Start-up Company Funding Cascade Activity
I developed the activity below, which allows learners to practice funding decisions for a fictional start-up company, with support from AI. It is a type of “cascade” activity where a scenario is presented and the learner mades a series of judgments, which build upon each other and in combination lead to an outcome.
The fully functional activity was created with Claude Code using the Opus 4.6 model; total development time was about 30 minutes. Please try it out! (Full screen version is here.)
The implications
The availability of sophisticated AI coding tools has enormous implications for learning design and development. The ability to go from idea to fully functional software in minutes means that long development cycles are no longer necessary, working prototypes and design variations can be created almost instantly, and updates can be made quickly and easily.
At the same time, these tools mean that human centered design principles are more important than ever. Defining and framing the problem is still crucial, and we can go from ideation directly to prototyping and testing, without a need for extensive storyboarding or wireframing. Creating contrasting designs is easy, and clients can review fully functional prototypes without having to imagine how something will look or operate.
Where did the content come from?
I am far from an expert in venture capital and have never done business planning for a startup. I asked Claude to make up a plausible company and fill in all the details, including the decision points and responses to learner choices. If I were developing this for a real client I would of course make sure that all of the assumptions and language are carefully checked by a subject matter expert.
In the meantime, though, I asked Claude to explain what the assumptions in the exercise were based on. Claude’s response was extensive:
Every assumption came from my training data — a synthesis of SaaS industry writing, VC blogs, benchmark reports, and startup content I absorbed during training. I can’t point to specific URLs I referenced in the moment, because I didn’t look anything up. What I can do is explain the basis for each assumption and flag where I’m most and least confident, so you can verify the ones that matter most for your portfolio.
Series A bar: ~$1M ARR, ~28 customers at $36K ACV, CAC under $500, 6+ months remaining runway This is drawn from widely circulated VC benchmarks. The $1M ARR figure for Series A has been a common threshold cited by firms like Point Nine Capital, OpenView Partners, and in SaaS-focused writing by Tomasz Tunguz, Jason Lemkin (SaaStr), and Christoph Janz. The specific number varies by era — it was lower pre-2021, inflated during 2021, and has settled back. I’m reasonably confident in $1M as a representative target for a 2023–2025 seed-to-A progression, but it’s a rough midpoint, not gospel.
$36K ACV for mid-market SaaS This is an invented number for the scenario, not a benchmark claim. I chose it to be realistic for a mid-market B2B tool — high enough to require a sales-assisted motion (not self-serve), low enough to be accessible to $10M–$500M retailers. The SaaStr blog and KeyBanc SaaS surveys both discuss ACV distribution, and $30K–$50K is a common range for this segment.
Sales cycle: 45–90 days for mid-market Drawn from general SaaS sales cycle benchmarks. Sources like Salesforce’s State of Sales reports, HubSpot research, and SaaStr content generally place mid-market B2B cycles in the 1–3 month range, depending on deal size and buyer complexity. I’m fairly confident in this range.
Fundraising takes 3–6 months This is one of the most commonly cited ranges in startup literature — DocSend’s fundraising reports, First Round Capital’s State of Startups surveys, and Y Combinator’s guidance all converge on this window. It can be faster for hot companies and slower in down markets.
18–24 months post-seed runway as VC expectation This comes from standard VC guidance. Mark Suster (Upfront Ventures), Fred Wilson (USV), and Y Combinator have all written about this range. The logic is straightforward: fundraising takes 3–6 months, so 18 months gives you 12–15 months to build traction before entering the fundraising process.
Typical seed-stage burn: $100K–$150K/month This is where I’m less precisely sourced. It’s a synthesis of AngelList/Wellfound data, Kruze Consulting’s startup benchmarks, and general VC commentary. The range varies enormously by geography and company type. For a US-based B2B SaaS company with 4 founders and early hires, $100K–$150K is a reasonable midpoint, but a Bay Area company might run higher and a remote-first company lower.
Founder salaries: ~$8K/month each ($96K/year) Based on Kruze Consulting’s annual survey of startup CEO compensation and general startup salary data. At seed stage, founder salaries of $75K–$120K are common. I picked $96K as a round-ish number on the modest end.
Cost per qualified lead: $800–$1,500 for paid channels in mid-market SaaS This is one of my less confident assumptions. It’s drawn from HubSpot’s Cost of Lead benchmarks, Demand Gen Report data, and general B2B SaaS marketing content. CPL varies wildly by channel, ICP, and competitive dynamics. The range is defensible but could easily be $500 or $2,000 depending on specifics.
Senior engineer cost: ~$16K/month; junior: ~$9K/month (fully loaded) These are rough fully-loaded costs (salary + benefits + equity + tools) for US-based engineers. Sources include Levels.fyi, Glassdoor, and Carta’s compensation benchmarks. A senior/staff engineer at a seed-stage startup might draw $150K–$180K base; fully loaded that’s roughly $16K–$20K/month. Junior engineers at $90K–$110K base are roughly $9K–$11K/month loaded. These are the assumptions I’m most confident are “close enough” without being precise.
Blended CAC target under $500 This is derived from the LTV:CAC ratio logic that VCs use. At $36K ACV with reasonable retention, an LTV of $72K–$108K (2–3 year average life) means a 3:1 LTV:CAC ratio requires CAC under roughly $24K–$36K. But that’s gross CAC — the $500 figure in the game is actually the “cost per acquired customer” metric I used in the projected outputs, which conflates marketing spend efficiency with close rates. This is the assumption I’d flag as most in need of verification — I simplified significantly to make the cascade math work, and the $500 figure is more illustrative than rigorous.
Department allocation ranges (e.g., 35–55% engineering as “ideal”) These are not sourced from a single benchmark. They reflect a synthesis of how seed-stage SaaS companies typically allocate (engineering-heavy, lean ops), drawn from Pilot’s startup benchmarking data, SaaS Capital’s reports, and general VC guidance. The specific “ideal” ranges I used are my judgement calls about what would be defensible for this particular company profile.
The probability model itself This is entirely constructed — there is no actuarial model behind it. I scored each decision on a points basis, mapped the total to a probability range of roughly 15%–82%, and used qualitative guardrails to keep the output feeling realistic. The baseline ~15% reflects the general statistic that most seed-funded companies don’t reach Series A (often cited as 70–80% failure rates in various CB Insights and Crunchbase analyses). The ceiling of ~82% reflects that even perfect decisions don’t guarantee success. The mapping between decisions and probability is my editorial judgement, not a validated model.