Do things that don't scale
Do things that don't scale — is one of YCombiantor's mantras. The essence of the mantra is to do unscalable things with your hands: conduct interviews, sell to customers, ship services manually, etc.
💎 Stripe case
Stripe is an online acquiring service and a YC alumnus. The founders of Stripe used to come to the office and install the code manually before the first thousand customers. During the S09 stream of the YCombinator accelerator, Patrick Collison conducted an incredible number of interviews and manual installations, and in his honor, YC coined the internal term "Collison Installation": if regular funders agree to the question: "Are you ready to try our beta?" and sent a link, Patrick Collison would say: "Okay, give me your laptop," and installed the stripe right away! It's not a scalable solution, and you can't grow a business on it. But after conducting many iterations of installing the code manually together with the client's engineers, they received a ton of feedback on their product.
The huge time costs that go into unscalable actions, you share knowledge about your customers, who they are, what is important to them, how they cope with their tasks, use your product, etc. The founder of Wealthfront (an investment service) used to sit down with his clients with a pen and paper and work with them to come up with their wealth management strategy. The founder of Food on the Table (meal planning) personally went to the supermarket with his clients and helped them choose products and plan their diets.
Start with a small segment
One of the mistakes startups/products make is to make a product for several segments, preferably to satisfy ALL possible customers. This leads to the fact that the scoop is constantly increasing, there is no filter for new wants, the complexity of the system and, as a result, the entropy of the code increases => each new feature leads to a rise in the cost of the next features => the product does not launch OR it is stillborn and is nailed in N months.
Start with a small segment whose needs overlap with those of neighboring segments. And this segment are the opinion leaders for the other segments. Logic:
- when you make a product for a small segment, the MVP and first version scopes will be limited + you can always choose the necessary 20% that will provide 80% of the value for this segment
- since the segment is small, you can literally conduct problematic/JTBD interviews with each representative of the segment in person and then decisive interviews, keep in touch with everyone and collect feedback
- Feedback from the segment will be consistent and easier to interpret and make decisions based on it
- product adoption among the segment that has a stronger pain/readiness to activate your product => faster first sales and faster feedback