the future.

They assemble the puzzle in a way create their vision of it; strategically, patiently, relentlessly.

A Simple Framework: Three Types of Puzzle Pieces

To make sense of the chaos, founders should look for three types of pieces:

1. Guiding Pieces — the edges; non-negotiable beliefs that shape direction

2. Emergent Pieces — customer driven truths that appear only after many conversations

3. False Pieces — assumptions that look right, feel right, but never actually fit

Your job isn’t to protect the original image.

It’s to identify which pieces reveal the real vision, and move faster than your assumptions.

When you do, something remarkable happens:
the picture begins to reveal itself with momentum.

If It Works…

The original image rarely wins.

But committing to the process of discovering the truest image takes you far.

What emerges is product inevitability:

• GTM becomes simpler: clearer ICP, sharper messaging, reduced CAC

• BD partners see obvious gaps: incentives for accelerating distribution by 30–60%

• Teams align around validated learnings: fewer wasted cycles, faster iteration

• Investors feel momentum, not promises: raising from strength, with probably 20–40% stronger valuation leverage.

When the puzzle evolves in sync with customer truth, burn decreases, accuracy increases, and revenue compounds.

If it works, the product becomes something no static vision could have predicted, because you have to be willing to assemble what is actually there.

Serving the Vision: How to Help Founders Win

1. Ship in Tight Loops, Validate in Real Environments

Reality should shape the roadmap faster than intuition.
Action: 1–2 week build → release → feedback loops.
Signal: ≥50% of shipped features generate learning within 14 days.

2. Build High-Leverage Surfaces, Not More Features

Scale comes from primitives, templates, and workflows.
Action: Identify the top 1–2 product surfaces customers reuse most.
Signal: Usage volume grows without adding new features.

3. Turn Customer Conversations into GTM Maps

Discovery is revenue research, not UX research.
Action: Track urgency, willingness to pay, and influence per call.
Signal: 2–3 repeat patterns shaping GTM each month.

4. Build Partnerships That Reduce GTM Friction

Distribution accelerates when partners carry momentum with you.
Action: Launch 2–3 pilot integrations or channel partnerships.
Signal: Partners generate ≥15% of qualified leads within 60–90 days.

Closing Shift

You don’t assemble the puzzle to prove the picture. You assemble the puzzle to discover it.

Every new piece is a signal. Every shift is a revelation.

Every evolution brings you closer.

The founder’s puzzle doesn’t get easier, it gets truer.

And piece by piece, belief continues to become capital.

If you are a founder trying to figure out how the pieces fit together, send me a note - JD

A lot of my conversations with founders revolve around fundraising. The conversation is almost never about the metrics they think matter. It’s about something quieter, harder to measure and something I heard greatly articulated in Turner Novak’s recent conversation with Adit Abraham, the YC-backed founder of Reducto.

Most founders obsess over when to raise, not why they should. Some of the most quietly powerful startups flip that script — they raise when their customers need them to.

That’s the insight from this conversation, how this team realized their most important metric wasn’t revenue or CAC. It was in the words of Abraham:

Their startup took off not because of a perfect pitch deck, but because the right people cared deeply. The customers weren’t just buying; they were building their own success around the product. And that energy, the authentic pull, became the gravitational force that drew investors in.

The Misunderstanding of Funding Momentum

Founders often think of venture capital as ignition fuel.

In reality, it’s often just an accelerant.

Investors would much rather fund kinetic energy rather than potential energy.

If your customers are already in motion — already integrating your product into their workflows or lives — VCs see that as proof of something rare: inevitability.

That’s why YC’s mantra, “Make something people want,” is often misunderstood.

It’s less about market validation, maybe its emotional validation.

The earliest signals aren’t standard metrics; it’s something like conviction per customer.

And conviction shows up in behaviors:

• They reference your product in their own decks.

• They bring you into their planning meetings.

• They offer to co-market before you even ask.

• They ask how they can help you scale.

When that happens, you’re not chasing traction, you’re proving it.

When Customers Become Your Growth Engine

Maybe more founders should think about “fit” not as product–market, but as customer mission alignment.

These founders build a relationship loop: they don’t just sell to customers; they recruit them as co-designers.

In a market where investor capital is selective, this kind of advocacy is the ultimate moat.

Early customers who feel emotionally invested can behave a lot like early employees: recruiting additional customers, evangelizing, defending, and supporting.

Research consistently shows that companies built on customer referrals outperform peers: Referred leads convert 3–5× better and retain at meaningfully higher rates over time DemandSage, 2024.

In early-stage venture, investors frequently discover founders through customer introductions or evangelist referrals — a growing trend as warm networks replace cold outreach in sourcing deals First Round Review, 2023.

That’s not coincidence.

That’s customer conviction turning into capital velocity.

The Emotional Signal Investors Actually Notice

What moves so many investors today isn’t a spreadsheet alone, it’s momentum you didn’t buy.

They look for unprompted behaviors: customers showing up in testimonials, tweets, Slack screenshots.

The new diligence isn’t just ARR; it’s the story customers tell about you without being asked.

When an investor hears:

“Our Fortune 10 client built their internal analytics around us before we were even GA,”

—they know you’ve hit escape velocity.

That’s the subtle truth from this conversation, and a key insight I shared with founders in my network:

VC funding is more often an after effect of customer energy, not a substitute for it.

Building for Pull, Not Push

This is the founder discipline that separates traction from noise.

Instead of asking, How do we get investors excited?, great teams ask:

“How strongly do our customers feel about what we’re building?”

Does that question make you uncomfortable? It should.

Because it forces honesty. Are you building something people truly need to build upon, or something you’re pushing into existence through pure will?

When you can validate the former — even in a handful of customers — you have something as valuable as a term sheet: signal.

If It Works…

If more founders centered fundraising strategy on customer conviction, the early-stage ecosystem would shift overnight.

Startups would grow from the inside out, through products that amplify their customers’ growth.

VCs would see higher conversion rates and lower burn multiples because capital would chase verified demand, not manufactured hype.

The result?

A healthier cycle: more sustainable valuations, better retention, and founders who raise from the comfort that they can, not because they must.

Even a modest increase in “customer pull,” e.g. 20% of users actively advocating, can lift early MRR growth by 2–3× and cut CAC in half.

The math is simple: emotion can compound faster than marketing spend.

Serving the Vision: How Founders Win

1. Quantify Emotional Demand

   Founders should measure more than usage. Track unsolicited referrals, co-marketing offers, or customers introducing you to peers. Those behaviors are your leading indicators of obsession.

   1. Action: Create a simple “customer-pull score” (% of customers proactively promoting).

   2. Signal: Score rises month-over-month before revenue does.

2. Build “With,” Not “For”

   Invite customers into the roadmap. Use shared Slack channels or design councils to co-create features. The more invested they are, the more they’ll root for your success.

   1. Action: Establish a 5–10 person customer council by next quarter.

   2. Signal: 50%+ attendance and at least one roadmap item shaped by feedback.

3. Build a Culture of Listening

   Every customer conversation should inform your build priorities. Don’t delegate, founders should hear the emotion firsthand.

   1. Action: Founders personally conduct 10 customer interviews per quarter.

   2. Signal: Product roadmap decisions tied directly to recurring feedback patterns.

4. Translate Love into Leverage

   Turn customer conviction into investor confidence by documenting real world use and enthusiasm.

   1. Action: Create a “customer-love” slide with screenshots, quotes, and proof of organic adoption.

   2. Signal: Investor inquiries increase post pitch.

Closing Shift

At its core, venture is about energy transfer, belief becoming capital.

But belief doesn’t start in boardrooms.

It starts in the trenches, with a few customers who can’t wait to build their futures on top of yours.

If you can capture that emotion, the money will find you.

Data & Context Sources

• Turner Novak – From Pivot to Fortune 10 Customer

• DemandSage (2024) – Referral Marketing Statistics

• Nector.io (2024) – Referral Marketing Strategies

• First Round Review (2023) – Drive Growth by Picking the Right Lane

• NFX (2023) – Why Market Pull Beats Market Size

Thanks for reading ⚡️The VC Concierge🗝️! Subscribe for free to receive new posts and support my work.

Subscribe now