Every business starts as a hypothesis: “I believe these people will pay for this thing.” The question is how long it takes you to test that hypothesis. Some founders take a year. You can do it in a week.
Not a sloppy, half-effort week where you “kind of look into it.” A structured, focused week where you run a real experiment with real people and end with a clear answer. I’ve designed this framework from years of working with startups at the Startup Burgenland accelerator and from running my own experiments at Vulpine Creations.
Seven days is enough time to go from hypothesis to evidence. Here’s the exact schedule.
The Rules of the Experiment
Before we start, three rules:
Rule 1: One hypothesis per experiment. Not “let me test whether people want this, whether the price is right, and whether my marketing works.” One thing. “Will freelancers pre-order a client follow-up tool at EUR 29/month?” That’s your hypothesis.
Rule 2: Define success before you begin. Write down the specific number that means “yes, this works.” Five pre-orders? Twenty email sign-ups? Three paying customers? Pick the number Monday morning and don’t change it.
Rule 3: Seven days, no extensions. The experiment ends Sunday night regardless of where you are. If you didn’t reach your success metric, that’s an answer. A disappointing answer is still an answer.
Monday: Define and Prepare
Morning: Write the Hypothesis
“I believe [specific people] will [specific action] for [specific offering] at [specific price] because [specific reason].”
Fill in every bracket. If any bracket is vague (“some people” instead of “freelance designers in the EU”), you’re not ready to run the experiment. Spend Monday morning getting specific.
Afternoon: Prepare the Assets
Build whatever you need to run the test. At minimum:
- A description of your offer (one page or less) that covers the problem, the solution, and the price
- A way to collect commitments (email form, payment link, sign-up page)
- A list of 30+ people you can reach who fit your target customer profile
If you’re creating a landing page, keep it dead simple. Five elements: headline, problem, solution, price, action button. Ship it ugly — you’re testing the offer, not the design.
Evening: Prepare Your Outreach
Write the messages you’ll send to your list. Customize for different channels (direct message, email, community post) but keep the core consistent: here’s the problem, here’s what I’m building, here’s how to get involved.
Tuesday-Wednesday: Customer Discovery + Early Outreach
Customer Conversations (Tuesday)
Reach out to 10-15 people from your list with a conversation request, not a sales pitch:
“I’m testing an idea this week and would love 10 minutes of your input. I’m exploring [problem area]. Would you have a few minutes today or tomorrow?”
Run 5-8 conversations. Use the standard discovery format:
- How do you currently handle this?
- What’s most frustrating about it?
- What would a good solution look like?
These conversations serve two purposes: they validate the problem (or reveal that it’s different than you assumed), and they warm up potential customers for the offer you’ll present later in the week.
Begin Outreach (Wednesday)
Now share your offer. Send your prepared messages to everyone on your list. Post in relevant communities. Share on social media.
Track every interaction in a simple spreadsheet:
| Person | Channel | Contacted | Response | Action Taken |
|---|
The data matters more than the feelings. If you sent 30 messages and got 2 responses, that’s data. Don’t interpret it emotionally. Record it.
Thursday-Friday: Push, Follow Up, and Adjust
Thursday: Follow Up and Expand
Follow up with everyone who didn’t respond. A simple “Hey, wanted to make sure you saw my message” converts a surprising number of non-responses into conversations.
If your initial outreach produced feedback that suggests an adjustment (wrong price, confusing description, different audience than expected), make the change today and reach out to a new batch of people with the updated version.
This mid-week adjustment is one of the most valuable parts of the experiment. It’s where validated learning happens in real time.
Friday: Final Push
This is your last full day of active selling. Send final follow-ups to anyone still on the fence. Make one more community post. Reach out to anyone on your list you haven’t contacted yet.
By Friday evening, you should have your maximum number of commitments (or non-commitments). The weekend is for analysis, not outreach.
Saturday: Analyze
Compare Results to Success Criteria
Pull out the success metric you defined on Monday. Did you hit it?
Above threshold: Strong signal. The hypothesis is supported. Plan your next steps for building and delivering.
Near threshold (within 30%): Moderate signal. Something is working but not everything. Review the data to identify what’s holding it back and design a follow-up experiment.
Well below threshold: Weak signal. The hypothesis as stated isn’t supported. But the specific way it failed tells you something.
Extract Specific Learnings
Don’t just record “it worked” or “it didn’t.” Answer these questions:
-
Which audience segment responded best? If freelance designers were enthusiastic but freelance developers were indifferent, you’ve learned about your real audience.
-
Which objections came up most? Price too high? Feature missing? Don’t trust a new product? Each objection points to a specific fix.
-
What language worked? The words and phrases that made people lean forward are your future marketing copy.
-
What surprised you? Often the most valuable insights are the ones you didn’t expect. A use case you hadn’t considered. A customer type you hadn’t targeted. A feature request you hadn’t imagined.
Sunday: Decide
Strong Signal Decision
If you hit your success metric, your next step is to deliver value to the people who committed. If they pre-ordered, build and deliver. If they signed up, create the first version. If they expressed strong interest, make a formal offer.
Plan your first week of delivery — what you’ll build, how you’ll deliver it, and how you’ll gather feedback from early users.
Moderate Signal Decision
If you were close but not there, design one more experiment to test the specific bottleneck you identified. Was it the price? Test a different price. Was it the audience? Test a different segment. Was it the positioning? Test different messaging.
One variable at a time. Don’t redesign everything — that’s not iterating, it’s starting over.
Weak Signal Decision
If the results were poor, you have three honest options:
- Pivot the offer and run another experiment targeting a different aspect of the problem or a different audience
- Park the idea and revisit it in 3-6 months with fresh eyes
- Kill the idea and move on to something else
Use the kill-or-commit framework to make this decision based on evidence rather than emotion.
Why One Week Is Enough
The one-week constraint works because it eliminates the three biggest time-wasters in validation:
Over-preparation. You can’t spend three weeks on a logo when the experiment ends in seven days. You do the minimum necessary and focus on what matters: talking to customers and making offers.
Analysis paralysis. The time pressure forces decisions. You can’t deliberate indefinitely about the “right” price or the “right” message. You pick one, test it, and learn.
Emotional investment. After one week, you’re not deeply attached to the idea. Walking away, if the evidence says to, is uncomfortable but not devastating. After six months of building, walking away is excruciating. The one-week format keeps the stakes proportional.
Takeaways
- One hypothesis per experiment. Test one specific thing. Don’t bundle multiple questions into a single test.
- Define your success number on Monday. Don’t move the goalposts mid-week. The number is the number.
- Spend Tuesday-Wednesday on discovery and first outreach. Real conversations plus real offers to real people.
- Use Thursday-Friday to follow up and adjust. Mid-week adjustments based on real feedback are where the best learning happens.
- Decide on Sunday. Strong signal: deliver. Moderate: iterate one variable. Weak: pivot, park, or kill. The data makes the call.
