5 Critical and Company-Altering Learnings from B2B Startups with Y Combinator Managing Director of YC Continuity Anu Hariharan (Pod 595 + Video)

Quick context: Who is Anu Hariharan, and what is YC Continuity?

YC Continuity is the later-stage/growth investing arm associated with Y Combinator, designed to support companies beyond the earliest accelerator
phase. Anu Hariharan’s perspective is shaped by pattern recognition: seeing what founders do before product-market fit (PMF), what they do
after PMF, and what they do when reality shows up and asks for receipts.

The big theme across her five learnings: in B2B, the fastest way to lose is to confuse momentum (fundraising, press, headcount) with value (customers
repeatedly paying, using, and depending on what you built).

Learning #1: Your product may not be as good as you think

This one stings because it’s usually trueespecially early. In B2B, your first “customers” may be friendly, curious, or hopeful. They might pay you
because they like you, because procurement hasn’t noticed yet, or because the status quo is painful and you seem promising.

But B2B greatness is not “promising.” It’s essential. It’s “if this tool disappeared tomorrow, my Monday would become a crime scene.”
And until your product reaches that level of pain relief, you should assume you need iterations.

What founders often get wrong

• Fundraising is not a product review. Investors are betting on potential, not validating usefulness.
• Early logos don’t equal PMF. A couple of well-known customers can hide weak retention or low usage.
• Scaling before quality is expensive self-sabotage. Hiring fast locks you into a product story you may need to change.

How to test if your product is actually “good enough”

A few signals are especially telling in B2B, because they’re harder to fake:

• “Very disappointed” test: Ask active users how they’d feel if they could no longer use your product. If fewer than ~40% say “very
  disappointed,” treat that as a warning lightnot a death sentence, but a “keep iterating” signal.
• Frequency of usage: If usage is rare, value is probably thin (or your product is a “once-a-year compliance ritual,” which is its own
  special business model).
• Demo-to-close conversion: If you’re doing demos and consistently closing well under 20%, either the product isn’t strong enough, the
  target customer is wrong, or the promise isn’t matching the delivered value.

A concrete example

Imagine a B2B startup selling an AI-powered workflow tool for customer support. You get early interest because everyone wants “AI.” But after the demo,
prospects ghost you. Why? Because the tool is neat, but it doesn’t reduce ticket backlog, cut handle time, improve deflection, or integrate cleanly with
the systems teams already use. It’s impressive in a pitch, underwhelming in production.

The fix isn’t “more marketing.” The fix is making the product so useful that customers keep using itand would hate to lose it.

Takeaway: Your job early isn’t to defend the product. It’s to stress-test it until the customer’s behavior proves the value.

Learning #2: B2B startups are underpricing

If you’re a B2B founder, there’s a decent chance you’re charging too little. Underpricing is common because founders fear losing deals, feel awkward
talking about money, or assume low price equals “easy adoption.” In reality, pricing is part of your positioning.

A price that’s too low can signal “toy,” “side project,” or “not enterprise-ready,” even if your product is solid. And it can quietly sabotage your
ability to support customers, invest in reliability, and build the very thing buyers say they want: a vendor that won’t disappear.

Why underpricing is uniquely dangerous in B2B

• B2B buyers often equate price with seriousness. Especially in categories tied to risk (security, finance, compliance).
• Cheap products can be expensive to adopt. If implementation takes effort, buyers want confidence you’ll be around.
• Low pricing blocks great hires. If your margins don’t support strong support and engineering, churn becomes a feature.

What “better pricing” actually means

This isn’t about randomly doubling your price and hoping for the best. It’s about aligning price to valueoften through:

• Value-based pricing: charge in proportion to the outcome you create (time saved, revenue protected, risk reduced).
• Packaging: separate core value from premium features (and avoid dumping every feature into one “please love us” plan).
• Willingness-to-pay testing: run structured price tests and listen for where buyers hesitateand why.

A simple rule of thumb (and how to use it without being reckless)

One heuristic shared in this conversation: if you raise price by ~30% and customers barely react, you were probably undercharging. The point isn’t that
30% is magical. The point is that you should be learning where the real value boundary is.

A concrete example

Let’s say you sell software that automates SOC 2 evidence collection for startups. If your product saves a security lead 10 hours a week and reduces
audit stress (and mistakes), $29/month is not a bargainit’s a credibility problem. A better approach might be pricing by number of systems connected,
number of frameworks (SOC 2, ISO 27001), or support tier.

Takeaway: Pricing isn’t just revenue. It’s a diagnostic tool. If customers are thrilled to pay more because the product is mission-critical,
that’s a signal you’re building something that matters.

Learning #3: You cannot serve all B2B customers

“Our customer is… businesses.” Sure. And my hobby is “breathing air.” B2B markets look uniform from far away, but up close they’re fractal chaos:
different budgets, workflows, regulations, integrations, security requirements, buying committees, and tolerance for risk.

Early on, trying to serve everyone creates a painful trap:
your product becomes a pile of compromises and your messaging becomes a vague fog of “streamline,” “optimize,” and “leveraging synergies.”
(If your homepage says “end-to-end,” you owe your team an apology.)

What to do instead: pick an ICP and commit

The solution is to identify your Ideal Customer Profile (ICP)the segment that gets value fastest, buys with the least friction, and
becomes your most reliable learning engine.

A strong ICP definition typically includes:

• Firmographics: industry, company size, geography, revenue range.
• Operational reality: existing tools, workflows, regulatory constraints, integrations needed.
• Pain intensity: how urgent and expensive the problem is right now.
• Buying path: who says yes, how long it takes, and what triggers a purchase.

“Know who your easy customers are”

The “easy customer” isn’t the one who compliments your deck. It’s the one who:
(1) clearly has the problem, (2) feels it urgently, (3) can buy without a six-month committee process, and (4) will pay for a solution even if it’s not
perfect yet.

A concrete example

Suppose you’re building a B2B analytics tool for sales teams. You might be tempted to sell to SMBs (“faster sales cycle!”) and enterprises (“big ACV!”)
at the same time. But the product requirements diverge immediately:

• SMB wants quick setup and simple dashboards.
• Enterprise wants SSO, audit logs, role-based access control, data residency, and integration with five internal systems you’ve never heard of.

If you attempt both, you’ll build neither well. But if you start with one segmentsay, mid-market SaaS companies with Salesforce and a RevOps functionyou
can build a clear “we solve this” story, win repeatably, and expand later.

Takeaway: Focus is not narrowing your ambition. It’s narrowing your execution so you can actually win.

Learning #4: Early sales should not be delegated

If there’s one place founders love outsourcing too early, it’s sales. It feels efficient: “Let’s hire a VP of Sales to figure out go-to-market.”
Unfortunately, that’s like hiring a chef to guess what you’re allergic to.

In the earliest phase, sales is not just selling. It’s product discovery, positioning, messaging, objection handling, and ICP validation
happening in real time. That learning loop is too important to delegate before it exists.

What “founder-led sales” really means

• You do the calls. Not forever, but long enough to learn patterns.
• You write the playbook. What resonates, what fails, which objections are real.
• You turn chaos into repetition. The moment sales becomes “boring,” you’re ready to scale it.

The “boring and repetitive” milestone

A practical marker discussed here: do enough customer calls (often on the order of 100–200 conversations) until you can predict the
flowwhat prospects ask, what they fear, what makes them buy, and what makes them bail. When your answers start sounding the same (because the market is
consistent), you can hire a sales leader to scale execution.

A concrete example

A founder selling a compliance tool hears the same three objections in every call:

• “Will this integrate with our existing ticketing system?”
• “How long does implementation take?”
• “Can you support our audit timeline in Q2?”

Once the founder can answer these clearly, show proof, and close deals with predictable steps, the company can document the motion and hire to run it.
Hiring before that point often creates a mismatch: the sales leader is asked to invent product-market fit, and the product team gets yanked around by
whatever deals sound loudest.

Takeaway: Early sales is the founder’s fastest route to truth. Delegate too early, and you lose the feedback that builds a real machine.

Learning #5: Plan to reach “default alive”

“Default alive” is one of those phrases that sounds like a zombie movie until you realize it’s actually a financial strategy. The idea (popularized in
the YC ecosystem) is simple:
a default alive startup can survive without needing more outside capital. A default dead startup can’t.

In B2B, this matters because the fundraising environment changes, buyer behavior changes, and the “next round” is never guaranteed. Companies often get
trapped in what Paul Graham described as the fatal pinch: spending has increased, investor standards get higher after you’ve raised,
and the story gets harder if growth isn’t strong enough.

Default alive is not “small dreams.” It’s leverage.

Being default alive buys you options:

• Time to iterate without panic.
• Negotiating power with investors (you’re raising from strength, not necessity).
• Customer trust (you’re stable, not a “maybe they’ll exist next year” vendor).

How B2B startups get to default alive (without becoming a miserly goblin)

Default alive isn’t just “cut costs.” It’s a deliberate mix of:

• Charging enough to support acquisition and customer success.
• Retention and expansion so revenue compounds (B2B is a long game).
• Operating efficiency so growth doesn’t require reckless burn.
• Focus so you’re not building five products for five segments.

A concrete example

A startup selling workflow automation to mid-market finance teams notices that customers who integrate with NetSuite retain far better than customers who
don’t. The default alive move is not “run more ads.” It’s:

• double down on the ICP where integrations and ROI are strongest,
• price around the value created,
• reduce churn by improving onboarding and reliability,
• and keep the team lean enough that revenue covers the cost to acquire and serve customers.

Now the company can keep operating even if fundraising takes longer. That’s not pessimism. That’s control.

Takeaway: Plan for survival as a strategy, not as an emergency. The companies that last are the ones that can fund their own learning.

Putting the five learnings into one practical B2B operating loop

If you want a simple mental model, it’s this:

1. Build something essential (not just interesting).
2. Charge like it matters (because it does).
3. Pick a segment you can win repeatably.
4. Sell it yourself until the motion is obvious.
5. Run the company like survival is a featurenot a phase.

None of these are “hacks.” They’re fundamentals. Which is annoying, because fundamentals don’t come with a dopamine hit. They come with outcomes.