How to Start.
How to Start a Startup
Today we’re talking about “how” to get started on your startup.
A lot of founders make the mistake of spending time on a name or company registration at the very beginning. When in reality, we just need to start talking to people.
1) First, we pick a problem, or two, that we are interested in researching. It might be a very broad problem, like “people spend too much time on email” or something more vertical like “Factories can’t stop downtime”.
2) Next, we need to survey people who have, or might have, this problem. Most founders settle for surveying 20–30 potential customers, the best founders usually talk to over 200.
3) Then, if our initial customer research indicates the problem is significant and ideally urgent for some segment of those surveyed, we can put together a demo. But don’t try to build a whole product at this point, use a tool and put together a simple visual workflow. It shouldn’t be more than 2 days of work, max.
Best of luck out there.
Starting Up: Finding a Problem
Today we’re continuing to talk about how to start a startup, looking more at how you find a problem to work on.
1. First, we select some problems we are interested in researching. We don’t need a detailed vision initially, but some rough ideas of a focus area should be written down. Don’t be afraid to spend a few hours brainstorming.
2. Next, we need to assess if our team is capable of solving these problems. If you’re building customer support software, most coders might be able to build a basic demo, but if we’re talking about biotech or space rockets, for example, then that will likely require specific skill sets on your team.
3. Lastly, we think about who might need our product and try to come up with at least 3 initial “customer profiles”. Who appears to have this problem? How many of them are there? What is our revenue potential if we sell to them? Ask those questions for each customer profile you initially create.
With your rough thesis in place and numerous avenues to explore — we are now ready to begin our customer research.
Best of luck out there.
Your First Customer Research
Today, we’re talking about your first customer research.
Once you’re armed with a rough idea of a problem to solve and who might need your help, we have to start talking to people.
1. First, remember to spread out your initial research across 3–5 different customer profiles.
Different types of customers in different locations can have very different problems, needs, and budgets. Don’t limit your conversations too early.
2. Next, make sure you’re asking for advice.
We’re not selling anything yet — instead, let people know that you’re building a product, especially for them, and you need their help — because of their strong experience or situation.
3. Lastly, talk to lots and lots of potential customers.
Most first-time entrepreneurs will only do 20–30 interviews, but successful serial entrepreneurs usually do around 250. That’s because they’re not willing to commit years of their life to a problem without knowing it’s important.
Best of luck out there.
Your First Demos and Trials
Today we’re talking about your first demos and trials.
Once you’ve completed enough customer research to confirm a problem and a customer profile to test, we now need to show people something.
Many founders make the mistake of trying to build a full app at this point, but that’s overkill.
1. 1st, we need a demo — this is mostly a walkthrough of what the product might do. For both software and hardware companies — this is a design rendering. You really don’t need to build anything yet.
2. Iteration. We show the demo to potential customers who fit our customer profile, and based on their feedback, we iterate on the demo. I suggest a group of at least 15 people actively providing this feedback as “design partners” or “customer advisors”.
3. Once we have a demo our potential customers are excited to try, then we can move them to a trial version. Although some products may require a basic MVP at this point, you’d be amazed at how much you can get done manually or using no-code tools.