The single most common reason early-stage startups die is not lack of effort. It is effort pointed in the wrong direction. Founders run hard, but they run at the wrong KPI. They feel productive because they're checking things off a list. The list is the wrong list.
Divya Bhat, a YC group partner, gave a Startup School talk that pulls this apart with the kind of bluntness only someone who has burned themselves badly can deliver. Her own first company, Jam Glue, hit millions of users on a free product, ran out of cash when a Series A fell through, was forced to switch to revenue at the eleventh hour, and only then started learning what their paying users actually wanted. She wishes she had done it on day one. Most founders are about to make the same mistake. Here is the playbook for not making it.
Speed in the wrong direction is just burning fuel
When she says move fast, she does not mean ship more features. She means: the faster you reach revenue, the sooner the company is default-alive instead of default-dead. The faster you get real signal from paying users, the faster every other decision sharpens. Speed is not motion. Speed is shortening the loop between "built a thing" and "a customer paid for it."
Most founders move fast in directions that feel productive but don't shorten that loop. They ship features nobody asked for. They polish onboarding for users who aren't there. They iterate on a landing page when the actual blocker is that nobody wants the product. Speed in the wrong direction is just burning fuel.
Vanity metrics make your mom proud
Vanity metrics are the numbers founders quote when revenue is uncomfortable. Pageviews. Signups. Followers. Press hits. They might make your mom proud. They might make your ex jealous. They do not build a business.
The trap is that they feel like progress. You can show them off. Investors and other founders nod approvingly. Meanwhile your actual revenue line is flat.
Divya's example: an early company of hers spent days agonising over which white-shoe law firm to retain. Long lunches, office tours, the whole performance. They felt important. They had not launched yet. Choosing a good lawyer is fine. Optimising the choice for two weeks is fake work.
Fake progress: the harder lie
Worse than vanity metrics is the category she calls fake progress - work that genuinely feels like work, that is intellectually demanding, that you can defend in a stand-up - but that does not move the needle.
Examples she names directly:
- Premature optimisation for scale you do not need
- Polishing a feature nobody is using yet
- Meeting potential investors when you are not raising
- Attending conferences outside a few specific industries
- Building cool, hard features without user demand
- Migrating ops out of spreadsheets because spreadsheets feel "unprofessional"
Spreadsheets are fine until they're not. If they're working, stick with them. The energy goes into figuring out what your users actually need to use your product daily instead of weekly.
The bottleneck is one specific question
She made the abstract concrete with Super Daily, an Indian grocery subscription startup that later sold to Swiggy. Operationally complex business: app, ops tooling, inventory, ground logistics. Endless things to optimise.
Their north star was growth. Their bottleneck, when they looked honestly, was that high-intent users were getting deep into signup and then dropping. Most founders here would have started A/B testing the signup screen. Super Daily asked the actual question: why are these users not converting?
Answer: a specific milk brand they didn't carry. Users wanted Brand X. Super Daily didn't have Brand X. Onboarding the brand, not redesigning the funnel, lifted conversion 50%.
The framework underneath: write down ideas, don't act. Rank by probability of success, sub-rank by complexity. Pick two. Ship. If the KPI doesn't move, ask why five times until you hit something that hurts to admit.
Five ways your brain protects you from the truth
Divya's list of mental traps - the comforting habits that drift you off the real problem - is worth pinning to a wall:
- Drawn to low-leverage tasks. They give a sense of accomplishment. Tangible. Checkable. Especially seductive when the company's future feels uncertain.
- Fooling yourself that things are working. Slow steady growth is easy to mistake for product market fit. It is not. Real PMF feels fundamentally different from "consistent but underwhelming."
- Perfectionism and indecision. Most decisions don't matter. For the ones that do, decide quickly and fix later. If a call is genuinely tough, it usually means either choice is survivable - so pick.
- Defending downside instead of chasing upside. Fixing little problems is satisfying. Innovation lives on the upside. Get more iterations on the upside, accept more false starts.
- Chipping at small problems while an existential one looms. You have 150 users. They've been 150 for three months. Don't build one-click ordering. Find out why nobody new is signing up.
Charge from day one
The single most expensive mistake she returns to: free products built for years, then forced to monetise too late. Scribd ran free for four years, started charging in year five, lost 90% of users overnight - and revenue grew "by infinity percent." They finally had a business. They also finally had real signal on what paying users wanted, which is fundamentally different signal from what free users want.
If you plan to charge eventually, charge now. Free user feedback is the wrong feedback. The exceptions are real (marketplaces, network effects), but the default is: get paid from day one or don't count those users as growth.
The discipline
Pick one primary KPI. Three to five secondary KPIs to keep yourself honest. Set a weekly target. Audit your calendar against it every week. Be ruthless about killing anything not on the path to that one number.
Move fast. But first, point the car the right way.
Source: Driving Growth Through Metrics and Ruthless Prioritization - Divya Bhat, Y Combinator
No comments:
Post a Comment