September 8, 2020
This week we’re having tea with Andrew Ofstad, Co-founder and Chief Product Officer at Airtable. Andrew previously led the redesign of Google Maps and worked as a PM on Android.
In this episode, Andrew shares:
The timeline of getting Airtable to market
Tradeoffs between going after a narrow use case and building a horizontal product
How Airtable thought about pricing their Pro account
This interview has been lightly edited for length and clarity.
Andrew Ofstad: I started in 2012 with my co-founder Howie, our CEO. Our third co-founder, Emmett, joined shortly after. Early days were just trying to set a solid foundation; we took our time learning about the space. I did a lot of reading from a lot of early computing pioneers, like Alan Kay, Doug Engelbart.
We both left our respective companies about the same time. I was at Google, Howie had started a YC company, Etacts, that was bought by Salesforce. At Salesforce he saw that a lot of useful business software is more or less just a database with some CRUD actions and views and workflow on top of it. We were inspired by the problem of taking this thing that we do as programmers, which is create useful software, and enabling anybody to do that.
We were all engineers. We actually knew each other in college and used to work on hack projects together and talk about tech a lot.
In some ways it validated what we were working on. A lot of the best ideas have been around for a while, and the time hasn’t been right. Or the execution hasn’t quite been right. We saw these old products like HyperCard, old database products that were end-user focused, the creation of the original GUI operating system. Moving from a command-line interface to something much more accessible to a broader audience. It showed us this was possible and kept us going.
There were practical things too. The way they invented the initial GUI operating system, and the early days of Mac, was a lot of prototyping, coming up with metaphors, and testing them out with users. The next day, rinse and repeating, making a bunch of changes, and testing that with users. We pretty much took that approach in the early days of Airtable: create a lot of prototypes, create a very early demo, try it, show it to people, get feedback, and iterate quickly.
We were building a full thing. It was basically a complete, front-end only product. We built the whole thing in our own Javacript framework on the web, in the browser. And we didn’t actually build the backend, so it would work really well and persist to local storage in the browser.
We tried to prove out the interaction paradigms and the design. The main problem was simplifying this very complex thing in the database. What’s the design to make that accessible to an end-user? Building out this spreadsheet interface, and this very simple table metaphor that people could easily interact with and understand, was the first thing we needed to de-risk.
We definitely were building the prototype with collaboration in mind. Pretty early on we built out a simple collaborative layer; it still did something hacky, it shared data through local storage. There wasn’t backend persistence, but it still allowed you to simulate the collaborative aspects of it. But we knew the collaboration aspect was important, and wanted it to be fully real-time.
After six months or so of just a purely front-end prototype, we started building out the server side. At that point we got initial customers using the product and validating parts of it, giving us feedback. It was over the course of a year, a year and a half, before we had a real alpha product for people to actually sign up for it.
We quit our jobs and within a couple of weeks had a very early prototype. It was probably three or four months later that we started showing it to investors and pitching the product. Even a year before that, while we were both working our full-time jobs, Howie had the broader idea and vision for the company. So the pre-thinking of the market and opportunity happened well before. If you look back, our Series Seed pitch deck was very similar to what our vision is today.
A lot of investors gave us that advice and said, “Hey, you should start with a very narrow use case.” But we were pretty committed early on to building a horizontal product. And there are trade offs there. It’s harder to market; it’s harder to say, “Here’s our customer, we’re building just for them”; things like how you prioritize features becomes a lot harder. But we thought that building a niche product would probably come at the cost of a broader horizontal opportunity.
That said, pretty early on we did start to see some use cases getting more traction and tailored our roadmap to that. So you see a lot of people using it for product management of some sort, or tracking things with dates. So let’s build a calendar view. And a kanban board versus a map view for geo-location services. We laid a pretty wide net but then started to see numbers and use-cases pop up and started to double down from a prioritization standpoint.
It was a bit of both to be honest. We would constantly get points of validation along the way. We got a very early alpha out and saw one non-profit start using it, tracking all their students, all their donations, basically their whole business in Airtable. They were just loving the product, and it was still pretty raw at that point, so that’s one milestone for us.
We keep adding on to the product, and I think we knew from the start the MVP for this thing was a pretty high bar. To some extent, you’re competing against spreadsheets that have been around for 30 years and are incredibly feature rich. We realized it would take a long time to get a real MVP out there.
Two or three years in, we start to see customers in different industries. In 2016, we see our first enterprise customer. You see that the vital mechanics of the product do work. And you see it spreading from team to team. So it’s this breadcrumb trail of validation along the way early on in the product. We do have these viral mechanics; the exponential growth curve is very flat and spiky. And we would have a press event and it would spike and then go flat for a long time. The actual metrics may not have looked that great, but we still had customer stories and evidence of people getting a ton of value out of the product. It just kind of hit that inflection point. But it’s definitely a very slow build.
Before we were charging for the product, that was a good indicator if people were using the product. Even if they’re just using it once a month, or once a day. Once we started charging for the product we added our Pro plans on, and I think people vote with their wallets so that was the biggest indicator and our main metric for a while.
It is a very sticky product. The challenge is you have to take Airtable as a set of building blocks and fit it to your use case. So you have to figure out the product and configure it for whatever you’re using it for, so that sometimes takes a while to get over. But once you’re over that hump, we found that people stay with the product and become huge evangelists.
Templates have been really important. People don’t use the templates verbatim, but they’re a good way to see what you could create, or inspire you. There’s also things we did just to make it feel like something more than a spreadsheet, which it definitely is. Things like color and select values, calendar views, and making it feel more like a rich application versus a text and number spreadsheet.
But there’s no silver bullet, I think some people are tinkerers, and they’re willing to experiment. Finding creators is important as well.
When we first launched the product, we did have a pricing page. We did think the product was premium and framed more along a product like Salesforce, versus a productivity tool Like Google Docs, or Quip. We ended up landing on $24/user/month month for the Pro Plan.
Honestly, we weren’t super scientific about it. We started putting any new feature we built only into the Pro Plan. With the rationale that it’s easier to move things to a lower plan than it is to take it away from the lower paid plan and move it up.
We had this pricing page where users could put a credit card in, but we hadn’t actually built the system to start billing them. Six months in we finally started charging them, gating more features, and enforcing some of our usage limits. It was a progression of adding in more features to that Pro Plan over time. We felt that what we had in our free plan was a pretty good set of things that could prove the value of the product and get people to fully build something out for their use case. The more advanced features in the Pro Plan were unlocks for larger teams scenarios and advanced functionality.
No, we were pretty hesitant to do things like make it free until you’ve added five people. Because then you have this incentive where people don’t want to add collaborators and spread the product. We were always sure to have things like shared views, where you can send a link to somebody.
The premium features do add advanced functionality and permission features that prevent people from doing certain things. But the Free Plan has all the features that allow people to use the product in a viral way.
It’s not something we predicted. I think it makes sense, given how much digital transformation has happened at companies, and how powerful software is. And I think it’s impossible for other companies to create the perfect vertical software for every single use case, given that every company does things differently.
Looking back it does seem like it’s inevitable. It’s really exciting, it’s a whole ecosystem with tools now. Companies like Zapier. There’s a lot of cool stuff out there, like Standard Lib and Retool.
Raise more money than you need, I would say.
Yeah I would say 30%. Maybe just make sure you have at least two and a half years. That’s something I would be comfortable with.
I’m kind of missing, with remote work, just a good whiteboarding experience. Especially with interviews, just brainstorming. Zoom has this crappy little whiteboard thing where you can invite somebody from other tools. Something that combines video plus free-form creative brainstorming in some way would be really powerful.
Yeah, exactly. And if it’s just like a collaborative, if each person just had an iPad and could easily sketch things out. That would be ideal. Someone should build that.
Want to hear more? Subscribe to Series Tea on YouTube, iTunes, Spotify, Stitcher, or your favorite podcast app.
You can also send us a note at [email protected] We’d love to hear what you thought of the episode, or who you’d like us to have tea with next.
The Mercury Team