Wayfair: Cost Structure & Culture

This is a snippet of the transcript, sign up to read more.

If I may ask, why did you leave?

It was like an "emperor has no clothes" situation. Niraj is a great commercial hyperscaler CEO but not technical. He didn't learn about technology after Steve Conine stepped away from managing that part of the business. Niraj has a lot of trust in his own judgment, which historically served him well, but it created a culture where those willing to challenge him left. He brought in senior executives from Walmart, McKinsey, BCG, Google, and Amazon, who were more inclined to placate him rather than challenge him. This led to chasing the shiny object of the quarter repeatedly, spreading resources too thin, and slowing progress. People avoided telling him about failures to avoid blame, creating a CYA environment. I had strong opinions on what Wayfair needed to focus on to succeed and tried to communicate and influence that direction. When I saw it was going nowhere, I realized I couldn't continue to give my all and energy there. You can't be a senior leader at a company if you don't believe in its direction. That's why I left.

This is a snippet of the transcript, sign up to read more.

Can you provide a tangible example of such a situation?

The second and more critical part of replatforming was data and database decoupling. Regardless of whether an application runs in the monolith or a decoupled service, it needs data. Wayfair was an extreme example of overusing relational databases, leading to a massive proliferation of databases across the company. This made it difficult to have a single source of truth for anything. Modifying existing data or database structures was challenging without disrupting a massive dependency chain of applications. Changes could break systems in unforeseen ways, leading to a mummification of the tech stack and hindering innovation. I have examples where this held Wayfair back, consuming time and energy without success. This is the most crucial part of Wayfair's tech stack replatforming. As the employee count decreased, the ability to maintain systems as they were became the focus, making new initiatives nearly impossible. For instance, the new Wayfair loyalty program was built in the monolith, not in the new replatformed stack, due to the time constraints I mentioned.

This is a snippet of the transcript, sign up to read more.

That's very interesting. I understand that in a fast-moving company, decisions are made at certain points. It's okay to make decisions to structure your data, catalog, or code in a certain way, only to realize later that things evolve and change, requiring adjustments. My question is, what took them so long to realize this?

I believe they never had effective technical leadership. When I joined Wayfair, the CTO was a man named John Mulliken. John Mulliken had been in a customer service or operations leadership role. He was a former BCG consultant with zero technology background, yet he was the CTO of Wayfair when it was a $5 billion company. During that time, Steve Conine was an engineer, but more of a hacker engineer. He's the type of CTO you'd want for the first three to five years, but he stayed much longer. Steve was always focused on enabling what the business wanted and supporting what Niraj wanted, figuring things out later. There was a succession of technical leaders who operated this way. Frankly, I don't think Wayfair has ever had a strong CTO. Jim Miller was probably the strongest, but it's hard to tell since he was interim. He had some good ideas but never really drove the organizational change management process to get things moving.

This is a snippet of the transcript, sign up to read more.