Evaluating a technical founder, and what to do when there is none
Evaluating a technical founder is the job of finding out whether the person who built the product can keep building it, can explain it to the people who will join, and knows what they do not know. A non-technical angel can do most of it with five questions, a reference method the angel checklists already describe, and one screen share. When there is no technical founder, which is common at pre-seed and is not fatal, the question changes to who is accountable for the product and whether the company has priced that gap. This page covers both cases and links the fuller team assessment used on a later-stage deal.
Published September 17, 2026. Editorial.
Key takeaways
- Five questions reveal a technical founder's depth without requiring the angel to judge code: what would you rebuild, what broke last, what did you choose not to build, who taught you the most, and what do you not know how to do.
- References work the way the ACA's 2007 guide describes for any founder: ask the given references for more names, and expect the reference chain to run to a former colleague who saw the founder ship under pressure.
- The ACA guide records that management is the single most important aspect of a deal for every investor it quotes; a technical founder is judged on the same things plus one, which is whether they can explain the product to someone who will inherit it.
- A company with no technical founder is a normal pre-seed company with a named gap; the flag is a company that has not named it, has not priced it, and treats a contractor as if they were a co-founder.
A technical founder is judged on the same things as any founder, plus one. The standard angel checklists cover the general things well: credentials, references, judgment, honesty, and whether you can work with the person for the next five to seven years. The extra thing is whether this person can keep building the product and hand it on to the people who will join. This page is about that extra thing, for an angel who does not write code, and about the company where nobody has the job.
What do the standard checklists already get right?
Management is the part of angel diligence with the most practice behind it, and the practice transfers. The ACA's July 2007 due diligence guide says that "the one thing that everyone in the venture industry is in harmonious agreement on is that management is absolutely critical and the single most important aspect of a deal". Its list is credentials, a legal background check, mindset, leadership, honesty and integrity, intelligence, and judgment.
Its reference method is the useful part. The guide quotes David Rose of New York Angels on talking to the references a founder provides and then asking those references for more names, and John May of New Vantage Group on taking that a step further, to the point that he "expects to have around 30 conversations". The Golden Seeds playbook from 2010 adds the instruction to look at the list of references and "think hard about who is NOT on the list that should be", such as the CEO of a startup the founder used to work at.
For a technical founder, that method works with one adjustment: the reference you are looking for is someone who watched this person ship software under pressure. A former engineering manager, a co-founder from a previous project, a senior engineer who reviewed their work. Ask each reference for the name of one such person. The chain usually gets there in two steps.
The UKBAA's September 2020 guide adds the question of what the team stand to lose if the business fails, and whether they can survive on a limited income. For a technical founder that has a specific form: could this person get a well-paid engineering job tomorrow, and what is keeping them here?
What five questions reveal depth without judging code?
These are the questions Reveneau's angel review puts to a technical founder in the first meeting, and none of them requires the person asking to read the answer's code.
1. "What part of the product would you rebuild, and why?" Someone who has built and watched users has a list, and the first item comes out fast with a reason. A pause followed by "nothing" is the answer of someone who has not built enough to have regrets. This question also appears in the one-hour technical check, where it has ten minutes.
2. "What broke last, and how did you find out?" Every product breaks. A founder who can name the last failure, say how they learned of it (a user, an alert, a test), and say what changed afterwards has run a product. A founder who says nothing has broken has either not launched or is not watching.
3. "What did you choose not to build, and what did it cost you?" Judgment shows in omissions. A technical founder with good judgment can name features they refused, and can name a customer they lost as a result. The answer tells you whether this person can say no to the sales side, which is the argument they will have every week.
4. "Who taught you the most about building software, and what did they teach you?" This is a proxy for whether the founder has worked alongside better engineers. A specific name and a specific lesson means they have. A vague answer means they may have learned alone, which is fine for a first product and a problem for a team.
5. "What do you not know how to do that this company will need in the next year?" The best answer names a gap and a plan: "I have never run infrastructure at scale, so the second hire is someone who has." The worst answer is that there is no gap.
Leo Polovets, who wrote the 2016 essay "Startup Technical Diligence Is a Waste of Time" that reached 287 points on Hacker News, described in the thread using his own technical background as "a baloney detector and a good differentiator from VCs who have non-technical backgrounds". The five questions are a baloney detector a non-technical angel can hold.
What does the screen share add?
One thing the questions cannot: whether the founder can explain the product to someone who will inherit it.
Ask the founder to open the code repository and, without going into the code, describe how the product is organised: where the customer-facing part lives, where the data is stored, what happens when a user does the main thing. You are listening for whether the explanation is ordered, whether it uses plain words, and whether the founder points at real things on the screen. A founder who can do this can onboard the second engineer. A founder who cannot will be the bottleneck for every hire, which is the single-developer risk in a different form, and the single-developer risk covers what to ask for in the round.
The commit history on the same screen tells you whether the founder is the person who wrote the product. If the founder claims the technical role and the log shows a different name, ask who that is.
What should you do when there is no technical founder?
First, do not treat it as disqualifying. Many pre-seed companies have a business founder and a contractor, an agency, or a technical friend, and some of them become good companies. The question is whether the gap is named and priced.
| Situation | What to ask | What good looks like |
|---|---|---|
| Contractor built the product | Who owns the code, what is the contract, how many hours a week, what else are they on | Assignment agreement signed, retainer in place, second engineer in the plan |
| Agency built the product | Who has the repository, who can deploy, what does the agency charge to change one screen | Code in the company's own repository, a named person who can deploy without the agency |
| Technical friend on equity | Are they committed full time, what is their vesting, what is their day job | Full-time with vesting, or an honest part-time arrangement with a hire planned |
| Business founder plans to hire | What month, what budget, who is helping them interview | A month, a number in the use of funds, and a technical advisor who will sit in on interviews |
The flag in every row is the same: a company that describes a contractor as if they were a co-founder, or a founder who says the technology is "handled" without being able to say by whom. The ACA guide's advice to check what the founder stands to lose applies to the technical contributor too. A contractor stands to lose a client.
Some companies fill the gap with a part-time senior engineer rather than a hire. Fractional CTO versus software development agency is the comparison, and the fractional CTO comparison page covers when that role is the right answer and when it becomes another bus factor of one.
What is different when the deal is bigger?
At pre-seed, this page is the team assessment. At seed with a team of four or five, the questions extend to how the team works together, who reviews whose code, and whether the technical founder has hired anyone before. Assessing the engineering team is the fuller method a firm would run on a later deal, and it is the page to read if the round is large enough that a wrong answer costs more than a review.
In an angel group, this assessment is usually split: a non-technical member runs the references and the general management questions from the ACA list, and the member assigned as technical reviewer runs the five questions and the screen share. Assigning the technical reviewer in an angel group covers how to divide it, and the pillar page puts the founder assessment next to the other checks.
What should you write down?
Five lines: the rebuild answer, the last-break answer, the gap the founder named, the name of the reference who saw them ship under pressure, and whether the founder could explain the product's organisation on screen in plain words. If there is no technical founder, replace the last line with who owns the code and what month the first engineer starts.
Best for
- An angel meeting a technical founder for the first time
- A group member running the management references and wanting the technical version
- An angel looking at a company with a business founder and a contractor
Avoid if
- The company has an engineering team of five or more, where the team assessment page applies
- You have a technical reviewer with code access who can judge the work directly
Verify before you commit
- Get one reference who watched the founder ship software under pressure, by asking the given references for names
- Confirm the name in the commit history matches the person claiming the technical role
- If there is no technical founder, see the assignment agreement for the code
Common questions
How can a non-technical angel evaluate a technical founder?
With five questions that reveal depth without judging code: what would you rebuild, what broke last and how did you find out, what did you choose not to build, who taught you the most, and what do you not know how to do that the company will need. Add the reference method the ACA's 2007 guide describes, adjusted to reach someone who saw the founder ship under pressure, and one screen share.
What references should an angel seek for a technical founder?
Someone who watched the founder ship software under pressure: a former engineering manager, a co-founder from an earlier project, or a senior engineer who reviewed their work. The ACA's July 2007 guide quotes David Rose on asking given references for more names and John May, who "expects to have around 30 conversations"; ask each reference for one person of that type and the chain usually gets there in two steps.
Is a startup with no technical founder investable at pre-seed?
Often, yes. Many pre-seed companies have a business founder and a contractor, an agency, or a technical friend. The question is whether the gap is named and priced: who owns the code under a signed agreement, who can deploy without the outside party, and what month the first engineer starts with a number in the use of funds. The flag is a company that says the technology is handled without saying by whom.
What does "what would you rebuild" tell an angel?
Whether the founder has built enough and watched enough users to have regrets. A real answer comes fast, names a specific part, and gives a reason grounded in users or cost. "Nothing" or a long pause followed by a marketing feature means the founder has not looked at the product with users, or the product is smaller than the deck suggests.
Why ask a technical founder what they do not know?
Because the best answer names a gap and a plan, such as never having run infrastructure at scale and therefore hiring someone who has as the second engineer. The worst answer is that there is no gap. The UKBAA's 2020 guide makes the same point for any founder: investors want to know the team understand and accept their limitations.
What should an angel ask when a contractor built the product?
Who owns the code and under what signed agreement, how many hours a week the contractor works on it, what else they are working on, what they are paid, and whether they will take a call. Then ask when a second engineer starts. A contractor stands to lose a client rather than a company, which is the difference the ACA's question about what the founder stands to lose is getting at.
What should an angel ask when an agency built the product?
Whether the code sits in the company's own repository or the agency's, who can deploy a change without the agency, and what the agency charges to change one screen. A company that cannot deploy without paying its agency has a bus factor of one and the one sends invoices. Good looks like code in the company's own repository and a named person who has deployed at least once.
How does the screen share help evaluate a technical founder?
It shows whether the founder can explain the product to someone who will inherit it. Ask them to open the repository and describe, without going into the code, where the customer-facing part lives, where data is stored, and what happens when a user does the main thing. An ordered explanation in plain words pointing at real things means the founder can onboard the second engineer.
What is a fractional CTO and does it fix the missing technical founder?
A fractional CTO is a senior engineer who works part time for the company on the technical direction and sometimes the hiring. It can close the gap at pre-seed, and it can also become another single point of failure if the fractional person is the only one who understands the product. Ask the same questions of them as of a contractor: hours, other clients, and who else can deploy.
What did the Hacker News technical diligence debate conclude?
Leo Polovets's 2016 essay arguing that startup technical diligence is a waste of time reached 287 points on Hacker News in July 2016. In the thread, Polovets described using his technical background as "a baloney detector" rather than to read code, and the top-voted reply argued that diligence earns its cost at Series A and later. Both positions support asking questions rather than reading code at pre-seed.
References
- Angel Capital Association, Best Practice Guidance for Angel Groups: Due Diligence, David Eyler, July 2007
- UK Business Angels Association, The due diligence process, September 2020
- Golden Seeds, Due Diligence Playbook, January 2010 (hosted by the Angel Capital Association)
- Hacker News, Startup Technical Diligence Is a Waste of Time, 13 July 2016 (287 points)
Related reading
Fractional CTO vs. software development agency: which do you need?
One owns your technical decisions. The other builds what you already decided. Most founders need to know which problem they actually have before they hire either.
How to choose a tech stack for a startup without shooting yourself in the foot
Most startups pick their stack for the wrong reasons. Boring and proven beats new and exciting almost every time. Here is how to choose for speed and hiring, not for a resume.
More in Is it real
Demo versus production: how to tell in a screen share
A demo is a rehearsed path through prepared data on an account the founder controls. Production is the same product under a real customer's account, with that customer's data, history and mistakes, on the address a customer would use. An angel can tell the two apart in a screen share without technical knowledge by choosing the account, choosing the task, and stepping one click off the rehearsed path. This page lists the tells, gives the questions, and explains why the distinction has become the most common technical finding at pre-seed, now that a convincing demo can be built in days with AI tools while a production system cannot.
MVP diligence: what having an MVP should mean
"We have an MVP" should mean that a specific set of first users has been given a working version of the product and the founder can say what they did with it. In angel diligence the phrase covers four different claims: a prototype nobody has used, a product used by people the founder knows, a product used by strangers, and a product strangers pay for. Each is a legitimate pre-seed position and each is priced differently. This page gives an angel the questions that separate the four, using Y Combinator's own definition of an MVP as the standard, and says what the answer should look like at pre-seed and at seed. For what an MVP is in the first place, the launching a product guide covers the basics.
The single-developer risk: bus factor at pre-seed
The single-developer risk is the risk that a startup's product stops when one person stops, because that person is the only one who can change it, deploy it, or explain it. The measure is the bus factor: the minimum number of people who would have to leave before the project stalls. A pre-seed startup with one developer has a bus factor of one. That is normal at this stage, and a 2016 study of 133 popular open-source projects found 65 percent had a bus factor of two or less. The job for an angel is to see the risk, ask whether written deployment steps and a second pair of hands exist, and price the round on the answer rather than pretend the risk is not there.