
At Ansa, we've observed that some of the most compelling company-building stories come from the peers of CEOs along with their direct reports. In this installment of C[X]O Confidential, we sat down with Ted Kornish, Co-Founder and CTO of Gravity Climate, to unpack his perspective on product velocity, building software in complex markets, and what a decarbonization company can learn from Costco.
Industrial decarbonization isn't the most obvious place to build software. What made the timing and market feel right?
Ted: A few things aligned. My co-founder Saleh deeply understands building compliance products for industrial customers. He also knows how valuable it is when that compliance product can also deliver standalone value. In this case, carbon accounting and energy efficiency fit together perfectly. For example, we can use the same data that powers carbon accounting to make or save millions of dollars for our customers, which adds up when margins are slim.
My background is more on the software side. I knew that AI could give new superpowers to the small sustainability departments tasked with processing unstructured data. Our third cofounder Jay had completed his PhD at Oxford and saw climate legislation snowballing into a megatrend, though not without fits and starts. We were at the tail end of the zero-interest-rate period and carbon accounting was flooded with new firms and new capital, so we knew we would have to stand out from Day 1.
You mentioned low-margin customers. Are they good beachhead customers?
Ted: They're extremely discerning with spend. If you can sell a high-ASP product to low-margin customers, you've got something. So, we focused on unlocking the right features and positioning early to make the value obvious. That discipline has served us well.
What changed to make the unstructured data problem solvable?
Ted: The big unlocks were better tooling and infrastructure. I built a BI company ten years ago—before Fivetran, Airbyte, or [Singer] existed. We had to implement all our own syncing from scratch. And so things that would normally take teams many months could be done by individuals from an integration standpoint, in matter of weeks if you built on existing technology, which is sort of related actually to the engineering decisions we made. We knew there was an opportunity to leapfrog old-school systems using new primitives, like AI. That shaped a lot of our early technical bets.
What were some of the most underappreciated engineering decisions that shaped Gravity's velocity?
Ted: Keeping the stack very, very simple to start and making every new technology earn its place. It's the opposite of resume-driven development – a term you'll sometimes hear people throw around when you're hiring software engineers who come in and tell you about all the cool technologies they've worked with, but they can't point to the business problems that they're solving. We do the opposite – we said we're going to pick only very boring, battle-tested software to build the product on an incredibly simple architecture for as long as possible.
The thesis is that we could reach tens of millions in revenue before we needed to start fragmenting the product into microservices and adding complex tools. We picked a very simple architecture, and that informed, really everything downstream. It's the hiring strategy. Itis the product velocity. It's the way that you're able to give ownership of projects to engineers. All the data lives in only two places: a blob store and a relational database. When you're able to keep your footprint so small, you can give engineers full stack features and, say, build this end-to-end. Do it yourself. Simplicity enables Extreme Ownership, which enables velocity for us.
You were previously the first employee at a company that sold to Tableau. What lessons from that experience shaped your approach this time?
Ted: Cleargraph was not sufficiently differentiated in a very competitive market. When we started Cleargraph, the thesis was: we're going to build an analytics tool, but we're going to do it with natural language as the interface, instead of drag and drop, and that will open up a new customer group. But it turns out: to build a BI product, there are a lot of prereqs, a lot of groundwork that needs to be laid in order to have anything interesting enough to build attractive features on top of. Having a natural language interface is not very useful unless you have database connections to all the databases, and you can run the right queries, and you can do it at scale, and you have schema management and all of these things. And so the lesson was, make sure to get the primitives right and be very thoughtful about the surface area you're going to need to build out before you can show off your differentiation.
At Cleargraph, we built a lot of stuff that wasn't differentiated to get to the cool things, and the cool things made a difference. Tableau bought that company mainly for its natural language IP, but most of our time wasn't spent on natural language. It was on the boring things that Tableau immediately threw away because they weren't different. So, at Gravity, we've been very careful to both drill the basics, but also to build aha moments early. You can call them what you want — 10x features, magic moments — but we've been careful to build the differentiated stuff from day one, because we know that the product won't survive long enough to get distribution unless we have differentiation immediately out of the gate.
One example: we built utility bill scan first, even though you could barely do carbon accounting on the platform at the time. Imagine having this “aha moment” in your quiver, where you're on a call with a prospect, and they're like, “why should I take you seriously?” It's because you can drag 100 utility bills in, and we can pull the data out, and we can do your carbon accounting in seconds. It creates a sort of credibility that gets you the next meeting, by which time you've built the next 10x feature to show them. It goes without saying that you need to constantly be shoring up the fundamentals, too, so your magic moments are supported by an extremely robust foundation of audit logs, calculation transparency, and all the more mundane items.
What's a belief you didn't hold five years ago, but now feel strongly about?
Ted: Overcapitalization is a risk. Five years ago, the prevailing wisdom was to raise as much as possible, but we've seen that more money doesn't always mean a better product.
One highly capitalized competitor has actually asked us how we build so quickly with so few people. The answer is constraints. We hire senior, full-stack engineers who can own entire features. We keep the tech stack simple, and that focus allows us to move quickly without bloat.
Who are some of the builders or thinkers you look to for inspiration?
Ted: Parker Conrad from Rippling. His idea of the "compound startup, "where a strong data model supports interlocking features that create a moat, really resonates. We think about carbon accounting, energy efficiency, and regulatory reporting as different use cases that all share a common data backbone. We love hiring people from companies that we admire, and we recently hired a former Rippling product leader.
Also, the team at Linear. Their obsession with quality and visual polish pays dividends in a crowded market. Looking premium gives our product credibility and helps it stand out from the noise, enabling it to win pilots even before users experience its depth.
Linear’s observation is you don't need to fragment into teams too early—you should first figure out what you want to deliver to your customer and then build your technical organization around that rather than following rote advice. As we consider growing the team, we're being deliberate about making sure that the magic we have working right now on the product side can be preserved for as long as possible. So we're hiring people who are senior and full-stack. They're high agency and have good taste. These are all the things we need to preserve in individuals.
And on the strategy side? I love Costco. Costco is more relevant to the decarbonization and vendor marketplace side of the business, but when we think about how we're adding decarbonization layered on top of our SaaS product. It has to do with focus and picking the right SKUs and not picking too many of them.
As I understand it, Costco has ~4,000 SKUs. Walmart has ~120,000. Target has ~80,000, and because Costco has so few, they can double down on the unit economics of these SKUs. They're often the largest single customer for many of their 4,000 SKUs, and so their buyers get insane preferential pricing. There are great stories of Costco’s merchandising team who are watching 15 to 20 SKUs instead of thousands or hundreds like someone at Walmart is. That means that if you're the person who is sourcing chocolate to all of Costco, you're looking at the commodity inputs to chocolate, and when the price of sugar goes down by 7%, you're calling Hershey's. You're saying, why are you not passing along some of these savings to our customer base? That's something we will do for our customer base.
How do you balance velocity with compliance when building in a regulated space?
Ted: A large part of it comes from the people, so you need to hire the right people who know how to comport themselves in that environment. “Move fast and break things” was cool at Facebook, but eventually got replaced with “Move fast with stable infrastructure”, a much-less-discussed slogan. You need people who believe the speed and quality aren’t a tradeoff.
We’ve brought in senior folks who’ve built HIPAA-compliant systems, who know that the iceberg of enterprise software is mostly below the surface. And we pair product with simply amazing professional services. Sometimes the best way to win a deal isn’t just great software — it’s putting the customer with a Rhodes Scholar whose family started the EPA, who can lend a voice of credibility, and is able to handle precisely the unique scenario that builds trust. We’ve got both great technology and great people for a reason!