The Founder Decision Framework: Evaluate Any Startup Decision in 3 Dimensions

Ask any founder how they made their last big decision and you'll get one of three answers: "it felt right," "the numbers said so," or "my advisor told me to." None of these are frameworks. They're outputs of a thinking process that the founder didn't watch happen.

After analyzing hundreds of founder post-mortems — the public ones and the private ones — a pattern emerged that's almost embarrassingly clean. Nearly every fatal decision could be retroactively classified as a failure in one of three dimensions. Not five, not twelve. Three.

We call this the Demand · Supply · Processing framework, or DSP. It is the underlying model behind every evaluation SarathiOS runs, and the most useful single tool we've found for breaking down founder decisions in real time.

The three dimensions

1. Demand — Is there real, sustained pull?

The Demand dimension asks: does the market want this badly enough to pay, return, and tell others? Not "will someone use it if it's free." Not "do people nod when I describe it." Real demand has three properties:

• Pay-to-use — somebody trades money or another scarce resource (time, attention, reputation) for the thing.
• Return-to-use — they come back without you nudging them.
• Recommend-to-others — they introduce someone else to it without being asked.

If any one of those three is missing, what you have isn't demand — it's curiosity, sympathy, or novelty. Each of these decays. None of them scale. Most founders confuse all three for demand at least once, usually fatally.

2. Supply — Can you actually deliver, profitably and repeatedly?

Supply is the dimension founders most underestimate. It asks: can the thing you sell be produced, fulfilled, and supported at a cost structure that survives the next 100 customers?

Supply is not just operational. It includes:

• Unit economics that hold up at the next order of magnitude
• Team capacity to deliver without the founder in every loop
• Quality that doesn't degrade as throughput goes up
• Suppliers, infrastructure, or platforms that won't constrain you

Supply problems are sneaky because they look fine at low volume — the founder is in every loop, every customer gets bespoke attention, every defect gets personally apologized for. At 10x volume, that model is structurally impossible, and the company hits a wall it didn't see coming.

3. Processing — Do you have the cognitive and operational capacity to execute?

Processing is the dimension nobody talks about. It asks: does your company — humans, systems, attention — actually have the bandwidth to execute this decision well, given everything else it's already executing?

A 6-person team can technically launch a new product line, sure. They cannot launch it while also onboarding a new enterprise customer, completing a SOC-2 audit, and replacing a co-founder. The decision isn't wrong in isolation. It's wrong given the processing budget the company actually has.

Processing failures are the most underdiagnosed cause of startup death, because they don't look like decisions failing. They look like everything getting slightly slower, slightly worse, slightly more conflicted, until one day a deal slips, a hire leaves, and morale collapses simultaneously.

How to run a decision through DSP

The framework only works if you can write down, in a single sentence, the answer to three questions. If you can't, you're not ready to make the decision.

Step 1: State the decision in active voice with a deadline

"We are going to hire two more engineers by July 1." Not "we are thinking about scaling engineering." Active, specific, dated.

Step 2: Score each dimension 1–10

For each of Demand, Supply, and Processing, write the single most honest sentence you can about where the dimension stands right now, then assign a score.

• Demand: Is there clear, paid, repeat pull for the output of this decision?
• Supply: Can we deliver the implied capacity profitably and repeatably?
• Processing: Do we have the human and systems bandwidth to execute this while doing what we already committed to?

Step 3: Identify the weakest dimension and ask the dangerous question

Whatever scored lowest is the dimension you're most likely to be wrong about. The dangerous question is: "if this dimension turns out to be a 3 instead of a 6, does the decision still work?" If the answer is no, you don't make the decision yet — you de-risk that dimension first.

Step 4: Set the falsification trigger

Before you execute, write down: "I will reverse this decision if X happens by date Y." This is the single most-skipped step in startup decision-making, and the one that distinguishes founders who survive premature scaling from those who don't. Without a falsification trigger, you will defend the decision indefinitely. With one, you will see the failure before it compounds.

The framework applied: three common decisions

"Should we raise a seed round now?"

• Demand: Do investors want to fund this thesis, now, or just startups in this space generally? (8/10 vs 4/10 is a different decision.)
• Supply: Can you actually deploy this capital faster than you can earn it? If not, you are raising to extend, not to deploy. Different decision.
• Processing: Will the founders have the attention to run a process and the company at the same time? Raising is a 3-month full-time job.

"Should we expand to a second market?"

• Demand: Is there evidence that the second market wants this — or are you extrapolating from one warm intro?
• Supply: Is your delivery model still profitable in the new geography / segment? Often the unit economics break.
• Processing: Can you operationally support two markets without doubling the leadership team?

"Should I fire my co-founder?"

• Demand: Do customers, the team, and investors actively want this change — or are you just exhausted by the conflict?
• Supply: Can the work this co-founder is doing be done by anyone else, or are you removing a critical function?
• Processing: Can the company absorb the org-wide processing cost of a co-founder departure right now, on top of everything else?

In every one of these, the framework doesn't tell you what to do. It tells you what you are still pretending not to know. That, more than any decision rule, is what good frameworks do.

Why three dimensions and not five or twelve

We've watched founders adopt 12-factor decision matrices, OKR-aligned weighted-scoring spreadsheets, and proprietary VC frameworks with twenty boxes. None of them get used past week three.

Three is the largest number of dimensions a founder can hold in working memory while in the middle of a stressful decision. That's the design constraint. A framework that can't be used in the moment of decision is not a framework — it's an artifact.

DSP also has the property that the three dimensions are orthogonal — they fail for different reasons, with different remedies. Demand failures are fixed by listening; supply failures are fixed by re-engineering; processing failures are fixed by sequencing. Mixing them produces the worst-of-all-worlds decisions you see in failed startups: "we'll fix retention by adding a sales hire," "we'll fix the org by launching a new product."

What this framework cannot do

DSP is not an oracle. It will not tell you whether your idea is good, whether your market is real, or whether you're the right founder to build the thing. Those are deeper questions. What it does is something more modest and more useful: it forces you to see the dimensions of a decision that your intuition is currently hiding from you.

Most founder mistakes — including the ones we've made — are not mistakes of evaluation. They're mistakes of incomplete evaluation. The founder evaluated demand brilliantly and never asked about supply. Or scoped supply meticulously and never thought about processing. Or processed the org change carefully and missed the demand signal that everything had quietly shifted.

DSP doesn't make founders smarter. It makes them harder to fool by themselves.

---

The DSP framework is the operating model behind SarathiOS. If you want to see it applied to one of the most expensive decisions a founder makes, read our breakdown of premature scaling, or the diagnostic for when to hire your first sales rep.