Playbook · Budget reality
What I'd actually do with $5,000 to launch one of these ideas
A real budget breakdown from someone who's built and shipped this kind of product before. Most of the money goes where you don't expect.
You unlocked a $5 dossier. You read it twice. The math looks plausible. The MVP is buildable. You have a day job or a runway, and somewhere around $5,000 you can put against this without your spouse asking questions. What do you actually spend it on?
Here's the honest answer, from someone who's launched five of these and watched maybe a hundred friends launch their own. Most of the budget does not go where new founders think it does.
The wrong mental model
New founders see $5,000 and start adding up tools. Stripe ($0 + fees), Vercel ($20/mo), Postgres ($25/mo), domain ($15), OpenAI/Claude API ($100/mo at minimum). Maybe a logo on Fiverr ($50). They get to $1,500 over a year on infrastructure and feel rich.
That's because they're thinking like a developer building a portfolio piece, not an operator launching a business. The thing that costs money in launching a B2B SaaS isn't the stack. It's customer discovery, distribution, and the part of your time you can't recover.
If you're a developer and you spent $1,500 on tools and the product didn't sell, you didn't have a budgeting problem. You had a distribution problem. And distribution problems are solved with calendar time, conversations, and one specific kind of money: money that pays you to NOT take freelance work for 90 days.
The actual breakdown
If I had $5,000 in cash and a Wishdeal Factory dossier I believed in, here's exactly where I'd send the money over the first 90 days:
Look at that allocation. $1,000 goes to talking to twenty potential buyers. $1,000 goes to NOT taking freelance work for a month. Together that's 40% of the budget on things that don't write a single line of code.
The actual tooling and infrastructure is $1,200. The marketing budget is $1,300, and most of THAT goes to physical/real-world presence (a conference badge, a photographer for the landing page) rather than digital ads.
Why customer discovery costs $50 per interview
The single best money you'll spend is paying twenty potential buyers $50 each for a 30-minute call. Yes, you're paying them. Yes, this feels weird. Do it anyway.
Here's the math: $50 to a senior person earning $150-300/hour means they take the call seriously. They show up on time. They prepare slightly. They give you real answers instead of polite ones. They might even take a follow-up message six weeks later when you have a demo.
Twenty calls over 30 days teaches you more about whether your idea is real than any amount of building. You'll hear the same complaint phrased five different ways, and suddenly you know the headline for your landing page. You'll hear three people independently mention the same competitor, and you'll know who your real comparison is. You'll hear someone say "I tried something like this and stopped using it because X" and that's worth a thousand dollars of pivot avoidance.
The dossier names the ICP for you. The dossier even names a few specific personas. Use that. Cold email people who match the persona, offer $50 for 30 minutes, accept that 80% will ignore you and 20% will say yes. You need 20 yeses, which means about 100 cold emails.
Why insulation money matters more than build money
The reason most ideas with good dossiers don't ship isn't because the founder didn't know what to build. It's because four weeks in, a freelance gig pays $4,000 for two weeks of work, and that money is real and the product is not. The founder takes the gig, loses momentum, never recovers.
$1,000 of insulation money is not a salary. It's the difference between saying "I can't, I'm in week 5 of a launch" and saying "well, I do need the cash..." It buys discipline in the worst possible moment.
If you don't have freelance temptation - if you're employed full-time or you're independently wealthy - skip this line. Move the $1,000 to two more months of paid customer interviews ($1,000 = 20 more calls).
Why a real photographer beats AI imagery
A $400 photographer shoot gives you 8-12 real photos of real people in real contexts. You use them on the hero, in case-study mockups (clearly labeled as illustrative), in the about page. Real photos communicate "this is a real business with real people" in a way that nothing else does.
AI-generated imagery in 2026 is uncanny-valley enough that buyers can usually tell. Stock photos look like stock photos. A custom 1-day shoot at a friend's office or a coworking space, with a few willing humans posing in genuine work scenarios, is the cheapest way to look real.
If you can't afford $400, do it yourself with a phone in good window light. The bar is "looks real and intentional," not "looks professional."
The paid distribution test that's worth $500
Over the 90 days, run exactly one paid distribution test with a clear hypothesis. Not "do Facebook ads work." Something specific: "I'll spend $500 sending qualified traffic from a niche newsletter to a landing page. If 3+ people convert to a demo call, distribution is solvable. If 0 convert, the offer is the problem."
The point of the test isn't growth. It's information. You're proving or disproving the assumption that distribution is winnable in this category at all. Skip generic Google Ads. Pick the channel the buyer actually pays attention to.
For B2B vertical software, that's usually: a relevant industry newsletter (sponsorship: $300-1500 for a single mention), a podcast ad (small podcast: $200-500), or a Reddit/HN post you actually wrote yourself (free, but you should value your time).
What about the conference money?
$800 buys you a conference ticket plus minimum travel, or a year of an industry association membership, or attendance at a couple of small in-person meetups in the buyer's space.
The reason this is in the budget: if you're building for a vertical you don't naturally belong to (the dossier is for a buyer you're not), conferences and association rooms are the fastest way to absorb the language, the unwritten rules, the actual current pain. You'll meet 5-10 prospects in person who will take your call afterward. You'll meet a couple of partners who can resell. You'll get a calibration on whether the pain is actually as bad as the dossier claims.
If you ARE the buyer for your own product (you're building a tool for indie devs and you're an indie dev), skip this. Move it to customer discovery instead.
What you do NOT spend on
- A logo. Use a simple wordmark in a clean font. You can buy a real logo when you're at $5k MRR.
- Branding work. Same logic. Look professional, not artisan.
- A team. If you can't ship the MVP solo or with one collaborator, your dossier is too big.
- Generic ads. Don't dump money into Facebook/Google ads with no specific hypothesis.
- An MVP that takes 3+ months. If the dossier's MVP isn't 30 days, either narrow scope or pick a different idea.
- Tools that solve a problem you don't have yet. Skip the analytics dashboard until you have users.
What this budget assumes
This is for a solo founder who can build the MVP themselves or with a small amount of help, who has at least 90 days of runway separately, and who's targeting a B2B vertical with a defined ICP.
If you're consumer, distribution costs go up dramatically and customer discovery looks different (focus groups not 1-on-1 interviews, smaller payment amounts). If you have a co-founder, you have more time but more coordination overhead. If you have a team, this budget is irrelevant - your problem is salary burn, not $5k allocation.
If you'd rather skip building and just hand the dossier to someone who'll run it for you, that's the operator partnership tier. Same dossier, different execution model.
The TL;DR
- 20% on tools and infrastructure - actually a lot less than people guess
- 20% on customer interviews - paid, structured, 30-minute calls with the named ICP
- 20% on insulation - one month of freelance-avoidance money
- 16% on industry presence - conference, association, or community membership
- 10% on real photography - to look like a business not a side project
- 10% on a single paid distribution test - with a falsifiable hypothesis
- 10% on buffer - because something always comes up
Read your dossier. Walk through this budget. If the budget doesn't fit the idea you've unlocked, the idea is probably too big or too small for the $5k constraint - and that's useful to know before you spend anything.