Several friends recently asked about how to pick startup ideas. Focusing on the right problem to solve is maybe the single most important determinate of startup success that’s within a founder’s short term control.
My mental model for picking startup ideas is a prospector in the old west. Each day you have to wake up and dig holes, hoping to strike gold. You can learn to spot signs on the surface of the earth that offer hints about where you might find gold, but ultimately you need to dig a hole to find out. The more you dig, the more data points you get and the better you get at reading the geology of an area.
I wrote this down this checklist for myself to make sure that I’m being efficient with my time and killing bad ideas before investing too much time into them. I thought it might be helpful to just share my framework more generally for people who are thinking about stating a new business soon.
The goal of the process is to de-risk the idea to make sure you’re working on a good one. What you’re looking for is a high-priority problem for customers that you can solve in a differentiated and much better way because of a new technology.
The Process:
Step 1: Keep a list of ideas. When you notice problems in day-to-day life, write them down. Staring at a blank page to brainstorm ideas is much harder. I just keep it in Apple notes on my phone and mine is now about 200 ideas long now since it’s been accumulating for years. The best way to find ideas is when you are frustrated about something painful or wrong in the world, or when you hear about someone else feeling that way.
Step 2: Pre-screen an idea (2-3 days)
Take 2-3 days to get smart on the idea. Read up on the market (size, segments, growth rate, value-chain, suppliers, incumbents, challengers, etc) and customers (top pain points, surveys, published interviews). I’ll keep a glossary of industry terms. Usually this is desk research, but it could also involve calling a friend or two who are experts. Write down notes so you can circle back later.
Criteria:
Bet on a high growth trend (anything growing +20% is interesting, and +100% is exciting). Rapid changes create opportunities if incumbents are slow to react, allowing a newcomer (a startup) to win market share and occasionally replace them. There are two predicable ways to monetize a trend:
New platform strategy: spot a frontier technology with growing adoption (eg PCs for Software, modems for internet, iPhones for mobile apps, payment/BaaS APIs for fintechs) and re-build an old business model on the new platform. Focus on industries that will be deeply changed by the new technology where incumbents will struggle to react. For example, Bezos saw the internet growing +2,300% in 1994 and decided to build an online bookstore, Amazon. Salesforce re-built CRM in the cloud. Uber re-built a taxi company on the mobile platform. Expedia built an online travel agency.
Picks and shovels strategy: spot the creation of a new market (due to new tech, regulatory changes, or shifting consumer preferences), and build infrastructure (“picks and shovels”). For example in the 1990s, PayPal was built on the back of eBay to support their seller ecosystem. Stripe and WeWork were old infrastructure re-branded for startups in the 2010s.
Find a small niche to start in (a small segment to monopolize with a 10x better product). TAM doesn’t matter for the first niche. It can be tiny. But you need a niche where you can be n of 1 to start. Once you get the product dialed in, expand. If there’s no un-served or under-served niche, it’s a no-go.
10x better product idea (new tech or regulation must unlock a 10x better solution to an old problem). Review the history of software in the category. Are they delayed in adopting the latest enabling infrastructure? For example in expense management, Ramp built a 10x better product than Expensify because it leveraged fintech APIs to do virtual card issuing.
Economies of scale or network effects ( as the company gets bigger, it becomes more valuable to customers ). This point is more academic than practical, and is true of most software companies by default.
Step 3: Expert calls (3-5 weeks)
Do 30 minute meetings with experts on the industry. Start with 2-3 that you’re close with, then go broad through cold outreach. There first few might be a wash since you’re triangulating the stakeholders in the buying decision. Usually 10-20 solid interviews is enough to get the big picture. The key groups are: customers, competitors, other founders, and investors. I source these through intros or by sending cold messages on LinkedIn.
Elimination criteria:
Doesn’t matter to the customer (must be a top 5 challenge). If the customers says the problem I’m solving doesn’t matter to them or doesn’t bring it up as a priority, then you’re working on solving the wrong problem
Product isn’t 10x better (product must have 10x ROI vs incumbent solutions). How to measure that at the idea stage? It’s a subjective judgement. Ask customers how the product could impact their business, and ask them to quantify the value. Ask: is it 10x faster, cheaper, or better?
Not sufficient willingness to pay (ACV above $12k/year). The ACV has to be over $12k/year to support a sales-driven go-to-market motion. This is much easier if the customer is used to paying for a similar service and has budget already. You can verify this by asking customers what pricing structure is most comfortable with them and how much they’d pay for it. I’ve found enterprise buyers to be surprisingly honest.
Long sales cycle (max is 3-6 months). Long sales cycles make it hard to get reps in and learn as an early startup. Longer sales cycles favor incumbents and mature companies. You can learn what the sales cycle will be by talking to sales people selling to the same customers. I’d rather go up market later on.
Mis-aligned incentives (buyer must be incentivized to purchase). Occasionally customers have mis-aligned incentives even though it feels like they should want the product. An example is selling productivity tools to industries that bill by the hour. Being more productive could decrease their top line.
Step 4: Shop a memo with close advisors/investors (1-2 weeks)
Draft up a 2-3 page memo or 10-slide deck on the idea and get feedback from a small group of people you trust. Make sure at least a couple of them have deep domain expertise. Treat them like your “board of directors” or “investment committee” of close friends and advisors.
Elimination criteria: Ask each of them: “do you think I should pursue this or not?” Guide them to be objective and criticize the idea. If the majority “I’m not sure about this one,” then dig into why they’re hesitating. If you think their fears are valid and they outweigh the pros, scrap it.
Step 5: De-risk the idea (3-4 month deep dive)
Secure 3-5 design partners (can be unpaid at first, but willing to pay in the future). The ask of them is that they start using the tool and give detailed feedback and access to their staff.
Develop a prototype of the product to get customer input.
Elimination criteria: If customers are unwilling to sign up as design partners, or if the product turns out to be technically much more difficult than expected, seriously reconsider the idea before investing more time into it.
If you make it through all that without eliminating the idea, it’s probably worth pursuing and investing years of your life into. I don’t have a crystal ball and it is impossible to predict the future completely, but at this stage at least you’ve done your homework and can be confident that you’re taking an educated bet.
I’d love to hear your feedback on this post. Please let me know what you liked and how it can be better. You can just send me a message or leave a comment.