How the most successful B2B startups came up with their original idea
Part one of my seven-part series on kickstarting and scaling a B2B business
đ Hey, Lenny here! Welcome to this monthâs â¨Â free edition ⨠of Lennyâs Newsletter. Each week I tackle reader questions about building product, driving growth, and accelerating your career.
If youâre not a subscriber, hereâs what you missed this month:
Subscribe to get access to these posts, and every post.
My paternity leave has come to an end, and Iâm kicking things up a notch. Over the next two months, Iâm going to share a multi-party in-depth playbook for kickstarting and scaling a B2B business. This series took hundreds of hours of work and is based on dozens of 1:1 interviews with the founders of two dozen of todayâs most successful B2B companiesâincluding Gong, Notion, Figma, Amplitude, Retool, Canva, and many more. In addition to these juggernauts, Iâve also pulled in a handful of my favorite up-and-coming B2B startups, including Linear, Vanta, Stytch, Zip, Census, Persona, and Hex.
This series builds on my two previous playbooksâfor consumer businesses and marketplace businessesâand brings me one step closer to my goal of giving every founder a tactical guide for turning their idea into reality. Even if youâre not building a B2B startup right now, youâll find inspiration, new ideas, and frameworks that will help you with the product youâre building right now. Here is whatâs in store:
Part 1: How to come up with a great B2B startup idea â This post
Part 2: How to validate your idea
Part 3: How to identify your ICP
Part 5: How to find product-market fit
Part 7: How to scale your growth engine
Youâll find never-before-shared stories, surprising lessons, and, as usualâa ton of tactical and actionable advice you can use to grow your own product today.
Hereâs a peek at a few of the more surprising takeaways:
The majority of founders had no special skill or background in the problem space they went after.
Most B2B startup ideas did not come from the founder feeling the pain at their last gig (though many did).
Every prosumer product (e.g. Notion, Figma, Airtable, Miro, Slack, Coda) took two to four years of wandering in the dark before they found something that worked.
Founders spoke to a median of 30 potential customers to validate their idea before committing.
~40% of startups pivoted at least once before landing on their winning ideaâoftentimes more than once.
About 20% were solo founders.
Cold outbound worksâitâs the second most common way to get your early customers.
This series should probably be a book, but instead Iâm sharing it with you all here. Iâm excited to hear what you think and evolve this work further. And remember: following these steps (or any steps!) wonât guarantee success. But itâll certainly improve your odds. Letâs get started with step one: coming up with a great idea.
A huge thank-you to Akshay Kothari (COO of Notion), Ali Ghodsi (CEO of Databricks), Barry McCardel (CEO of Hex), Boris Jabes (CEO of Census), Calvin French-Owen (co-founder of Segment), Cameron Adams (co-founder and CPO of Canva), Christina Cacioppo (CEO of Vanta), David Hsu (CEO of Retool), Eilon Reshef (CPO of Gong), Eric Glyman (CEO of Ramp), Guy Podjarny (CEO of Snyk), Jori Lallo (co-founder of Linear), Julianna Lamb and Reed McGinley-Stempel (co-founders of Stytch), Mathilde Collin (CEO of Front), Rick Song (CEO of Persona), Rujul Zaparde and Lu Cheng (co-founders of Zip), Ryan Glasgow (CEO of Sprig), Shahed Khan (co-founder of Loom), Shishir Mehrotra (CEO of Coda), Sho Kuwamoto (VP of Product of Figma), Spenser Skates (co-founder and CEO of Amplitude), and Tomer London (co-founder and CPO of Gusto) for contributing to this series. Art by Natalie Harney.
What makes a great B2B startup idea
If you boil it down, there are essentially three elements to a great B2B startup idea:
Ingredient 1: The problem is important to people
There need to be a lot of people willing to spend a lot of money to solve the problem. Most startups fail not because the idea isnât good but because the market for the solution is just too small. If you want to build a venture-scale business, a rule of thumb is that there needs to be a clear path to $100m in revenue per year, and eventually a path to $1B a year. Hereâs a great example of what you want to see, as told by Ryan Glasgow, CEO of Sprig:
âRobinhood agreed to a large contract even though we were early. Theyâre like, âWe love it. Weâll install tomorrow.â They installed it as we were building the first version.â
Hunter Walk has a great framework called LUV. Also, hereâs my take on what makes a venture-scale idea, and a guide to help you think through your market size.
Ingredient 2: The market is underserved
Existing solutions need to be doing a subpar job of solving the problem. The founder of CRED, Kunal Shah, has a great framework for evaluating potential new solutions, called Delta-4:
What would a customer rate the existing solution on a scale of 1 to 10?
What would they rate your solution?
Your solution needs to be +4 better for anyone to care about your product.
This also means that if the existing solution gets a 6 or higher, itâll be very hard to replace (e.g. Excel). Hereâs what youâd love to see, as shared by Tomer London, co-founder of Gusto:
âAs soon as we asked them the simple question of how they feel about their current payroll provider, they started cursing. More than half of the people we talked to just started cursing, unprompted. Two people voluntarily told me, âI use [competitor name], and my password is fuck[competitor name].ââ
Ingredient 3: Youâre excited about solving this problem
Finally, the problem needs to be something you are willing to spend many years of your life solving. This advice from Dylan Field, co-founder of Figma, captures the sentiment perfectly:
âThere are actually a lot of good ideas out there. Itâs kind of the weird part, especially if youâre searching for one; it feels like itâs not the case, but there are so many different markets that are underserved. The more important thing, actually, is to find something that you are personally passionate about, because any good company takes a long time to build.
If you are, letâs say, three to four years in on an idea that you hate, youâre just going to burn out and youâre going to quit. It wonât feel good and youâll be hating life. Donât just go for an idea because itâs kind of working. Go for an idea that you really care about, because even if it doesnât work, youâll still learn from it and youâll still have one.â âDylan Field, co-founder and CEO of Figma, via Elad Gil
Takeaway: Find an idea that solves a problem that is important, underserved, and that youâre excited about. Youâll hear a lot more about how to do this, and what this looks and feels like, below.
âItâs essential to work on something youâre deeply interested in. Interest will drive you to work harder than mere diligence ever could. When in doubt, optimize for interestingness.â
âPaul Graham
Finding a great B2B startup idea
Across all of the interviews I did, I found three reliable paths for discovering a great idea:
Past pain: Identify a large pain you experienced at a previous companyâthen build a solution.
Ponder and probe: Pick a space youâre interested in, then whiteboard and tinker while talking to dozens of potential customersâlooking intently for pain and pull.
Present pull: Identify something youâve built that is showing signs of pullâand pivot fully to that.
Hereâs an overview of how todayâs biggest B2B companies found their ideas:
No matter which path you take, you are looking for two things: pain and pull. Pain tells you thereâs an opportunity to solve a problem, and that itâs important. Pull tells you that youâre actually solving the problem.
Letâs explore each of the three paths, illustrated by founder stories.
Path 1: Past painâIdentify a large pain you experienced at a previous company. Then build a solution.
When I first started this research, I was expecting this path to be the way that all great startups found their ideas. Surprisingly, it only accounts for about 40% of great ideas.
Letâs consider five successful startups that found their big idea following the path of past pain: Gong, Retool, Linear, Persona, and Hex.
The CEO of Gong experienced the pain of being unable to understand what was going wrong within the sales process at his previous company:
âAmit, now my co-founder, ran a company in the BI space called Sisense. He ran into a problemâwhen sales didnât work wellâand it was very hard to understand why. He realized that essentially all of the knowledge about what was working and not working was hidden inside peopleâs heads: âThe CRM was showing me stuff, but it wasnât anything meaningful. Yes, you didnât close that deal, but why?â
He left Sisense and looked around to see if anybody was addressing this problem. I was on sabbatical doing nothing, learning deep learning, so we figured: letâs try to come up with a system that takes the stuff from salespeopleâs heads, captures the information, and gives visibility and guidance to the rest of the organization.â
âEilon Reshef, co-founder and CPO
David Hsu, the founder of Retool, found himself building the same product over and over at each company he worked at:
âPerhaps a little-known fact is that I had actually started a couple of other companies and products before Retool. And every time the team built any of these products, we had to go build our own tools for it. Eventually, when you build enough of the same thingsâin this instance, internal tools for different projectsâyou kind of realize they actually all have the same building blocks. We thought, âThereâs got to be a better way of doing this.â And weâre lazy engineers and we thought other engineers would be lazy as well. So thatâs how the idea [for Retool] first came about.â
âDavid Hsu, founder and CEO
The founders of Linear felt deeply disappointed every time their companies defaulted to Jira:
âLinear emerged out of the pain we all felt with issue tracking project management, when it comes to building software. I was at the time at Coinbase, Karri [Saarinen, co-founder] was at Airbnb, Tuomas [Artman, co-founder] was at Uber. We were all at these growth-stage companies and saw how they both grow and operated as they grew.
I very vividly remember the time around 2016 at Coinbase, when we were closing in on 100 people and we hired a VP of engineering. There was a desire to start unifying the teams using a single tool. And Jira started coming up in the conversations.
I was maybe a little bit naive and tried to keep my team using Trello, but I very quickly realized that it is a losing battle. Youâve just got to go with it and thereâs no other solution, basically, on the market. And so we ended up going with Jira. But it just stayed with me as this nagging feeling of disappointment.
Separately, Karri at Airbnb had created a Chrome extension to re-skin their Jira, and Tuomas was building developer tools at Uber, using Fabricator. We all had similar experiences of âThis doesnât feel right and we should do something about it.ââ
âJori Lallo, co-founder
Personaâs founder noticed how inadequate identity tooling was when having to build this technology at Square:
âI worked on identity over at Square for a fair bit of time. We actually used to joke as a collective team over on Square, âIâm sure in 10 years there will be someone who does this well for us.â Importantly, while we were building it, we noticed that there was this disconnect between existing identity vendors thinking that there was going to be one technique thatâll solve all of a customersâ identity problems, while in reality, every business had a multi-modal approach: a bunch of different techniques to verify different people in different contexts. So we saw an opportunity to go after this problem in a unique way.â
âRick Song, co-founder and CEO
Hexâs founder realized that no one was going to build what he needed, so he had to do it himself:
âI actually started this journey as a buyer. I was looking for something like Hex and I couldnât find it. It took me a few months of going around and talking and asking people like, âOh, what are you using for this? What are you using for that? Have you found anything good for this?â Everyone said, âNo, but if you do, let us know.â It kind of clicked one day. I was like, âWell, maybe we should do this.â
I didnât set out to be a founder. There are a lot of people out there who are like, âI want to be a founder. I just need an idea and a co-founder.â I was almost the opposite. The idea was very obvious to me. I had two great co-founders that I had met at Palantir whom I loved working with. It almost took us a while to be like, âAll right. I guess we have to do this. Donât we? No one else is doing this, while we have to do it.ââ
âBarry McCardel, co-founder of Hex
Takeaway questions to reflect on:
What did you or others build at previous jobs that proved to be incredibly valuable to the company, or your teammates?
Whatâs a tool you wish you had (and would have paid a lot of money for) at your previous companies?
What internal products have you built again and again at past companies?
Path 2: Ponder and probeâPick a space youâre interested in, then whiteboard and tinker, while talking to many potential customers looking for pain and pull.
A surprisingly large number of founders came up with their startup idea by doing something thatâs often discouraged: sitting around and whiteboarding. As youâll see below, thereâs a wrong way of whiteboarding and a right way (i.e. ideating and building in a silo versus quickly talking to potential users to validate your idea). But most importantly, if you want to find a great idea, again, focus on finding a problem space with three traits: (1) itâs important, (2) itâs underserved, and (3) youâre excited about it. Ideally your idea is also rooted in some experience youâve had in the space, but it doesnât have to be.
Figmaâs founders narrowed their focus to two areas they saw potential in and then started tinkering with the one they were most excited aboutâdesign:
âAfter we got the Thiel Fellowship, I called Evan [Wallace, co-founder] and said, âHey, you know, weâve got actual funding now: $100,000 over two years. Do you want to go do this thing for real?â
When youâre starting a company, itâs really useful to ask the question: why now? Itâs a really useful framework. It can be societal; maybe thereâs some new cultural trend. Perhaps itâs regulatory; some law has been passed or repealed. I like the technological version. For us, we saw drones in 2012 and WebGL as a few technologies that were happening. Because of them, new possibilities were suddenly there.
On the drone side, we didnât get very far. Evan in particular was not that interested in doing hardware. Heâs like, hardware sucks, run debug cycles are really long, and then we focused on WebGL instead.
We were also excited about computational photography for a bit, but quickly realized that with photos, that ends up leading to a consumer sort of application. And the entire purpose of WebGL was to be in the browser. Why would you do anything in photo editing if youâre not on the phone? And so we felt like we were kind of building in the wrong place and then eventually sort of shifted our attention to design. I had been a design intern at Flipboard, and that kind of helped me realize what would be possible there.â
âDylan Field, co-founder and CEO, via Elad Gil
Vantaâs Christina Cacioppo similarly narrowed down her thinking to two spaces that felt ripe for opportunity that she was excited about (security and collaboration), and kept talking to customers to find the biggest source of pain and pull:
âOur first couple of ideas were just total garbage. There was a lot of whiteboarding. I think the failure mode in this part of the process with the whiteboard is that itâs just all so abstract. There were no customers or users.Â
Back then, I tended to be a shiny-object person, so left to my own devices, my natural inclination was just to go find all the shiny objects on the surface. So I was, okay, no, donât do that. Pick one or two spaces and go really deep. Youâre not allowed to look at anything else. Not allowed.
I chose (1) team collaboration tools, because I thought, I donât know if I truly love this but I know it from my Dropbox days, and (2) security, because it seemed interesting and I wanted to learn it. They both seemed big and important, and interesting, and I had all kinds of security and compliance challenges at Dropbox. I also felt like Iâd be happy learning about that space, and I also wanted to work with startups.
Within collaboration, we were asking ourselves, âWhat are the macro trends and new technologies shaping the world?â Itâs late 2016, and so . . . voice! Team collaboration software is also emerging, and so the answer is B2B Alexa. At the whiteboard stage, it makes so much sense. And in reality, zero sense. We recognized that if you can do it on a whiteboard, someone has probably done it. Thereâs no $20 bills on the sidewalk.
Also, when youâre changing context from B2B Alexa to other ideas like, Iâm going to make a better wiki, to solving security, to whatever, youâre spreading yourself too thin and youâre not going deep enough to actually learn anything that someone else hasnât.
So within the security bucket, I started asking all the startups I knew about security, and they all looked at me really guiltily and said, âWe donât really do anything. We know we should, but we donât.â Everyone wanted to do security, but it was hard to prioritize until customers asked for it. So I started doing it for them scrappily and manually, and that led to what Vanta is today.
When I tell founders this, sometimes theyâre like, âWell, how did you know security was going to work?â You donât. If you know the answer to that, go for it. But itâs only obvious if you look back.â
âChristina Cacioppo, founder and CEO
Notion started with the general idea of a no-code app builder and eventually (four years later!) noticed pull for a specific set of features:
âEarly Notion was basically a no-code website builder. Some of the videos actually exist in YouTube. It didnât go anywhere for years. It was very unstable, the tech stack we had used in the early iterations of it was fairly buggy, and user feedback was not great. But as we peeled the onion, talking to users and using it ourselves, we realized that for the people that continued to use it, they liked some elements of the editor and the collaborative features that the product had. And through that, we realized that we should double down on docs and wiki.
The turning point was in early 2017 (four years in), when we launched Notion 1.0, focused around just docs and wikis. At the same time, we made a few different templates ourselves so that you can click a button and start using the product. Instead of telling people to build software using our tool, you can now click a button and get this modern docs/wiki product, without any work. It proved to be a great wedge into the market, because notes is the simplest unit of work, and it got people in the door and using the product. And then once people actually started using it, some people went down to the lower stack and they were like, âHuh, I can actually change these things. I can modify it, I can build my own template.â We took a little bit of a roundabout way of getting there, but thatâs the story.â
âAkshay Kothari, co-founder and COO
For Zip, Rujul Zaparde and Lu Cheng went through a variety of ideas, looking for a big and underserved market they were excited about. They eventually focused on procurement (and their company is currently valued at over $1B):
âThis was actually our sixth or seventh idea. Some earlier ideas included allowing people in India (and then other countries) to be able to buy U.S. equities in a fractional way, and an accounting staffing platform (which generated $200k of revenue!).
Eventually, as we were thinking through our next pivot, the advice we got from our Y Combinator partner was to look for a market that was large and had entrenched companies that werenât great. This led us to the procurement space.
And then we thought back to this problem at Airbnb. In hindsight, we both adjacently ran into it a couple of times, but we didnât realize it until later.â
âRujul Zaparde and Lu Cheng, co-founders
For Databricks, Ali Ghodsi noticed a new technology emerging that was underutilized and kept talking to users until he realized how to turn it into a real company:
âWe were doing research for a bunch of years at U.C. Berkeley and had started seeing what was happening in Silicon Valley tech companies, how they were solving data scalability problems, and how different their approach was from what we were seeing in the rest of the industry. At that time, the rest of the industry was leveraging this technology called Hadoop, and that was the best thing since apple pie for them. But we were seeing that the Silicon Valley tech companies were using a different technology (what is now Apache Spark) and were in fact doing AI and machine learning on data, and they were getting magical results. We knew this technology could be game-changing for the rest of the industry.
Our initial thought was, âCan we just take what weâre seeing over there, open source it, and publish the research, and then the whole world will adopt it and weâve changed the world?â We didnât really intend to start a company. But then over the years, it was frustrating to see that we had found this software, and we had open-sourced it, and we thought it was so amazing, but it wasnât getting the uptake that we thought it deserved. And people kept telling us, âThis is some academic project. How do we know that we can really rely on it? Is it really enterprise-ready?â And so on. There was frustration around 2012 where we thought, if we start something, if we ourselves get behind it and we raise some money, maybe we can really actually ourselves make it happen. So that was the journey.â
âAli Ghodsi, co-founder and CEO
Takeaway questions to reflect on:
Whatâs a trend thatâs emerging but underserved (e.g. security, collaboration)?
Whatâs a transformative technology thatâs emerging thatâs underutilized (e.g. data scalability)?
How many potential customers have you spoken with about your idea?
How important to potential customers is the problem youâre exploring, on a scale of 1 to 10?
How underserved is the current market, on a scale of 1 to 10? And where would your solution rank? Look for a delta of at least 4.
Path 3: Present pullâIdentify something youâve built that is showing signs of pull. Pivot fully to that.
A final path is simply to keep building what youâre building, but pay special attention to features or functionality that are showing strong pull from existing users. This may lead to shifting your approach slightly, or it may lead to a completely different product.
In the case of Amplitude, the big idea came from noticing other founders wanting access to their internal analytics tool:
âWe started with a different company before Amplitude, called Sonalight, which was a voice recognition application that allowed you to send and receive text messages by talking to your phone. One of the things that was really clear to us at the time was that you should look at what people are doing in your product in order to figure out how to make it better. Figure out where people are getting stuck, what keeps them coming back, what they like to use, and what they donât like to use.
We tried out tons of products on the market. I remember Flurry, Google Analytics, Adobe, Kissmetrics, and others. And none of them were able to answer the questions I just posed. And so we said, âAll right, well, weâve got to build it ourselves.â A bunch of engineers with a lot of hubris, haha. So we ended up doing that. And as we shared those insights with other companies that we knew (we were at Y Combinator at the time), they were like, âWow, thatâs amazing. I really need to understand that about my business.ââ
âSpenser Skates, co-founder and CEO
Sprig followed a very similar storyâRyan Glasgow was building one product, but then noticed a side project seeing much more pull:
âââI was exploring a startup idea, but I wasnât quite getting the pull I expected. I went back to what I knew, which is product management, and I started asking users tough questions, like âWould you pay for this?â and âWhy do you not use this?â In this discovery process, I wanted to add more rigor, knowing that this was a high-stakes decision (e.g. Iâd be spending 10-plus years of my life on it), and I found a technique called outcome-driven innovation, which adds rigor to product innovation. I was like, âWow, this is mind-blowing. What if I could build a tool to help other people go through the idea maze and not build something before they de-risk the idea?â
So I built a little SDK that people could plug into their apps, with a very simple front end, that would just rotate through three questions about functionality, usability, and quality (the core questions in the framework I mentioned). It used Airtable as the back end, actually; it was so simple.
I shared this with other founders that I knew who were also trying to get their business off the ground, and gave them this spreadsheet of quality, usability, functionality, using this framework. And they were like, âWow, Iâm learning so much. Iâd be your first customer if this was a product.â It immediately clicked. They would say, âIâll pay you for this.â âThis was the market pull I was looking for, and I quickly pivoted to surveys.â
âRyan Glasgow, co-founder of Sprig
The Loom founders went through two completely different products before seeing obvious pull within one unexpected feature of their app:
âLoom was the result of two previously failed attempts in the video space. Our first was a marketplace for companies to hire subject-matter experts (i.e. product managers, designers, engineers, etc.) for feedback on checkout flows, UX, sales funnel, onboarding, and more. While we saw initial revenue here, this failed to get the traction to warrant additional build time (we were already four months in at this point). The second was a SaaS tool that any company can embed on their website and get instant feedback from their users in the form of a user test. Similar to the first iteration, this failed to get any real momentum after three months, although it led us to a remarkable insight from one of our few design partners.
Three months into this second idea, we had an aha moment when a client used the product to record a video of himself summarizing all of the user tests his team had collected. Thatâs when it occurred to us that there could be something here.
A month later, we launched on Product Hunt and had thousands of people whoâd downloaded the extension by dayâs end. That made it clear to us that we should double down on this new direction, and weâve never looked back.â
âShahed Khan, co-founder
With Segment, nothing was working for a year and a half, and while throwing ideas against the wall, one idea finally stuck:
âWe originally applied to YC with the idea for a university classroom lecture tool (we were college students at the time), and we built that out over the course of the summer. It failed because we arenât good at identifying real problems, and we decided that one reason is that we didnât understand our users.
We spent about 15 months pivoting and building various analytics products to help better understand users. None of them really took off, and we keep chasing the next idea that will get us users. As a growth hack, my co-founder Ilya built a little library called Analytics.js. The idea was that users can use it as a âdrop-in replacementâ for Mixpanel or Kissmetrics and send us the exact same data as you would to each of them via one library. Users didnât care much about our tool, but they seemed to like the idea of Analytics.js.
Finally, we have about six months of runway left and we havenât launched anything. My co-founder Ian [Storm Taylor] thinks that the idea behind Analytics.js could actually be a big deal. We could effectively be the âAPI layerâ over all these annoying and similar, but inconsistent, data tools.
Weâre split on the decision. Peter thinks itâs the worst idea heâs ever heard and that thereâs zero chance a company could be made from 100 lines of JavaScript. We all agreed to build it for a week and launch it on Hacker News. That day, it goes straight to the top of HN, and the rest is history.â
âCalvin French-Owen, co-founder of Segment
And famously, for Slack, the big idea came from noticing one feature being incredibly popular within their ill-fated video game:
âWe came to the conclusion that Glitch [the game he was working on] was never going to be the kind of business that would have justified the $17.2 million in venture capital investment [that we raised]. It might have been a neat project for a half-dozen people if we had spent a million dollars to get there, but by the end of 2012, there were 45 people working on it, we had spent many millions of dollars, and it just wasnât ever going to scale. So we decided to shut it down without knowing what we were going to do next.
The company still had millions of dollars left in the bank, and one of the possibilities was to return that money to the shareholders and call it a day, maybe start something else up. But because we had that money, we had the flexibility to shut it down in what we felt like was a humane way. We spent a long time working on reference letters for the employees who built the website, essentially to make sure that they all got jobs.
We also gave users the choice to get a refund for everything they had spent, to let us keep the money, or donate it to charity. We kind of had our hands full for a little while in shutting the game down, and while we did that, we were thinking about what we might want to do next. There were all kinds of ideas.
It took us a little while to settle on the idea that would become Slack.
That was really born out of the style of communication that developed while we were working on the game. We used an older technology called IRC, and because IRC is very limited, over the years we added the little features here and there that we wanted.
For example, in IRC, if youâre not online at the same time as me, I canât send you a message. I have to wait until you are also connected. So one of the first things we did was build a way of archiving messages so that you could catch up when you came back online, and in those archives we wanted to be able to search them, so we added search, and so on. There was no good iPhone client, so we made an HTML5 front end for our archive viewer.
This interesting dynamic happened. By the time we shut down the game, again there were 45 people at the company, we had been in operation for three and a half years, and we had a companywide email list. After more than three years, it only had 50 messages on it, so about one every three weeks. That wasnât a deliberate decision, that wasnât ideologically driven. But it just happened that everyone paid attention to IRC, and the more people paid attention to it, the more information we routed to it; and the more information we routed to it, the more people paid attention to it. So eventually, everything from database alerts to daily sales figures were being pumped into IRC. Every time someone uploaded a file to the file server, that would be posted into IRC. While we werenât successful in making the game, we were very efficient in being unsuccessful to make the game.â
Takeaway questions to reflect on:
Of all the things youâve built, what one feature is showing the most pull?
What problem are you currently facing with your startup that you wish someone would solve? Explore solving it yourself.
Whatâs a side project youâve been wanting to build that you havenât?
To summarize, if youâre trying to come up with a great B2B startup idea:
Look for ideas that are (1) important, (2) underserved, and (3) youâre excited to solve.
Pick one of these routes and ask yourself these questions:
Past pain
What did you or others build at previous jobs that proved to be incredibly valuable to the company or your teammates?
Whatâs a tool you wish you had (and would have paid a lot of money for) at your previous companies?
What internal products have you built again and again at past companies?
Ponder and probe
Whatâs a trend thatâs emerging but underserved (e.g. security, collaboration)?
Whatâs a transformative technology thatâs emerging thatâs underutilized (e.g. data scalability)?
How many potential customers have you spoken with about your idea?
How important to potential customers is the problem youâre exploring, on a scale of 1 to 10?
How underserved is the current market, on a scale of 1 to 10? And where would your solution rank? Look for a delta of at least 4.
Present pull
Of all the things youâve built, what one feature is showing the most pull?
What problem are you currently facing with your startup that you wish someone would solve? Explore solving it yourself.
Whatâs a side project youâve been wanting to build that you havenât?
You know youâre on to something when you notice pull and pain.
Now that you have a framework for coming up with an idea, the next step is to figure out if the idea is worth going all-in on.
Next: How to validate your idea
Have a fulfilling and productive week đ
Appendix: Bonus origin stories not included in above (for brevity) but also incredibly interesting
Past pain
Snykâs Guy Podjarny noticed a couple of emerging trends and pain points in two previous jobs and began tinkering with the idea:
âBetween 2002 and 2010, I helped build and manage some of the very first Application Security (AppSec) products, which customers used to find and fix AppSec vulnerabilities. Since it was very expensive to fix flaws you found at the end of a year-long release cycle, we tried to get security to âshift leftââto find issues during development where itâs cheaper to fix them.
I left that path to found my first startup, Blaze, building a real-time web page compiler that made websites faster. Blaze was acquired by Akamai, where I became the CTO of the web performance business. DevOps played a key role in that journey, and the perspective helped me appreciate the impact it will have on the industryâand the opportunities it brings.
These two lenses made me realize two things:
At the pace of DevOps, getting developers to embrace security has gone from an efficiency and costs play to a necessity. Itâs the only way for software to be secure.
If you build the right solution, developers can and will embrace security. DevOps had proven that by getting developers to embrace ops, with tools like New Relic, Heroku, and others paving the way.
So after leaving Akamai, I set out to build a developer tooling company that tackled security.â
âGuy Podjarny, founder and CEO
All three of Canvaâs founders noticed a gap in the design space:
âMel [Melanie Perkins, co-founder and CEO] was teaching design through university, including how to use different types of design software, and through that experience, she realized that it was quite complicated and really hard for people with no design background to start approaching these tools and actually get something useful out of them.
And Iâd created a lot of creative tools over the previous decade, which I found unlocked peopleâs creativity in a whole bunch of different ways, like with music and drawing, and other creative tools. Iâd worked at Google on a product called Google Wave, which was all about communication and collaboration, and again, being able to unlock peopleâs creativity through new technology and new experiences.
When I met Mel and Cliff [Obrecht, co-founder and COO] in 2012, the idea of unlocking design and creativity really captured all of our imaginations, and so we decided to band together in July 2012 and start building that first product. We eventually launched Canva in August 2013, about a year later.â
âCameron Adams, co-founder and CPO
Same story for Stytch:
âWe had a monthly coffee chat on our calendar after Julianna [Lamb, co-founder] left Plaid to go to Very Good Security. We were always just catching up and often complaining about work, what we were banging our heads against the wall on. And that week, we were both banging our head against the wall on authentication. It was the number one thing I worked on at Plaid. And so I was particularly frustrated. And then Julianna had just finished a project to rip out Auth0. A project had gone on for six months before she joined, and then it was taking so long that theyâd stopped doing it and then later picked it up again. It was like a meme within the company because of how annoying a project it was and how much time it was taking.
Julianna was talking about how frustrating it was internally there, I was talking about how we just had a really bad experience with Auth0 and AWS Cognito at Plaid. And we were just surprised that everything assumed youâre using a password in terms of how the APIs were designed, but also, you didnât have much flexibility.
And so that was the first time we were both complaining about the same thing, which led it to being the first company idea where I think what we said was: we just presumed there wouldâve been a Stripe for authentication in terms of how easy it was to work with an API, to do this rip and replace Julianna was doing or add the password-less features that we were trying to add in the Plaid.
So December 2019 was the coffee. We had left our jobs at Plaid and Very Good Security at end of April, early May, I want to say, to go full-time on this. And we raised our seed round in June of 2020.â
âReed McGinley-Stempel, co-founder
The founders of Snowflake noticed looming challenges with companies relying on Hadoop:
âIn 2012, Thierry Cruanes and I were working at Oracle, and there was not a single day without discussions on how Hadoop [a developer of open-source computer software] was about to make [our work] completely obsolete. I even remember one day interviewing a young engineer who [was] so excited about Hadoop that he didnât listen to anything we said about Oracle. We realized then that we were missing something.
This made us think a lot about the future [of what we were doing]. Big data was taking over the world, and data warehouses as we knew them were having a really hard time competing. They wereâand often still areârigid, expensive, and difficult to use. At the same time, we were convinced that Hadoop was not a good solution either. Hadoop systems were really too hard to use for mostâvery inefficient, slow, and missing key features that I would consider must-haves.
This is the point we realized we could [build] a new type of data warehouse system . . . that would master the elasticity of the cloud. [What we built ran] 10 times faster than any other system for the same cost [and freed] users from any management tasks.â
âBenoit Dageville, co-founder, via Yahoo Finance
Rampâs founders found that credit card companies were solving the wrong problem, through working at Capital One and their previous startup:
âRamp is a continuation of a lot of things weâve been working on for a while. I had a company before this called Paribus, which helped people save money on things they were buying online. Within the year, we had about a million customers. Then we got what was a life-changing offer from Capital One to buy the company and went for it. So that was about 2016.
We then spent the next two and a half years there scaling the company. We learned a lot more about how to turn data into savings at scale, and about all the ways that you can use fragments of data processes to automate processes, save people money, and save people time in various ways.
We also ended up in the credit card division within Capital One, where we learned about how credit cards work, how rewards work, and all that. I just remember we talked to customers and asked them, âDid you want points, cash back, something different for rewards?â And what jumped out is people didnât really want any of that. They wanted to be better-off. And we were just struck by this idea of rewards versus not spending that dollar in that first place.
Pretty quickly we got obsessed with an idea: what if your card was smart enough to help you spend less? It actually might be better for the customer. You might be able to have a better shot at breaking out of the price-based competition.
And fintech was becoming a thing then, where you could actually start these businesses relatively cheaply.
There was only so much we could do inside of Capital One, and we helped what was Capital One Shopping become a thing and grow, and now itâs a big part of what they do, but we also concluded we couldnât quite do what we were hoping to do inside of the company. So we left and then set out to build Ramp. Thatâs the story.â
âEric Glyman, co-founder and CEO
Ponder and probe
Codaâs Shishir Mehrotra had been thinking about this idea for years:
âCoda started as a series of brainstorms. And at the time, I was reasonably happy at Google. I wasnât going anywhere. But I had this long list of ideas of just stuff Iâd like to see different in the world. This idea is one thatâs been on my list for 20-plus years. And in some way or the other, Iâve actually worked on it three different times.
I had a good friend, Alex [DeNeui, eventual Coda co-founder], who was busy starting a company and had decided to pivot to a new idea. He asked me for help brainstorming what he should work on, and we ended up talking about the core ideas behind Codaâthat the line between docs and apps is artificial, and that it was time for a new blinking cursor. And in that process, I ended up basically falling in love with the idea again myself. Despite all reasonable indicators of why you shouldnât start a company, I just couldnât stop thinking about it. And so I jumped in, and it was kind of funny, when I was talking to my wife: we were driving up to Napa for this wedding, and weâre in the car, and I was like, âI think Iâm going to do this company with Alex.â And she says, âYeah, I already knew that.â And I said, âHow?â And sheâs like, âBecause you took every free moment and youâre spending all that time with him on this idea.ââ
âShishir Mehrotra, co-founder and CEO
Gustoâs three founders had secondhand experience with the challenges of running a small business and after ideating, focused on payroll as a great place to start:
âMy parents have a clothing store. Itâs a small business, and I just grew up there in the store every day after school, going and organizing and selling and answering the phone calls and helping in the store. So I have this kind of small-business experience throughout my family. My grandfather has another store down the street. My aunt had another store. All in the same street in Haifa. So I really grew up around these small businesses, and one thing thatâs really, really, really clear is that they donât have the tools that big companies have in order to be successful. If you think about it, they are the ones who actually need the tools more than anyone else.
So this idea of, what can you do to help this underserved group who need the help and what can you do to level the playing field for small business, is something that was always top of mind for me. Josh [Reeves], Eddie [Kim], and I [co-founders] all share a similar story with small businesses in our families. So when we got together, we wanted to build something.
We kicked around a bunch of ideas for a couple of months. We had this massive whiteboard, and we kind of plotted a lot of different problems that we felt ourselves personally in our lives in different moments and different things, but very quickly we got focused on small businesses and then started talking with people.Â
We learned that a third of companies make payroll mistakes every year, and they get fines for it every year. Thatâs insane. The system is incredibly broken, so thatâs where the idea came from.â
âTomer London, co-founder and CPO
For Census, Boris Jabes noticed a big pain point at his previous job and couldnât stop thinking about it:
âWhere the idea came from (eventually): I sold my previous startup (Meldium) to a company called LogMeIn, and I ended up working with a bunch of sales and marketing people there. Working with them was a struggle because they didnât seem to have the same fountain of usage data that we did on the product and engineering side. That planted a seed in my mind. I thought, âThere really should just be one version of who is your user, what did they do, and what do we know about them?â My brain kept coming back to it, and my co-founders found it intriguing as well. But who was it for? In what form? That part was super-vague.
We started to get together in the metaphorical basement, which was my house some days and my co-founder Anton [Vaynshtok]âs house other days. We kept asking, âWhatâs something that is customer data- or CRM-related?â For a couple of months, we worked on ideas in the domain, even going as far as building the prototype of a CRM for marketplaces.
We realized it was better for us to say, âAll right, marketplace CRM, letâs do this.â We did it for two months. We didnât want to raise on this right away. Instead letâs see if there âis a there there.â We were looking for the combo of something we liked and excited potential users before building in earnest. A sensation that thereâs a big problem that I would be willing to spend the next decade of my life pursuing.
Looking back, sitting in a room and trying to come up with a great startup is really bad. We played that exercise for a month, and itâs just bad because we should have been talking to people who had problems. What I realized was all our ideas sounded great, but they were not connected to a specific user pain. For us, writing code, even though itâs not the most efficient use of time, helped us explore ideas best. Maybe we did it in a less efficient way; it forced us to talk to people and build a mental model of the real world.â
âBoris Jabes, co-founder of Census
đŁ Join Lennyâs Talent Collective đŁ
If youâre hiring, join Lennyâs Talent Collective to start getting weekly drops of world-class product and growth people who are passively open to new opportunities. I hand-review every application, and accept less than 10% of candidates who apply.
If youâre looking for a new gig, apply to join! Youâll get personalized opportunities from hand-selected companies. You can join anonymously, hide yourself from companies, and leave anytime.
If youâre finding this newsletter valuable, share it with a friend, and consider subscribing if you havenât already. There are group discounts, gift options, and referral bonuses available.
Sincerely,
Lenny đ
@lenny.. I canât thank you enough for putting this together.. I have been looking for exactly this kind of playbook.. as I also aspire to build a B2B SaaS startup.. this is like Bible to me.. thanks a ton .. đ..
So many inspiring and interesting stories in this post, amazing work Lenny!