Validate before you build

Who this is for

You have an idea and the skills, or the AI tools, to build it quickly. The open question is no longer whether you can build it but whether anyone wants it. Guests who spent months or years on a first product call that their biggest mistake: Dmytro Krasun spent five months perfecting a first launch (Dmytro Krasun · Dmytro Krasun (Nov 2024) ▶ 04:25), and Cole spent about half a million dollars on two failed projects before a game built in two weeks worked (self-reported) (Cole · Cole & Ian (Apr 2025) ▶ 10:51).

(Cole · Cole & Ian (Apr 2025) ▶ 12:17)

The route at a glance

1. Check that competitors already make money, then read keyword and trend data.
2. Talk to target users about what they do now, not what they think of your idea.
3. Write down a threshold and a deadline before you look at results.
4. Put up a landing page or fake demo before writing product code.
5. Post about the idea first; build what the audience asks for.
6. Take payment: pre-sell, or charge from day one.
7. Ship a one-feature MVP in about two weeks and iterate.
8. Do things that don’t scale for the first customers.
9. Read fit from paying, retained strangers before you scale spending.

Steps 4–6 are alternative tests rather than a strict sequence; most guests used two or three of them.

Step 1 — Look where money is already being made

Build in markets that already generate revenue rather than inventing a category (Deven Bhooshan · Deven Bhooshan (Sep 2025) ▶ 02:53). Samuel Rondot confirms that an idea already makes money and checks where its traffic comes from: growth driven by ads can be copied quickly, growth from SEO alone cannot (Samuel Rondot · Samuel Rondot (Jul 2025) ▶ 04:17). Then check search demand: Nicolai Klemke looks for about 1,000 monthly searches with low difficulty (Nicolai Klemke · Nicolai Klemke (Jul 2025) ▶ 05:39).

Thresholds are personal, not benchmarks. Max will enter a market if the top competitors earn 100–200 EUR a month (Max · Max (Dec 2025) ▶ 04:04); Lotts Ezeike looks for 100K–500K MRR (currency unstated) (Lotts Ezeike · Lotts Ezeike (Nov 2025) ▶ 07:41).

Step 2 — Talk to users about behaviour, not opinions

Talking to users is the step builders avoid (Tibo · Tibo (Dec 2025) ▶ 04:47). Never ask “would you use this?”. Ask how people solve the problem today and what it costs them (Gaurav · Gaurav (Aug 2026) ▶ 10:50). Umberto Mezzadra interviews 5–10 target users without saying why he is asking (Umberto Mezzadra · Umberto Mezzadra (Mar 2026) ▶ 12:04); Gaurav’s team logged about 2,000 calls, trading seven days of extra access for each booking (self-reported) (Gaurav · Gaurav (Aug 2026) ▶ 07:29) (Gaurav · Gaurav (Aug 2026) ▶ 07:09).

Common mistake: counting friends and family, who will encourage you and won’t buy (Angus Cheng · Angus Cheng (Jul 2025) ▶ 03:16). The presenters’ view, from Pat Walls, is that when anyone can build quickly, talking to people is the edge (presenter · Brian Shin (May 2026) ▶ 15:26).

Step 3 — Set the bar before you look

Brian Shin’s team set a dated target (10 hosts committing to use the app at real events) and did not build until they hit it (Brian Shin · Brian Shin (May 2026) ▶ 10:45). John Rush wants 100 sign-ups, then five discounted pre-sales (John Rush · John Rush (May 2025) ▶ 04:04). Gil Hildebrand works backwards from about $20,000 of runway to 50 buyers and the list size he needs (Gil Hildebrand · Gil Hildebrand (Dec 2025) ▶ 10:32).

Common mistake: emotional attachment.

(Loïc Berthelot · Loïc Berthelot (Nov 2025) ▶ 12:47)

Step 4 — Landing page or fake demo before code

A waitlist page now takes about an hour (Steph France · Steph France (Jun 2025) ▶ 01:35). CJ Zafir reports 1,800 sign-ups in two weeks from a demo tweet and a page (self-reported) (CJ Zafir · CJ Zafir (May 2025) ▶ 06:28). The bolder version is a fake demo: Timo Nikolai ranked a page with a fake AI-voice demo and a book-a-call link (Timo Nikolai · Josef Buettgen & Timo Nikolai (Sep 2025) ▶ 03:04), and Alejandro made an app demo from existing footage that drew 80,000 views (Alejandro · Alejandro & Mario (Dec 2025) ▶ 00:13).

Common mistake: treating sign-ups as proof (see the waitlist warning below). And a page with visitors but no sales may have the wrong framing rather than the wrong idea (Jack Friks · Jack Friks (Aug 2026) ▶ 06:27).

Step 5 — Post the idea first

For consumer apps, many guests post content about the problem and build only what the audience responds to (Content-validated app idea (post the idea before building it)). Louis says his ten earlier apps, built before any marketing, got no users (Louis · Louis (Dec 2025) ▶ 00:00). Jack Friks posts 60-second talking-head videos and builds if commenters say they need the app (Jack Friks · Jack Friks (Mar 2025) ▶ 05:20). The signal is comments asking “what’s the app?”, not “lol” (An Nayal · An Nayal (Jun 2026) ▶ 07:06). Test in cheap formats first and step up only the winners (Content testing ladder (tweet → thread → video)) (Alex Lieberman · Alex Lieberman (Nov 2023) ▶ 17:22).

Common mistake: a weak test video. If the content is poor, you cannot tell whether the idea failed or the video did (Alejandro · Alejandro & Mario (Dec 2025) ▶ 04:44).

Step 6 — Take payment

This is the most widely shared belief in the cluster: only payment validates (Pre-sell before you build).

(Timo Nikolai · Josef Buettgen & Timo Nikolai (Sep 2025) ▶ 06:15)

Hildebrand pre-sold about $20,000 before building (self-reported) (Gil Hildebrand · Gil Hildebrand (Dec 2025) ▶ 05:34). Colin McIntosh offered sign-ups a $69 launch price instead of $100 (Colin McIntosh · Colin McIntosh (May 2024) ▶ 06:00). Otherwise, charge from day one and put a buy button on the scrappiest version (Ayush · Ayush (Jan 2026) ▶ 19:05). Tiny first payments count: Ken got $10 a few days into an alpha (Ken · Ken (Sep 2026) ▶ 06:28).

Common mistake: trusting free usage. Leandro’s free products lost their users as soon as he charged (Leandro · Leandro (Jun 2025) ▶ 14:15).

Step 7 — Ship a one-feature MVP in about two weeks

Two weeks is the build time founders quote most often (Samuel Rondot · Samuel Rondot (Jul 2025) ▶ 11:23). Cut to one feature and skip password resets, settings pages and logos (Samuel Rondot · Samuel Rondot (Jul 2025) ▶ 12:10); launch within a month (Dmytro Krasun · Dmytro Krasun (Nov 2024) ▶ 04:41). Treat the MVP as disposable: Shin’s team threw its validation prototype away and rebuilt (Brian Shin · Brian Shin (May 2026) ▶ 05:53) (Lean MVP loop: ship small, get feedback, iterate or move on).

Common mistake: perfectionism, as in Krasun’s five-month first launch. The presenters push speed hard (presenter · Diego Roshardt (Jul 2025) ▶ 03:02).

Step 8 — Do things that don’t scale

Deliver by hand first. Sell a “productised SaaS” that is fulfilled manually behind the scenes (Andre Haykal Jr. · Andre Haykal Jr. (Jul 2024) ▶ 04:07), or take orders through a form with no backend (Nico · Nico (Mar 2025) ▶ 07:19). The presenters add that this phase is for the first customers only (presenter · Joseph (Mar 2026) ▶ 08:57).

Step 9 — Read fit before you scale spending

The signal is paying, retained users from outside your network; Krasun’s bar is ten of them (Dmytro Krasun · Dmytro Krasun (Nov 2024) ▶ 04:11). Alex Lieberman defines fit with numbers (50% would recommend it, 10-month average tenure) (Alex Lieberman · Alex Lieberman (Nov 2023) ▶ 10:17). When trials are high and paid conversion is low, one of the co-founders (Eyal or Yahli) says to fix the product before marketing (unnamed guest · Eyal & Yahli (Nov 2025) ▶ 07:08).

Where guests disagree

• Must the signal be money? Hildebrand says yes (Gil Hildebrand · Gil Hildebrand (Dec 2025) ▶ 08:06); Shin says any pre-set metric that truly represents usage will do (Brian Shin · Brian Shin (May 2026) ▶ 09:01).
• Refundable or not? Timo takes a fully refundable $500 deposit (Timo Nikolai · Josef Buettgen & Timo Nikolai (Sep 2025) ▶ 05:52); Mezzadra sold a no-refund lifetime deal on an unfinished app (Umberto Mezzadra · Umberto Mezzadra (Mar 2026) ▶ 06:19).
• Free beta? Launch free briefly, then charge (Thomas Mahony · Thomas Mahony (Mar 2026) ▶ 03:55), versus never give free accounts (Mike · Mike (Nov 2025) ▶ 06:35).
• How rough can v1 be? Ken says if you’re not embarrassed by it you shipped too late (Ken · Ken (Sep 2026) ▶ 12:56); Lane Wagner wants “minimum quantity, not minimum quality” (Lane Wagner · Lane Wagner (Apr 2025) ▶ 05:52).
• Stuck: product or distribution? Fix the product (Lane Wagner · Lane Wagner (Apr 2025) ▶ 10:36), versus the presenters’ view that you should fix distribution first (presenter · Flo (Apr 2026) ▶ 04:49).

What can go wrong

• Waitlists are slow and leak. Pat says a waitlist takes six to twelve months to build and most members will not buy (presenter · Lara Acosta (Jan 2026) ▶ 07:33).
• Refund exposure. Hildebrand’s guarantee meant he would lose all the pre-sale money if he failed to deliver (Gil Hildebrand · Gil Hildebrand (Dec 2025) ▶ 05:53).
• Fast validation suits small products. Julien Nahum notes that building in a week doesn’t work for projects one person can’t build quickly (Julien Nahum · Julien Nahum (Nov 2024) ▶ 04:25).

How strong is the evidence

Strong overall. Independent guest sources per step, with repeat guests (such as Marc Lou, Colin McIntosh and Jack Friks) counted once and host lines excluded: market research 16, conversations 18, thresholds 12, landing page or demo 13, content first 14, payment 25, MVP 35, manual first 12, fit 11. No step rests on a single person.

• Build times, sign-ups and revenue are self-reported, and build time is measured inconsistently (coding time versus idea to store).
• Survivorship: fast MVPs that failed are rarely featured.
• Pat Walls promotes validate-before-build and speed in many episodes (presenter · Gil Hildebrand (Dec 2025) ▶ 06:09). These are the presenters’ views, not independent support.
• Several points come from co-founders the transcript cannot tell apart (the Once, Pushscroll and Eyal-or-Yahli episodes).
• Possible commercial interest: Starter Story is owned by HubSpot. CJ Zafir’s episode includes his praise of Starter Story itself, and the channel carries sponsor reads from no-code and AI builders. Tools named here (Sensor Tower, Ahrefs, Bolt, ElevenLabs) are guests’ choices, not tested recommendations.

Go deeper

Researching demand before you build · Talking to customers · Validation targets and kill criteria · Waitlists and fake demos · Content first, product later · Sell before you build · MVP and shipping fast · Do things that don’t scale · Product-market fit · Copy a proven product · Content-validated app idea (post the idea before building it) · Pre-sell before you build · Lean MVP loop: ship small, get feedback, iterate or move on · Content testing ladder (tweet → thread → video) · How did you validate the idea? · How did you build it and how long did it take? · How did you get your first customers (or first dollar)? · Job → side project → quit · Services → productised → product