When Are You Actually Failing?

You close your laptop on Friday and think: "Are we actually failing, or is this just what early looks like?"

At idea stage, it feels dumb to say you're "failing" when there isn't even a product. At MVP stage, every user says "this is cool," but nobody uses it twice. Is that failure, or "normal"? At go-to-market, you have happy users, but every new customer seems to arrive through some bespoke, painful hustle. Is the product bad, or just the sales motion?

Later, you see startups raising big rounds while losing millions per month. The headlines call them "rockets." Two years later, they're dead.

When did they actually fail?

Founders torture themselves with one vague word—failure—while the underlying risk keeps changing. Investors don't help: one tells you "it's too early to judge," another says "if it's working, it would be obvious by now." So you bounce between "we're onto something" and "we're wasting our lives," often in the same week.

The real problem is not that you're failing. It's that you're using the wrong definition for the stage you're in.

Failure Changes Shape at Every Stage

"Failure" is not one thing in startups.

As you move from idea → MVP → go-to-market → growth → scale, the main way you can die changes: from wrong problem, to weak pull, to no channel, to bad economics, to losing control. If you don't update your failure definition at each milestone, you either kill good companies too early or keep zombie ones alive for years.

The fix is to pick one stage-specific failure metric before each phase and stare at it every week.

Stage One: Idea & MVP — If There's No Sharp Pull, You're Failing

Idea Stage: Failure = Nobody Cares Enough to Change Behavior

At idea stage, your only job is to prove a real, sharp problem exists.

Not "people say it's a problem." "People are already contorting their lives around it." Talk to 20–30 potential users (not your friends). You're failing the idea stage if almost nobody is already hacking together a workaround, almost nobody can put a dollar value or time cost on the pain, and almost nobody says some version of "If you solved this, I'd try/pay immediately."

If 25 conversations in a row feel lukewarm, you don't have a "messaging problem." You have the wrong problem.

Your metric here can be simple: Idea-stage failure line: "Out of the last 10 qualified people I talked to, how many got visibly animated and asked to try it when it exists?" If that number is 0–2 consistently, kill or pivot the idea.

MVP Stage: Failure = People Don't Come Back on Their Own

Once you have something usable, the main risk shifts to engagement.

Compliments do not count. Screenshots in your pitch deck do not count. Behavior is the only vote. Signs you're failing MVP stage: new users try it once and never return, your usage comes only when you nag people manually, and people say "I've been meaning to use it more" (translation: it's not important).

Concrete benchmarks (will vary by product type, but ranges help):

For a weekly-use product (team tool, habit app, etc.), you want ~20–30% of signups to still be active after 4 weeks. Active means log in and do the core action at least once a week. For B2B workflow tools, you want the same users inside the customer account using it multiple times a week, not just "we tried it once in a meeting."

For the Sean Ellis survey ("How would you feel if you could no longer use this product?"), if <40% say "very disappointed," you're probably not there yet.

MVP-stage failure line: "4-week retention below 20% for 2+ months in a row despite serious attempts to improve it."

If you're below that and your only plan is "launch on Product Hunt and hope," you're not early—you're failing this stage.

Stage Two: Go-to-Market & Growth — If You Can't Win Customers Repeatably and Profitably, You're Failing

Once a small group of users loves the product, the risk changes.

Now the question is: Can you get more of these people, on purpose, at a sane cost?

Go-to-Market: Failure = No Repeatable Channel

You're testing channels: cold email, outbound calls, content, paid ads, partnerships, whatever.

You're failing this stage if almost all new customers come from founder's network, PR, or random luck. When you double spend in your best channel, your cost per customer explodes. You have no idea which 1–2 channels are actually working because everything is "a bit messy."

Two simple definitions:

• CAC (customer acquisition cost): how much you spend to get one paying customer (ads + sales time + tools).
• Conversion rate: what % of leads become paying customers.

Rough sanity checks for early B2B SaaS (benchmarks vary, but directionally): landing page to signup should hit at least 3–5%. Signup to paying should hit at least 15–25% if sales-assisted, higher if self-serve. CAC payback (months of gross profit to earn back CAC) should ideally land at <12–18 months.

So if you spend $500 to acquire a customer who pays you $100/month with ~80% gross margin, that's $80/month profit → ~6–7 month payback. That's plausible. If CAC is $2,000 for that same customer, payback is >2 years.

That's probably not fundable.

GTM-stage failure line: "We don't have at least one channel where we can 3–5x spend while keeping CAC roughly stable and payback under ~18 months."

If after 6–12 months of honest testing you still have no such channel, the distribution model—not just the copy—is failing.

Early Growth: Failure = Leaky Bucket + Bad Unit Economics

Once you have a working channel, your risk shifts again: Are we growing a real business or just buying vanity metrics?

Basic concepts:

• Churn: % of customers who leave each month.
• Gross margin: % of revenue left after direct costs (hosting, support, COGS).
• Burn multiple: net burn / net new ARR (for SaaS).

For a healthy SaaS-ish business in this market, rough ranges: monthly logo churn ideally <3–5% for SMB; lower for bigger accounts. Gross margin >70%. Burn multiple <1.5–2 during growth. >3 is a red flag.

You're failing this stage if every new dollar of revenue makes your losses worse, you can't explain when or how the company becomes cash-flow positive, and churn is so high you have to sprint just to stay flat.

Growth-stage failure line: "Churn + CAC mean we never realistically earn back what we spend to acquire customers."

If the only way the model "works" is with heroic future assumptions, that's not growth; it's denial.

Stage Three: Scale — If You're Losing Control, You're Failing Even If the Numbers Look Good

Later, the main risk isn't "does anyone want this?" or "does the math work?" It's "can we run this without it collapsing?"

You can fail at scale while revenue is still rising.

Watch for operational chaos: outages, missed deliveries, angry customers, rising refund rates. Watch for team churn: key people leaving faster than you can replace them. Watch for runway: burning $1M/month with <12 months of cash and no clear path to profitability or a strong next round.

Your dashboard might show record revenue while your culture and operations are burning down behind it.

Scale-stage failure line: "Quality, culture, or cash runway are degrading faster than top-line is improving."

If you ignore that, the market will eventually notice for you.

So What? Define Your Failure Line Before Each Stage

The pattern across all of this:

Early on, failure = wrong problem / no pull. Next, failure = no repeatable way to get and keep customers. Later, failure = the economics or operations don't survive scale.

Before you enter each stage, write down on one line:

"In this phase, we are failing if __."

Make it a single, measurable thing: 4-week retention, CAC payback, churn, burn multiple, runway—pick one that matches your current risk. Review that number every week.

If you're below your line for 4–8 weeks with no clear improvement trend, don't just "give it more time." Change something big: pivot the product, switch channels, cut burn, or, sometimes, shut it down.

Lessons for Founders: Define Failure Before It Defines You

The takeaway isn't that you should obsess over metrics or panic at every dip. It's that clarity saves time—and sanity.

Here's what actually matters:

1. Pick the right failure metric for your stage. Wrong problem, weak pull, no channel, bad economics, or operational chaos—only one matters at a time.
2. Track it weekly, not monthly. If you wait a quarter to check retention or CAC, you've already wasted months of runway.
3. Set a failure line before you cross it. Decide now: "If we're still below X after Y weeks, we pivot or quit." Don't negotiate with yourself in the moment.
4. Change something big if you're stuck. "Give it more time" is only valid if you have a clear hypothesis and fresh tactics. Otherwise, you're just hoping.
5. Study your stage, not someone else's. A growth-stage company's problems are not your problems. Don't import their playbook.

In future pieces we'll go deeper on how to pick the right metric for your exact stage and how to know when to pivot vs quit. For now, stop arguing with yourself about whether you're "failing."

Decide what failure is for this milestone, then go beat that one number.

‍