What It Takes to Reach $100M+ Revenue as a Startup | CircleCI | EO

What CircleCI Learned on the Way to $100M in Revenue

CircleCI passed $100m in revenue selling the manufacturing system for software, and its CEO and CTO explain what got it there. Most startups die before finding product market fit, so the job is to take many shots and aim at markets large enough to offer many routes, while letting customer conversations and pricing change as the company learns. A proven niche invites a Fortune 10 competitor with far more money.

Takeaways

• Give yourself many shots at product market fit. Most startups fail and very few succeed, so Jim Rose says founders should keep development and feedback cycles short and ship early. The killers are big, well-funded startups that spend quarters perfecting technology nobody may use.
• Target markets big enough to allow many routes. A $10m market forces you to capture all of it, while a $100bn market offers many routes to the same $10m of revenue. Rose says pursue opportunities you can reach quickly to learn whether the problem is real.
• Product love is the real go-to-market leverage. Rob Zuber says direct sales is expensive and self-serve, product-led growth works better early on. Companies should watch the funnel for where users drop off, and keep the paranoia that found product market fit.
• Customers describe solutions, not their problems. Jim Rose says talk to as many customers as possible, and that negative feedback produces the most valuable conversations. As the company gains expertise it must move from consuming customer information to a dialogue as an expert in the problem.
• The co-founder relationship is hard to exit. Rob Zuber calls it a very long relationship that is difficult to exit, and says he would be honest about what motivates and frustrates him before starting one. Jim Rose compares a team to a basketball lineup whose pieces must fit together.
• A Fortune 10 rival will copy your niche. Jim Rose says technology cycles have shrunk from three to six months to five to ten days, and generative AI teams now plan in six-week cycles. A startup may get 20% of the way before handing the baton to a larger organization.

CircleCI's shift to usage-based pricing

1. Rob Zuber joined CircleCI in 2014, when the pricing model was already established.
2. As customers grew comfortable with elastic compute, they moved from fixed contracts toward paying for an EC2 by the hour and then by the minute.
3. Customers asked why they paid for fixed bandwidth on the weekend, which pointed toward capacity and usage models.
4. CircleCI moved its entire pricing model to usage-based, a long transition. Zuber says you cannot switch off the old platform until the last customer leaves it.