Applying Stripe’s lessons to Customer Support

Why support is the next payments

John Wang
8 min readJul 12, 2021


When I left Stripe and started a company in the customer support space, most people thought I was crazy: “why would you leave that rocket ship?” But in many ways, Stripe taught me to see the long game and to invest in the opportunities that no one else paid attention to.

Think back to 2010: the Collison brothers saw a massive problem to be solved in an industry that was considered a backwater by most startups. Back then, it was all about the flashy consumer apps, social networks, and buzzy tech, but for Stripe, it was about staying focused on the opportunity no one could see.

I strongly believe there’s a new wave of change coming to the customer support industry and that it has many of the trappings of the payments industry a decade ago. The two are quite analogous:

A massive, growing market

Proactively delighting customers earns trust, which earns more business from those customers, even in new business arenas. Take a long-term view, and the interests of customers and shareholders align.
Jeff Bezos, Founder of Amazon

Customer support is a massive, growing market and a differentiator for the world’s best businesses. Much like payments in 2010, a confluence of macro factors today have led to the growing importance of customer support:

  • Consumer expectations have shifted. Amazon changed the way people shop online by opening people’s eyes to what outstanding customer support looks like. Consumers now expect to receive refunds for damaged items, get tracking updates, and have missing items taken care of. Consumers also make purchasing decisions based on the quality of support they receive. In 2020, Microsoft found that 90% of consumers use customer support as a factor in whether or not to do business with a company. A decade ago, customer support was hardly a factor in purchasing decisions, but it has now become one of the dominant inputs.
  • The world is moving online. The COVID-19 pandemic accelerated the push to drive life online. Online transactions have been trending upwards for years and now account for over 20% of all retail transactions. Anecdotally, the best brands are advancing with online commerce squarely at the forefront of their minds. Nike met its goal of moving 30% of its total sales online almost 3 years ahead of time.
Graph from From
  • Companies have budget to spend. There’s a massive amount of money in customer support. In the United States, there are twice as many customer support agents as truck drivers. It’s a $105 billion market in the United States alone, and a lot of global spending occurs in Asia and South America. The software industry for customer support is also growing, evidenced by the fact that Salesforce’s customer support product now brings in more revenue than its namesake sales product.

In 2010, the world was becoming more digital and Stripe rode on the coattails of the growth in internet businesses. Today, those internet businesses have progressed and must offer superb customer support to stay competitive.

Complex, outdated tools and processes

[I was] baffled at how convoluted and awkward [online payments] appeared to be. It seemed like a prevailing ecosystem designed to reduce the number of Internet businesses.
Patrick Collison, CEO of Stripe

Building a business in 2010 that took payments online was hard. You had to walk into a bank to fill out an application, pay thousands of dollars in fees, and connect a couple of arcane payments gateways. The whole process could potentially take months, which is what Stripe aimed to solve.

Likewise, scaling up your customer support team today is incredibly difficult. Tools like Zendesk and Intercom send messages but they don’t help you manage your team because they’re the communication layer and not the intelligence layer. Most businesses still rely on spreadsheets to manage a labyrinth of remote workers, all of whom have their own preferences and skills. Modern support teams now work across a wide variety of channels and platforms (like chat, phone, Twitter, WhatsApp) and work in different timezones. What’s more, each agent can have different specializations and skills: from answering developer focused support questions for a company like Stripe to figuring out how to refund a transaction on GoFundMe’s website.

Actual industry-leading customer support software. Many solutions are still desktop based and fighting to move onto the “cloud”.

Most industry-leading customer support tools were built for use in call centers, where the dominant worldview is one of command and control. A hierarchy of managers instruct workers how to maximize efficiency, leading to incredibly specific requirements. It’s pretty depressing to visit an actual call center: bathroom breaks are scheduled down to the 3 minute interval, and taking too long would be punished with a lower performance score. Most call center software is sold to Fortune 500 companies, costs millions of dollars per year, and requires a team of people to implement and maintain them. This is the world that has spawned most of the support tooling we see today.

One of our values is that you should be looking out for each other. Everyone should try to make the lives of everyone else who works [at Slack] a little bit simpler.
Stewart Butterfield, CEO of Slack

Many modern support teams want to scale their operations up, but not by setting up with a cell center with thousands of people. Most fast growing companies don’t see support agents as cogs in a machine, but instead want a system that is rooted in empathetic support. There’s a new era of companies that are pushing for this, and the old tools aren’t ready to handle this shift.

Fundamental infrastructure for the internet

The same way that Google exists as a foundational component of the Internet around information retrieval, it felt like there should be such a foundational component for economic infrastructure, and that was sorely lacking.
Patrick Collison, CEO of Stripe

Payments and customer support are both areas that are fundamental to operating an online business. They are both necessary and a driver of differentiation among companies:

  • Foundational. Every single online business needs to accept payments and support its’ customers, there’s no way around it. Imagine buying something online and not having the ability to contact the company about a lost package or a defective product. The concept is very much akin to sending a check through the mail for an online purchase: it may have worked for mail order catalogs in the 1990s, but having no support today would lead to a quick death.
  • All companies face the same problems. Everyone uses the same credit card network because payments systems aren’t created for a specific industry. A company selling balloons can use the same system as a company selling enterprise software. Likewise, customer support remains largely the same across different companies. Whether you’re Glossier selling lipstick or Monzo providing banking services, you must forecast incoming support requests, portion it out to the right people, and solve those requests as quickly and accurately as possible. Some requests may take longer or require multiple steps, but the concept of putting people in the right place at the right time remains the same for all businesses.

A hard, non-obvious problem

Even the smartest, most imaginative people are surprisingly conservative when deciding what to work on. People who would never dream of being fashionable in any other way get sucked into working on fashionable problems.
Paul Graham, Co-founder of YCombinator

Stripe started in a non-obvious place: no one really talked about APIs and few people cared for payments. The sole payments related business in the top 100 most valuable startups in 2010 was TrialPay, valued at $200 million and sitting at number 58. Only the small number of people who had tried starting an online business understood the problem of payments and it wasn’t obvious to the majority of Americans or the even the majority of people in Silicon Valley.

In the same vein, few people in Silicon Valley have expertise in customer support. I’ve heard anecdotally that engineers will literally quit if they have to work on internal customer support tooling because they don’t believe it to be interesting or impactful. Even fewer people in the management of most technology companies have answered a support ticket themselves.

But if you‘ve ever seen a large support team run, you’ll realize that support is both important and ripe for innovation. My cofounders at Assembled built out the support tools team at Stripe and found massive gains from targeted changes. At first, people laughed at them for working on support tools, but towards the end of their tenure, their small team of 2 people was replaced by a full team of dozens of product managers, engineers, and business people.

So what’s one to do?

“Change will not come if we wait for some other person, or if we wait for some other time. We are the ones we’ve been waiting for. We are the change that we seek.”
— Barack Obama

At Assembled, we’re building the intelligence layer for customer support. We’re rethinking the way that modern companies scale out their customer support teams. Instead of starting with a hierarchical call center approach where thousands of people have their every move directed to the minute, we’re starting with the belief that capable, autonomous humans given the right information will provide great customer experiences. We’re building software that helps forecast tickets and optimize schedules to where and when agents are most needed.

My cofounders and I spent years at Stripe seeing support problems firsthand. We’ve talked to hundreds of different companies about how they think about customer experience and we’re now applying our knowledge by building the best software for any company that is scaling their support team.

If you’re excited about the customer support revolution, come join us or email me at

Thanks to Megha Narayan, Nimalan Mahendran, and Olivier Godement for feedback on drafts of this.