Common wisdom in startups is "ideas are easy, execution is everything" The idea being, if someone else can execute an idea better than I can—just by me telling them about it—I shouldn’t work on it! I still think this is true, but now ideas (at least, big ones) are much harder to come by.

Simple ideas are easy. "Uber for dog walking," "Airbnb for cars," "vetcove for industrial materials," etc. If you can explain the idea in one sentence, sure, the idea doesn’t matter. It’s all about the execution.

Years ago, there were many ideas like this. The mobile + cloud platform shifts were playing out, and there were numerous "obvious" businesses to be built. Combine a one-liner, some ambition, the right set of talents, time freedom, and you could build a great company. It wasn’t easy, but it was straightforward.

There aren’t a lot of simple ideas left.

VC dollars have been at work for many decades, luring founders of all types out of their comfortable jobs to crush both obvious and hidden B2B software opportunities. Barring an obvious "why now" moment (regulatory change, technical innovation, black swan event, etc; AI being the most recent example) large[1] obvious ideas are already built.

What remains are hard-to-find ideas. Ideas that don’t fit into a one-liner explanation. Ideas that someone couldn’t execute on even if you told them about it in detail.

These ideas only emerge when you’ve toiled away in a weird domain where you’ve accumulated enough knowledge to undercover a "secret". This secret is an idea that can only be described in a well-written ten-page memo, not a one-line description. And, even after reading the memo, someone outside the domain could only start to understand how the idea could be executed.

Only someone with the hard-earned hidden knowledge can create a business out of this important domain secret. These type of ideas aren’t worthless. It doesn’t mean you shouldn’t share them publicly, but they are hard to come by. That deep domain knowledge is critical to unlocking B2B SaaS opportunities at this part of the system. Moreso than in the past, it makes sense to partner with folks who have this domain knowledge.

For instance, I’ve been working to find my next startup idea in the health tech realm. Even after working in the space for many months, if you asked me what opportunities exist in applying AI to RCM, I could give you some basic contours but couldn’t explain to you what to build and why you would win with a particular strategy. It’s really complex.

Here are some of the categories that I’ve seen these secret ideas come from:

• Deep understanding of the customer. Not just their needs (any good product person can get those) but the culture and way of thinking that dominates an industry. Certain types of people are attracted to specific industries—the people shape the industry and the craft shapes the people. If you aren’t from that industry, you can’t easily understand the people dynamics from the outside.
• Deep understanding of technology. I mean technology in the broadest sense. Understanding what is happening on the edge of science, or digging up interesting research and inventions that were already invented but forgotten about (my favorite part of the Rainmaker and Valor Atomics stories) enables you to uncover a novel way to solve an existing problem.
• Intuitive domain knowledge. Composing an elegant solution requires understanding the nuances and challenges unique to the industry enough to identify what solutions are tractable and how to avoid dead ends. This is hard because you need to keep an outsider’s mindset while being an insider. It’s impossible to see what’s broken if you believe "it’s the way it’s always been done". For instance, building SaaS tax software without ever running a tech company and understanding the accounting/compliance side of the problem would make architecting an elegant solution to the problem much harder[1].
• Cross-disciplinary knowledge. Without this, you can’t compose a creative solution. If you become too focused on your domain, you won’t know enough about the outside world to invent.
• Desire to disrupt. This is partly why ERPs and CRMs have been so resilient to outside disruption. Everyone knows they are terrible, but those who know enough about them to design an alternative are happy billing hundreds of dollars an hour to do average work. There’s no desire to invent. This is partly the brilliance of NetSuite/SalesForce/Oracle—they know that building a partner network that is happy and thriving is a strong moat around their business.
• An understanding of your unique knowledge. It’s easy to think that what you know isn’t that special. Everyone must know what you know. It’s obvious or easy. Smart people are especially bad at this—they compare themselves to other smart people, or others who are even more knowledgeable than them and discount their understanding. Being humble yet knowing where you have truly unique knowledge is hard.

More so now than five or ten years ago, this detailed understanding of a domain is the seed of a great business.

[0]
There are lots of "small" opportunities (< 10M ARR TAM) out there. They are not interesting to VCs, or to most experienced founders, so they are left unsolved. Think of plugins to extensions on any large platform, scanner driver software (one of my favorite MicroConf businesses), amazon price tracking software, or other similarly-sized IndieHacker businesses. All neat lifestyle businesses, but very different from a "large" business.

[1]
Some examples of companies that required deep domain knowledge to take the risk of building:

• Stripe. You would have thought this was a solved problem, but payment APIs were terrible from the developer’s perspective before Stripe. You had to have conviction that developer’s will hold more decision-making power as all companies become tech companies.
• Heroku. After AWS was launched cloud felt like a solved problem. However the UX was terrible and developer time was a scarce resource. If you weren’t working directly on deployment, you would have thought this was a solved problem (what’s amazing is this is still an unsolved problem since Heroku has been acquired and is effectively dead).
• Vetcove. Purchasing for vets is poorly orchestrated and there’s economies of scale being left on the table. Without working with vets this would be impossible to uncover.
• Valor Atomics. Molten salt reactors weren’t deployed and can float in the ocean to get around regulation. Without going deep on energy generation you’d have no idea this was possible.
• Rainmaker. Weather modification was invented in the 50s, never deployed, and is well researched. The west is drying out. Without going deep on water conservation there’s no way to run into this.
• Anrok. Tax is still an unsolved problem and the current software either operates like a monopoly (Avalara) or is dying (TaxJar). Unless you run an internet company that deals with state tax, and have tried multiple vendors, you wouldn’t run into this problem and would assume it’s solved.
• Stedi. Without running into EDI, and attempting to implement it as a founder or non-consultant type, you wouldn’t believe that EDI is as terrible as it is and understand how terrible the existing competitors are.