How to Sequence Startup Experiments by Business Type

A hardware startup doesn't validate the same way as a SaaS company. A two-sided marketplace doesn't validate like a direct-to-consumer service. The ten experiments are the raw material. The sequence: which runs first, which runs last is where the craft lives.

Over enough projects, certain sequences recur. What follows is the starting template for each of the four we use most often. Starting points, not recipes.

Who owns what

Inside the studio, different functions lead different experiments.

• Problem Hunting owns the ones that surface and understand customers (interviews, search trends).

• Rapid Prototyping and Testing owns build-and-measure (landing pages, Wizards of Oz).

• Branding and Storytelling owns narrative (explainer videos, pop-up stores, campaigns).

The sequence tells us who leads which step.

Sequence 1: Consumer hardware

Hardware used to be the hardest thing a small team could try. Not anymore. Off-the-shelf components, contract manufacturers happy with low volume, and crowdfunding as a pre-sale channel have collapsed the starting cost.

What still holds: hardware punishes mistakes. Software can be patched. Molded plastic can't. The sequence pushes mistakes into the cheap part of the process, before tooling locks in.

Customer interviews → search trend analysis → cardboard / paper prototype → 3D-printed prototype → explainer video → crowdfunding campaign → pop-up store.

Cheap-to-expensive, story-before-manufacturing. By the time a dollar is spent on tooling, real customers have committed at this price.

Sequence 2: Consumer software

Modern tooling dropped the cost of launching software near zero. The constraint isn't how to build — it's whether anyone will use what's built.

Front-load demand aggressively.

Customer interviews → online ads → simple landing page → email/social campaign → clickable prototype → mock sale or pre-sale → Wizard of Oz.

Demand first, experience second, delivery last. Most software teams do the opposite: six months on the product before anyone says they'll pay. That's how most software teams die.

Sequence 3: Consumer services

Service businesses are geography-sensitive. A meal delivery service that works in District 2 doesn't automatically work in Hanoi. The sequence picks one geography and goes deep.

Tightly-scoped customer interviews → search trend analysis → online ads (geo-locked) → simple landing page → email/social campaign → pre-sale → concierge delivery.

Geography, reachability, commitment, delivery. Only after the pattern proves out in one area, with paying customers getting actual value, do we consider scaling to a second.

Sequence 4: B2B2C intermediaries

Intermediary businesses sit in a structurally powerful position and rarely use it. They can go directly to end consumers, produce evidence, and walk into business partner conversations with data instead of a pitch. Most B2B2C founders skip this, walk in empty-handed, wonder why the partner doesn't return the call.

Load the team up with consumer-side evidence before the B-side conversation starts.

End-consumer interviews → simple landing page → explainer video → pre-sale → concierge delivery → buy-a-feature exercise → data sheet → partner and supplier interviews.

Consumers first, partners last. By step eight, the team isn't pitching. It's showing evidence. A founder walking in with signed-up pre-sales and concierge testimonials negotiates from a different position than one with a slide deck.

Rules for using sequences in practice

• Pick by business model, not by interest. A B2C service team running a hardware sequence will produce the wrong evidence in the wrong order.

• Reorder aggressively if the hypothesis stack demands it. Sequence is the default. If the riskiest assumption is at step seven, it gets tested first. Risk ordering beats sequence ordering, always.

• Not every step needs a full six-week cycle. Sometimes two or three compress into one. Landing page test and Meta ad test often share a week.

• Unusual business models deserve custom sequences. Marketplaces, deep tech, regulated industries have specific dynamics the standards miss. Go back to the Pitch, look at the Critical Assumptions, design from there.

Closing the loop

Nine posts in, a pattern should be visible.

We started with a philosophy: ideas are hypotheses, experiments are how we take them seriously. We narrowed to three sectors serving the modern Vietnamese family. Added a strategic lens for picking ideas (Blue Ocean) and two canvases for articulating them (VPC, BMC). Introduced the Pitch that shapes each idea into something testable, the Six-Week Cycle that moves it through evidence, the Library of ten experiments, and now the sequences that string those experiments together.

None of this is ritual. The canvases aren't sacred. The cycle isn't a religion. The sequences aren't a recipe book. They're scaffolding so the team can argue about the same idea, with the same definitions, against the same evidence, and move forward honestly.

The question that opens every project is still the simplest: is this a new game, or are we joining an old fight? If it's a new game, can we name the customer (a Vietnamese family), the pains, the value, the model? Can we shape it small enough to test in six weeks? When the evidence comes back, are we disciplined enough to listen, even when the answer isn't what we wanted?

Twenty blue-ocean companies in ten years is the mission. Every survivor was pressure-tested against every step in this playbook. The nine posts you've just read are what those survivors have in common. The rest (the ones that didn't make it) taught us the method by dying cleanly.

That's the whole game. Now go run it.

← Previous: The Experiment Library
🏠 Back to the start: Inside a Startup Validation Studio