OpportunityMarket home

How to start a startup: the order the steps actually go in, and the questions everyone asks first

Almost every account of how to start a startup begins at the idea. That is the second step, and starting there is why so much of the first year gets spent building something for a market that was never checked.

On this page

  1. 1The order the steps go in
  2. 2Why step one is not the idea
  3. 3One paying customer before a company
  4. 4Starting alone, and starting at night
  5. 5Where this site fits in the sequence

The order the steps go in

Six steps, and the order is the entire content of this section. Every one of them is cheap to do in this sequence and expensive to do in any other, because each step produces the input the next one needs.

  • Find an opportunity — a demand that exists, is underserved, and someone pays for. Not an idea, a condition of the market.
  • Check it against evidence you did not create: distinct people, a cost they are paying, and a trend that has not already turned.
  • Only then choose a solution. Ideas are cheap once the opportunity is fixed, and you can afford to pick a boring one.
  • Sell it before you build it. A promise someone pays for, or refuses to pay for, is worth more than any amount of encouragement.
  • Build the smallest thing that discharges the promise, for the people who already paid.
  • Incorporate, raise, hire — after there is something to incorporate around. These are consequences of a business, not preconditions for one.

Why step one is not the idea

An idea and an opportunity fail differently, and that difference decides how much a mistake costs. A wrong idea aimed at a real opportunity is recoverable: you change the product and the demand is still sitting there waiting. A brilliant idea aimed at an imagined opportunity cannot be rescued by execution, funding or persistence, because there was never anyone on the other end.

Since ideas are the abundant input and verified demand is the scarce one, spending the scarce effort first is the only sequence that makes sense — and it is the reverse of how the story is usually told.

One paying customer before a company

The registration, the brand, the deck and the co-founder agreement are all work that feels like progress and answers no question. The one thing that answers a question is a person handing over money for a description of something that does not exist yet, because that is the only signal in the whole process that cannot be produced by enthusiasm.

A refusal is worth almost as much, provided you ask why and the answer is about the problem rather than about you. Both outcomes cost a conversation; building for six months to find out costs six months.

Starting alone, and starting at night

Neither is a handicap at this stage. The first two steps are reading and arithmetic, and they do not need a co-founder, a budget or a full-time commitment — they need a few hours a week for a month. The point at which a job becomes a genuine constraint is the point at which customers need answering during business hours, which is later than most people quit.

Choosing a co-founder before there is an opportunity also picks the person for the wrong reason — availability instead of fit to a problem neither of you has identified yet.

Where this site fits in the sequence

Opportunity Market covers steps one and two. It collects public complaints, groups the different phrasings of one problem into a single cluster, counts distinct people, and scores each cluster on demand, pain, momentum, competition, monetization potential and evidence quality — with every score linking back to the posts behind it. It does not generate ideas, and that is deliberate: step three is the cheap step.

What the index is measuring most demand for right now

Measured from public posts, updated continuously — every row links to its evidence

1
Negative Reviews of Subscription Packages
SaaS
50.7
2
Frequent Errors and Slow Access
Banking & Payment Apps
50.5
3
Banking: Failed Update Version
Banking & Payment Apps
50.5
1.0%
4
Zalo messaging app issues and errors
Mobile Apps
50.9
4.4%
5
Logistics: Late Deliveries and No Pickup
Delivery & Logistics Apps
56.1
1.1%
6
Search: Failed Login and Authentication
Mobile Apps
50.1

Frequently asked

Where do I start if I want to build a startup?
With an opportunity, not an idea. Find a demand that already exists and is underserved, check it against public evidence from distinct people, and only then choose what to build.
Do I need an original idea to start a startup?
No. Originality is a property of ideas, and ideas are the abundant input. A boring solution to a demand you have verified beats an original solution to one you have not.
How much money do I need to start?
Nothing for the first two steps — finding and checking an opportunity costs reading time. Costs begin at the point you build, which should be after someone has already paid.
Should I register a company first?
No. Registration answers no question about whether anyone wants the thing. Do it when you have a customer, a contract or a legal obligation that requires it.
Can I start a startup while working full time?
Yes, through the first several steps. Finding and checking an opportunity takes a few hours a week for about a month; a job only becomes a real constraint once customers need answering during business hours.
Do I need a co-founder?
Not to begin. Choosing one before you have identified an opportunity selects for availability rather than fit to the problem, which is the wrong criterion at the wrong time.
How do I know if my startup idea is good?
You cannot evaluate the idea; you evaluate the demand behind it. Count distinct people with the problem, check whether any of them pay for a worse answer today, and look at whether new complaints are still arriving.
What is the most common reason startups fail?
Building for a market that was never checked. It is the failure that execution cannot fix, and it is decided in the first month, long before the product exists.

Keep reading

Every guide

How to find, test and value an opportunity — the whole series.