The Airtable success story is one of the stranger ones in enterprise software. Not because the product is unusual, though it is, but because the company spent years being fundamentally difficult to describe. Not a spreadsheet. Not a database. Not a project management tool. Not exactly a no-code platform either, at least not in the way that term usually gets used.
Howie Liu, the co-founder and CEO, has given versions of this explanation in dozens of interviews and never quite cracked it into a single sentence that clicks for everyone. He tried “a spreadsheet that works like a database.” He tried “a platform for building connected apps.” He tried analogies to Microsoft Access for people who remember that era. None of them fully land until you actually use the product.
That gap between how hard Airtable is to describe and how obviously useful it becomes once you’ve used it is actually the whole story. The company built something that took years to explain and five minutes to get addicted to.
The Kid Who Sold a Startup to Salesforce at 21
Before Airtable there was Etacts. Howie Liu co-founded it in 2009 at the age of 20 while he was still figuring out what came after college. Etacts was a personal relationship manager, a tool for keeping track of your professional contacts and interactions in a way that felt less like a CRM and more like a memory extension.
It was a genuinely interesting idea with a genuinely small market. Within a year Salesforce had acquired it. Howie was 21. He went to Salesforce, spent time on the social CRM product, learned how enterprise software actually worked at scale, and started noticing the gap.
The gap was this: people inside large organizations were constantly building their own workarounds. Spreadsheets that had gotten so complex they were effectively databases. Email chains that were tracking data nobody had bothered to put into a real system. Manual processes that existed because the official software didn’t quite fit the actual workflow.
This was not a small observation. It was an observation about the fundamental gap between what software engineers can build and what everyone else can do. Anyone who can code has a superpower inside an organization. They can build tools. They can create systems that don’t exist yet. The people around them, who may understand the business better than any engineer does, are stuck waiting for IT or paying consultants or making do with spreadsheets.
Howie’s original mission statement for Airtable was about closing that gap. Democratize software creation. Give the people who understand the business the ability to build the tools.
Two Years of Stealth and a Slow Build
Airtable was founded in 2012 with Andrew Ofstad, who had led product on the Google Maps redesign, and Emmett Nicholas, a seasoned engineer. They raised a $3M seed round from Caffeinated Capital and Freestyle Capital and then went essentially dark for two years.
This was a deliberate choice, and an unusual one in an era when startups were being encouraged to launch as fast as possible and iterate in public. Airtable spent those two years building the core relational engine that the product would run on. The technical problem they were solving, making relational database logic feel as accessible as a spreadsheet, required getting the foundations right before anything on top of them could work properly.
When they launched publicly in 2015, three years after founding, the product was ready. Not perfect, but ready in the sense that it was fast, stable, and genuinely did something that nothing else did well.
Andrew Ofstad, describing the early days, said the vision was always broad. They looked at computing history, at products like Microsoft Access and Lotus Notes that had tried to make structured data accessible and failed to break out of technical audiences. They interviewed people who had built those products. They looked for where those products had gotten it wrong and what the right version would look like.
The answer they landed on was not a new interface bolted on top of familiar database logic. It was something that looked like a spreadsheet, felt like a spreadsheet to start with, but had relational power underneath it. You could start with a simple grid and, as your needs got more complex, discover that the database capabilities were already there. The floor was low enough that anyone could start. The ceiling was high enough that real builders could go deep.
$30M in Revenue Without a Single Sales Rep
The early growth story at Airtable is one of the cleaner product-led growth examples in enterprise software.
They did not hire sales reps for the first few years. Not as a philosophical statement, though Howie did believe the product should sell itself, but as a practical reflection of where the growth was actually coming from. Individual users were discovering Airtable, building something useful, and sharing it with their teams. Teams were using it, expanding to adjacent teams, and eventually the product was embedded across whole departments before anyone in procurement had approved it.
This bottom-up motion is called land and expand, and Airtable executed it unusually well. The freemium tier was genuinely useful. Not crippled-demo useful. Actually useful for real workflows. People could do real work on the free plan, which meant the adoption rate was high and the conversion from free to paid happened naturally when users ran into the limits that the paid plans removed.
By the time Airtable brought on its first sales reps at around the $30M ARR mark, the reps were not opening doors. They were fulfilling demand that already existed. The product had already landed inside thousands of organizations. Sales was the mechanism for formalizing and expanding what was already there, not the mechanism for creating initial awareness.
Howie described this first sales motion as “demand fulfillment.” The reps were taking inbound signals from organizations where Airtable had organic adoption and helping those organizations consolidate, expand, and buy properly. It’s a lower-pressure, higher-conversion sales motion than traditional outbound enterprise selling. And it compounds over time as the organic base grows.
The Horizontal Bet That Investors Kept Fighting
One of the most consistent pieces of investor pressure Airtable faced in its early years was to pick a vertical. Focus on marketing teams. Focus on product teams. Pick one use case, nail it, build the go-to-market around it.
Howie consistently refused.
The argument for staying horizontal was that Airtable’s value was in its flexibility. The product could track inventory, manage hiring pipelines, run editorial calendars, coordinate film production, track cattle on a ranch, organize family archives, and manage software development all at the same time, often inside the same organization. If you narrowed the positioning to one vertical, you locked out everyone else and you also described the product less accurately than it deserved.
The argument for focusing was that horizontal products are hard to market, hard to position, and hard to build enterprise sales motions around. When you talk to everyone, you often end up resonating with no one.
Airtable navigated this tension by staying horizontal at the product level and becoming increasingly vertical at the go-to-market level. The product remains a general-purpose platform. The sales and marketing motion increasingly speaks to specific use cases, specific personas, and specific workflows. “Airtable for Marketing” or “Airtable for Product Operations” as sales conversation starters that sit on top of a platform that can do much more.
This hybrid approach, horizontal product with vertical selling, is now a recognizable pattern in enterprise software. Airtable was one of the earlier companies to prove it could work at meaningful scale.
From $11 Billion to Layoffs to Cash Flow Positive
The 2021 funding round was the peak of the peak. Airtable raised $735M in a Series F led by Salesforce and Franklin Resources, hitting an $11.7B valuation. Revenue that year was around $156M, which meant investors were pricing in a 75x revenue multiple. That multiple is a number that only makes sense if you believe the company is on its way to becoming something much larger and will grow fast enough to justify it.
What happened next is what happened to most high-multiple SaaS companies between 2021 and 2023. The market turned. Capital got expensive. Growth-at-all-costs gave way to profitability-or-die. Companies that had used cheap capital to blitzscale into headcount were suddenly looking at cost structures that didn’t match a more constrained revenue environment.
Airtable had scaled to around 1,300 employees at the peak. That was too many for the business they were actually running. In December 2022, the first round of layoffs cut roughly 20% of staff. In September 2023, another 27% reduction, 237 people, reshaped the organization again around a clearer enterprise focus.
The secondary market valuation fell from $11.7B to somewhere between $3.8B and $5B depending on whose estimate you use. That’s a painful correction on paper. Howie’s description of it in early 2026 was direct: the stock price fell about 66% from peak but the company is “throwing off cash,” has roughly half the $1.4B raised still in the bank, and grew ARR 27% to roughly $478M in 2024 while achieving cash flow positivity.
The numbers underneath the valuation correction are actually good. Enterprise revenue grew 100% year over year in 2023. Net dollar retention from enterprise customers sits at 170%, which means the customers they have are spending 70% more each year on average. For comparison, Asana’s NDR is around 130% and Monday.com’s is around 120%. Those are both healthy numbers. Airtable’s enterprise expansion rate is exceptional.
The valuation correction was a market story. The underlying business grew through it.
The AI Refounding
In 2024 Howie did something that takes a particular kind of self-confidence for a sitting CEO to do. He described what he was doing as “refounding” the company. Not pivoting. Not launching a new product line. Refounding. Starting over around a new paradigm.
The new paradigm was AI. Not AI as a feature added on top of the existing product. AI as the default interface through which users interact with the platform. Instead of building in a grid, users would describe what they wanted to build and the system would generate it. Instead of configuring automations manually, users would explain the outcome they wanted and the system would configure the path.
He went back to being an individual contributor himself. Writing code daily. Using AI hourly, not just daily. He told employees to cancel all their meetings for a week and just experiment with AI tools. The reorganization split the company into two groups: a fast-moving team shipping AI features on weekly cycles, and a slower-moving team making long-term infrastructure bets. Thinking fast and thinking slow, applied to org design.
Cobuilder, launched in July 2024, let users describe an app in natural language and have it generated automatically. It became Airtable’s fastest-adopted feature to date. In October 2025, Airtable hired the former head of engineering for ChatGPT’s business products at OpenAI as CTO and acquired DeepSky, an AI agent startup. The signal was clear. Airtable was not going to be a no-code tool with AI features bolted on. It was going to be rebuilt as an AI-native platform from the inside out.
What Made Airtable Work
The product insight was the right one at the right time. The gap between what spreadsheets could do and what databases required had existed for decades. Microsoft Access had taken a run at it in the 1990s and ended up as a product that IT departments used and everyone else avoided. Airtable came at the same gap with a completely different UX philosophy and hit it at a moment when cloud software, collaborative editing, and mobile access made a browser-based tool viable in a way that hadn’t been true before.
The go-to-market was disciplined in a specific way. They did not try to sell top-down into enterprises before the product was ready for enterprise. They let bottom-up adoption build the beach-heads and used sales to formalize what was already there. That sequence, product first, sales second, is harder to execute than the reverse but produces a qualitatively different kind of customer relationship.
The horizontal bet was correct but required patience. Staying horizontal long enough for the use cases to emerge organically meant the market defined the product’s applications rather than the founders trying to predict them in advance. Cattle ranchers using Airtable. Lawyers using Airtable. Elderly people organizing family archives. None of those were in the original pitch deck. All of them are real. The breadth is the proof that the original insight was right.
And the restructuring in 2023, painful as it was in terms of headcount and valuation optics, got the company to cash flow positive on a revenue base that was growing 27-30% annually with enterprise expansion rates that outperform larger competitors. That is a company that is running correctly. The $11.7B peak was market froth. The business underneath it was real.
Howie sold a startup to Salesforce at 21, spent three years building Airtable in stealth before launching, grew to $30M in revenue without a single sales rep, got pushed to $11B and then came back down with a better company underneath. Now he is coding again daily, using AI hourly, and describing the whole thing as a refounding.
Whatever the valuation is on any given day, that is a company that knows what it is.
Leave a Reply