Workflow automation, fourteen years in, and the code red that reset everything: Wade Foster of Zapier

Workflow automation is a category most people encounter as a product. Wade Foster has been running the company at the center of it for fourteen years, which makes this less a conversation about automation than about what it takes to still be here.

Foster is co-founder and CEO of Zapier, which connects thousands of tools and powers billions of actions a month. Most technology startups do not survive past year five. Fewer grow into a multi-billion dollar business. Fewer still do it on as little funding as Zapier raised.

Deciding on a ten year horizon

Asked when startup problems became stewardship problems, Foster is clear the startup problems never left. They are as hungry as ever for a better product and new customers.

What changed, somewhere around the pandemic and possibly around having children, was the time horizon on decisions. He also credits listening to the podcast Acquired, and its deep dives on enduring businesses, where the recurring theme is how principled those companies are.

His observation about why so few companies go long is unsentimental. People are trying to make a quick return, or flip the thing, or they get tired, bored or distracted.

And his own arithmetic is why he thinks he can. He started the company at 24. Fourteen years in he still feels young, and another fourteen is not an insane proposition.

The concrete example is the pricing and packaging change in January 2024, which he describes as very customer-friendly and which gave a substantial amount of revenue back to customers.

His reasoning is the clearest statement of long-horizon thinking in the archive. Continuing as they were might have made more money in the short term. Giving customers more value should compound better over a decade, earning more market share and leaving them in better shape ten years out. The next two years might look worse, and he was comfortable with that, because the decision was for ten years from now.

What little funding buys you

Zapier went through Y Combinator and raised a seed round of just over a million dollars afterward. That is the only primary capital raised.

Michael's question is how much of the long-horizon thinking comes from not having to optimize for the next two years.

Foster's answer is balanced. Plenty of venture-backed and public companies think long term. And he knows many peers making short-term decisions to reach the next quarter or the next fundraising milestone, and struggling visibly with product, sales, marketing and pricing choices as a result.

Pricing and packaging is his example of where the pressure bites hardest. You can make more money by raising prices on existing customers and doing things that are not especially customer-friendly. Whether it is right over the long haul is a different question, and everyone struggles with it, Zapier included.

His conclusion: without a board member or investor over their shoulder, they probably have a little more fortitude to go long.

The code red

Asked whether there was a moment the company might have lost its way, Foster names now, and then the spring of 2023.

They called a code red when GPT-4 arrived, because the improvement over the previous model was so large and had taken only six months. Extrapolating those two trend lines, they concluded there was both an opportunity and a risk.

So they paused everything in the company for a week and ran a hackathon, with everyone building against the APIs, playing with the tools, and understanding what was newly possible.

His reasoning was categorical: if this was the next chapter of automation, and Zapier intends to be the automation company, then AI is priorities one, two and three for the foreseeable future.

His honesty about where things stand now is refreshing. He thinks they figured some things out, and does not believe anyone has figured it all out. Even among the largest model companies, the lead keeps passing between them, and it is genuinely unclear who is ahead at any given moment.

Why AI differs from previous waves

Foster has been through several technology shifts, and his explanation of why this one is different is specific rather than dramatic.

AI is a new way to automate, which hits the core mission and premise of the company. Mobile did not, partly because Zapier was very young at the time, and partly because the automation work was for knowledge workers sitting at desktops and laptops logging into software. Mobile existed and people did things with it, and it was not the predominant form factor for how those workers automated their day.

AI, by contrast, is an entirely new way to deliver automation. It arrives at the doorstep in a way previous shifts did not.

Try everything, for now

Asked what they have said no to, Foster's answer is that the philosophy at the outset was to try everything.

His justification is that there is a great deal still to learn and figure out, and people can build things very quickly now.

He is open that this draws internal criticism, with people asking which of these is actually going to be the one.

But his position is that this is a moment requiring an experimental mindset and beginner's mind, and that they are not at the waypoint where you know the winner with certainty, say no to everything else, and focus on the one thing.

He is also willing to say, borrowing his own phrase from elsewhere, that we do not have anything approaching a finished intelligence in a box yet.

Advice for companies that are not technology companies

Michael's question is a good one on behalf of listeners outside technology: what does getting AI into the workplace mindset look like when a hackathon does not obviously translate?

Foster distinguishes the cases first. For digital knowledge work, AI will be critical. If you are moving steel beams around in the physical world, AI will be part of it in a supporting role rather than the central one.

His advice for leaders in those organizations is direct. You need to be the kind of leader who is finding out what is possible with the technology. Put your hands on it. Work it out yourself.

Because done well, he thinks it can genuinely disrupt older industries: eliminating waste, improving the margin profile, lowering costs for customers.

And his warning is the reason to act rather than observe. There is an emerging pattern of pairing a founder deep in AI with a domain expert who knows an established industry thoroughly, and having that pair build the new version of the leading company in that space with a substantially better experience.

If you are in one of those industries, his point is that you would rather do that yourself than let someone else do it.

What took longest to see

Asked what problem took longest to recognize as holding growth back, Foster names moving upmarket.

He believes there was an opportunity around 2017 and 2018 to go aggressively after mid-market and enterprise, and they did not take it. They stayed in their product-led, self-serve roots, closer to a consumerized approach to business software.

It took until roughly 2020 or 2021 to recognize the size of the opportunity they were sleeping on, and his description of why is candid: they were being dogmatic, and in some ways, he says, chicken about it. The cost was valuable time.

The Belichick problem

His explanation of why they hesitated is the best analogy in the episode, and it generalizes well beyond Zapier.

There is a comfort in the business you know. You are good at it. It is familiar.

His comparison is Bill Belichick and the Patriots. An iconic coach, arguably the best of all time, who by the end had not really updated the style of play, and stopped winning as much.

His generalization is the part worth keeping. You come up in an era, there are things you are good at, and you know you can win a particular way. But the world does not stay static. People study how you work. They pay attention to what works and what does not, and they develop new techniques, because they are trying to beat you.

If you are fortunate, you can run that for a decade. Eventually there is an expiration date on the approach.

Which produces his instruction for leaders: even at the top of your game, keep looking for an edge, keep innovating, keep asking what the next act is and how you stay on top. Because that is how you earn customer loyalty repeatedly.

And his account of how they finally moved is simply recognizing the fact. There was a large opportunity out there, and they were being comfortable. So the question became why they were being comfortable, followed by a decision to push and challenge themselves. Which, as he points out, is also more fun.

The 5 things I took away from this conversation

1. Give value back if you are measuring in decades. Zapier handed real revenue back to customers on the reasoning that market share compounds better than margin does over ten years, and that two worse-looking years were an acceptable price. Very few companies can make that trade, and the ones that can rarely do.

2. Pause the company for a week when the trend line moves. The code red after GPT-4 was not a memo. It was stopping all work, running a hackathon, and requiring everyone to build against the new tools. That is what treating something as priorities one, two and three actually looks like.

3. Try everything is a legitimate strategy, for now. Wade takes internal criticism for not naming the winning bet, and his defense holds: we are not at the point where anyone knows. Premature focus in a genuinely unresolved market is a different failure from lack of focus.

4. Comfort is the thing to interrogate. They lost several years on the move upmarket, and the diagnosis was not analysis or resources. It was being comfortable and, in Wade's own word, chicken. Asking why something feels comfortable is a better prompt than asking whether it is working.

5. There is an expiration date on how you win. The Belichick point. Your approach works, competitors study it, and eventually it stops working, no matter how good it was. Which means looking for the next edge while you are still winning, not after.

FAQ

What is workflow automation? Connecting the tools a business already uses so that actions in one system trigger actions in another without manual work. Zapier connects thousands of applications this way, and Foster's framing is that AI represents an entirely new way to deliver that same automation rather than a separate category.

How is AI changing automation? Foster describes it as a new delivery mechanism for the core problem, which is why it affects his company's mission more directly than earlier shifts like mobile did. Zapier's response was to make AI its top priority indefinitely rather than treating it as an additional feature area.

How should a non-technology company approach AI? Foster's advice is that the leader has to use it personally to understand what is possible. He distinguishes digital knowledge work, where AI will be central, from physical work, where it plays a supporting role, and warns that outsiders pairing AI expertise with domain knowledge are already targeting established industries.

How do you make decisions on a ten year horizon? By accepting that the next two years may look worse. Foster's example is a customer-friendly pricing change that returned substantial revenue, made on the reasoning that greater customer value compounds into market share over a decade. He credits limited outside capital with making that easier to hold.

Why do companies miss the move upmarket? Comfort with what already works. Foster believes Zapier had the opportunity around 2017 and lost several years to being dogmatic about its self-serve roots. His broader point is that every successful approach has an expiration date, because competitors study and eventually counter it.

Also mentioned

• Zapier, its Y Combinator origins and single seed round
• The 2023 code red, and the company-wide hackathon that followed GPT-4
• The January 2024 pricing and packaging change that returned revenue to customers
• The Acquired podcast, and what Foster took from its studies of enduring companies
• Bill Belichick, and the expiration date on any winning approach

Listen to the full episode

Wade Foster on Between Two COO's

Between Two COO's is hosted by Michael Koenig. Subscribe on Apple Podcasts, Spotify, or wherever you listen.