Code Brain | Where Do Great B2B Startup Ideas Come From? (Part 2)

In our previous post, Where Do Great B2B Startup Ideas Come From?, we broke down what makes a strong B2B startup idea and walked through the first of three paths to finding one: "pain you've personally experienced." Today, we continue that thread, using Figma, Vanta, Notion, and others to illustrate the second path.

Think and explore: Pick a domain that interests you, analyze it to find a market gap, then go talk to ten potential customers and listen carefully to what they need and what frustrates them.

Everyone likes to say startup ideas shouldn't come from sitting in an office analyzing spreadsheets, yet plenty of founders' ideas do exactly that. The examples below show what proper desk research looks like — and what it doesn't (for instance, you should think and build in a closed environment first, rather than immediately interviewing potential customers and killing the idea). Of course, landing on the right idea still comes back to what we discussed earlier: finding a problem market with three key traits: 1. People care enough to pay for it; 2. Service levels are low; 3. You're genuinely interested in it.

Ideally, your idea should also draw on experience you have in a given area, though that's not strictly necessary. Figma's two founders narrowed their focus to two promising domains, then chose the one that interested them most: design.

"When we got the Thiel Fellowship, I called Evan [Wallace, co-founder] and said: 'Hey, we've got funding — $100K for two years. Want to actually do this?'

When you start a company, one useful question is: why now? It's a very powerful framing. The answer might be social — a new cultural trend. It might be regulatory — some law passed or repealed.

My favorite answer is: a technology breakthrough.

For us, it was seeing several technologies in 2012 — drones and WebGL among them — that suddenly made new things possible.

We didn't get anywhere with drones. Evan wasn't interested in hardware; he thought hardware was lame, the iteration cycles too long. So we pivoted to WebGL.

We were also interested in computational photography, but quickly realized that path led to a consumer app, whereas WebGL was browser-based. Why edit photos on anything other than your phone? We realized we'd picked the wrong direction and started shifting toward design. I'd done design work during my internship at Flipboard, which helped me see that design was a viable path."

— Dylan Field, co-founder and CEO, Figma

Vanta founder Christina Cacioppo similarly narrowed her scope to two areas of interest: security and collaboration. Then she kept talking to customers to find pain and demand.

"Our early ideas were a mess. We did a lot of desk research behind closed doors. I think the flaw in that approach is it's too abstract — no user or customer perspective.

At the time, I was someone easily distracted by shiny new things. If left to my own devices, my nature was to only see the surface-level glossy stuff. So I told myself: no, don't do that. Pick one or two areas and go deep. Don't let yourself look at anything else.

I chose (1) team collaboration tools — I wasn't sure I was genuinely interested, but I'd been exposed to the space since my time at Dropbox. And (2) security, because it seemed interesting and I wanted to learn about it. Both had massive market potential, and I had interest in both. Plus I'd experienced all kinds of security and compliance challenges at Dropbox. I felt I'd enjoy learning about security, and I wanted to work with startups.

For collaboration, we asked ourselves: 'What are the macro trends and new technologies shaping the world?' It was late 2016, so the answer was 'voice!' Team collaboration software was rising, so our answer was B2B Alexa. It made perfect sense on paper and was completely meaningless in reality. We realized that if an idea only comes from desk research, someone probably already tried it. There's no money lying on the sidewalk.

In that same context, if we'd pivoted from B2B Alexa to something like 'let's build a better wiki' or 'solve security issues,' it would have scattered our focus and prevented us from going deep on something others hadn't.

Once we pivoted to security, I started asking every startup I knew about their security thinking. Every single one answered with embarrassment, admitting they'd done nothing in security despite knowing they should. Everyone wanted to do security right, but it was hard to prioritize before customers demanded it. So I rolled up my sleeves and did piecemeal security work for them — and that's how Vanta started.

When I tell this to other founders, sometimes they ask: how did you know security would work? I didn't. If you can foresee an area succeeding, just go do it. But that kind of 'foresight' usually only appears in retrospect."

— Christina Cacioppo, founder and CEO, Vanta

Vanta is a SaaS company that provides security compliance solutions for tech companies, helping them streamline and improve compliance processes.

Notion's original idea was a no-code app builder — it took four years before the team identified their core product.

"Early Notion was a no-code website builder. You can still find videos of it on YouTube. For years, nothing worked. It was unstable, our early tech stack was problematic, user feedback was poor. When we dug into conversations with users and used the product ourselves, we realized the users who stuck around loved the editor and collaboration features. That made us realize we should focus on docs and wikis.

The turning point came in early 2017, four years after launch, when we shipped Notion 1.0 focused on docs and wikis. We also built templates that users could deploy with one click. The product shifted from letting users build software through our product to one-click creation of modern docs and wikis. That proved to be a strong market entry point, because notes are the simplest atomic unit of work — easy to attract users. Once users got serious about it, some would dig into Notion's underlying structure and discover they could modify and create their own templates. We took some detours to get here. That's the Notion story."

— Akshay Kothari, co-founder and COO, Notion

Zip co-founders Rujul Zaparde and Lu Cheng experimented across many domains, looking for a market that was large, poorly served, and genuinely interesting to them. They eventually landed on procurement (the company is now valued at over $1 billion).

"This was our sixth or seventh idea. Earlier concepts included letting people in India (and elsewhere) buy fractional shares of US stocks, and an accounting services marketplace (which generated $200K in revenue).

Later, when searching for our next direction, our Y Combinator partners suggested we find a large market where incumbents were mediocre. That led us to procurement.

We'd experienced procurement's shortcomings at Airbnb. In retrospect, we'd both indirectly encountered procurement problems, but only realized it later."

— Rujul Zaparde and Lu Cheng, co-founders, Zip

Zip is a company that provides procurement workflow services for businesses. Employees can use Zip to initiate procurement or vendor requests, which are then automatically routed through the correct approval chains across procurement, finance, IT, data security, legal, and other relevant teams.

For Databricks, Ali Ghodsi spotted an undervalued emerging technology and kept doing user interviews until he was convinced he could build a company around it.

"We'd been doing research at UC Berkeley for several years and started noticing changes at Silicon Valley tech companies. Their approach to data scalability was different from what we saw in other industries. At the time, other industries were using Hadoop — for them, that was the best solution. But we saw tech companies in Silicon Valley using a different technology stack (what became Apache Spark), and using data to train AI and do machine learning with remarkable results. We realized this new technology could drive massive transformation in other industries.

Our initial thought was: 'We could open-source these new technologies we see, publish our research, and companies worldwide would adopt them — that way we change the world?' We didn't set out to start a company. But over the next few years, we discovered and open-sourced software that didn't follow the path we expected. It was remarkable software, and we were frustrated. People would say it's just academic software — how do we know it's reliable, can it work in enterprises? By 2012, we were frustrated enough to think: if we raised some money ourselves to support it, maybe we could make it happen. That's the Databricks founding story."

— Ali Ghodsi, co-founder and CEO, Databricks

Databricks was founded by members of AMPLab who worked on Apache Spark. It provides data services for enterprises.

Reflections

1. What trends are emerging but poorly served (e.g., security, collaboration)?
2. What commercially viable technologies are rising but undervalued or underdeveloped (e.g., data scalability)?
3. How many potential customers have you discussed your startup idea with?
4. How important does your potential customer consider the problem you're solving, on a scale of 1 to 10?
5. How low is service quality in the current market, on a scale of 1 to 10? What's your solution's score? You need at least a 4-point gap to have a real shot.

Follow Source Code Capital for more.