Wayfair Tech Blog

Wayfair Descends on Google Cloud Next 2023

When Wayfair thinks of the cloud, we see a massive world-class “development accelerator” that dramatically reduces the friction for developers to provision, build, and create the things they want without having to toil with the underlying infrastructure to “wire up” those concepts. When we think of cloud innovation, we think of Google, which is at the forefront thanks to its focus on innovation.

That’s one of the reasons Wayfair makes the trek to Google Next each year. There’s widespread agreement among our team regarding the benefits of attending this event. For example, we learn about new Google Cloud Platform (GCP) services and capabilities and what’s on the roadmap. Members of the team also get the opportunity to present the innovative initiatives they are working on that leverage Google services (e.g., Duet AI for code, Vertex), as well as the chance to hear about what other leading technology companies are doing.

Another reason to attend is to connect with Google as a partner rather than a customer. The Google relationship is extremely important to us, and Wayfair’s technology transformation and these meetings are incredibly fruitful, giving us the face time to explore ways we can help drive their roadmap and identify opportunities for joint success that will pay dividends for both sides as well as the larger vendor community.

These views are shared by all members of the Wayfair team who attended this year’s event. But our event speakers also walked away with some of their own unique views on why it was important to attend, what they spoke about, their big takeaways, and more. Here’s a look at some of what they shared.

Matt Bruno, Associate Director, Production Operations & Site Reliability

Why was it Important for You to Attend?

For me, it was important to demonstrate to Google that we are not just a customer buying their solutions. We are a good partner that is co-invested in the relationship and building things back into the ecosystem.

What were your big takeaways from this year’s show?

Aside from getting a look into Google’s product roadmap, we got the chance to see where other technology shops are in their adoption of Google solutions, which allows us to see where they are versus Wayfair. What I found is that we are a lot further along than many other companies. It was an invigorating realization to see where we stand versus these companies and then get a glimpse at where Wayfair could be a year from now.

What did you present at Google Next?  

I lead production operations and site reliability at Wayfair, so my job is to keep everything running and manageable. For my session, I spoke about metadata and how we corral, measure, and keep tabs on the infrastructure on our platform so that we can manage it at scale—I like to use the analogy, how do you tag all of your sheep so that you know the flock is altogether or if you have lost one or two.

In the session, I discussed some of the things we have rolled out, such as the helper pieces we have built. From there, I touched on how we have managed it, kept it running, and made it measurable and cost-contained as Wayfair scaled into a multi-billion dollar company with billions of impressions and a huge platform footprint. In essence, the session allowed me to show how Wayfair is now sitting at the “big kid table” with other large-scale companies.

You can view Matt’s session here.

John Keane, Director of Software Engineering

Why was it Important for You to Attend?

Wayfair is beginning a journey where we will take all of the live traffic offline and put a Federated GraphQL layer between the client and the data. To do this manually would take years. Our plan is to take advantage of Google’s Generative AI capabilities to help us transform the code we have in our legacy SQL instances to a new API layer.

We will then use the same Gen AI capabilities to pick all those select statements and transform them into live code that we can push out to teams so they can quickly roll out those changes. This project has the potential to accelerate our technology transformation from a period of five years down to a one to two-year window. Wayfair’s partnership with Google is vital to this effort, so attending this event is critical.

What did you learn? Takeaway?

Generative AI was a huge part of the conference, and we are clearly at a transformational moment in our lives. I also learned that our work is truly unique in the industry. You go into these events thinking that all businesses are essentially in the same place and doing many of the same things. What I learned is that what Wayfair is doing is unique and novel and together Google and Wayfair are blazing a trail for the industry.

What did you present at Google Next? 

I am the Director of Engineering for Global Supplier Technology. In that role, I support the platform team, including product catalog, pricing capability, supplier services, and integrations. My talk centered on the technology transformation journey that I mentioned earlier, delving into how we are migrating all the live traffic, building the new infrastructure programmatically through Gen AI, taking all the live traffic and migrating it, helping feed the data out of the legacy databases and transforming it onto newer modern single databases build on Google Cloud, Spanner, and others. A key point of emphasis throughout the session was that none of this would be possible without the Google partnership.

You can see John's presentation here.

Mark Quigley, Director of Engineering Enablement

Why was it Important for You to Attend?

For me, it’s the technical conversations we can have. These are critical. Aside from connecting with Google, our technology leadership team talked to the technology leadership team at Uber and compared notes on how we see these technologies unfold, the ways they can be adopted, and what we think is a good use case for them. It’s a sanity check on our expectations as well as our aspirations.

Another important reason to attend was to sit in on some of the scheduled sessions. I attended one focused on developer productivity, which is what I do in my role at Wayfair. I not only got to see the presentation, but the Q&A afterward gave me a chance to see how other companies, such as Spotify, are thinking about productivity and, in this instance, how they are leaning into hard metrics for things like NPS in a way that’s fairly predictive. This gives me something to think about and consider if Wayfair should be weighing these metrics more than we have been.

What were your big takeaways from this year’s show?

That’s easy, Generative AI. Gen AI is going to have as significant an impact on the world as the Internet itself has. Naturally, as with the Internet many years ago, Gen AI will follow the same hype curve in that it will be over and under-hyped. But at the end of the day, Gen AI will have a huge impact on how we work individually, work with others, and more. But it’s not going to eliminate huge swaths of jobs. In fact, we will see new jobs coming in, such as prompt engineering.

What did you present at Google Next? 

It looked at how Duet for Code is helping engineers at Wayfair. It’s still early days for us, but we are already seeing results in the POC and pilot. That includes reviews from developers in the 60% to 70% range—what this means is that 60% state they are working on more meaningful things when using Duet, and 70% feel that they are dealing with less boilerplate-type projects. These are the types of results that we are seeing so far. Net productivity gains are also in the 15% to 50% range. That’s significant. If you take anybody in any role and ask them how you can get them to be 15% better and faster without having to work any harder, that would be a big challenge. That’s what Duet AI does for us.

You can view Mark’s session here.

Emily Roth, Senior Software Engineer

Why was it Important for You to Attend?

I wanted to learn about the GCP product roadmap to see what’s coming and help influence those improvements that will meet Wayfair-specific needs. The opportunity to network was also important. This year, I got to hear from other customers about how they are applying Generative AI products and tooling in their companies. We were also able to share our experiences and successes, which gives others in attendance the opportunity to apply what we’ve done, and then maybe next year, we can learn something new from them.

What were your big takeaways from this year’s show?

From the big-picture perspective, we saw the sheer speed of Google’s AI innovation. The projects we talk about today are very revolutionary. Then you get to see what’s coming down the road, and you realize how quickly everything is advancing and how that will allow us to push the needle even further in the not-so-distant future.

On a more granular level, I learned more about Google’s Duet AI. In my role, I work on the AI code generation side and ran a pilot program for Duet AI’s code assistant evaluation. At Google Next, I attended some sessions where I learned about new Duet AI use cases. For example, in addition to using Duet AI to write code, engineers are using it to plan out architecture diagrams, think about the problem, and how AI can help all parts of their software development lifecycle.

What did you talk about?

I spoke about the Duet AI pilot experiment, where I evaluated if our engineers were more efficient and their jobs more satisfying when using Duet AI. I presented our findings and how Duet AI improved productivity and increased different metrics that we were looking at, specifically developer satisfaction. These findings ultimately led us to recommend adopting Duet AI across engineering.

You can view Emily’s session here.

Zubin Rupani, Engineering Manager

What was presented at Google Next? 

Our presentation focused on our effort modernizing our centralized logging platform from Elasticsearch to Google Cloud Logging (GCL) - an innovative strategy and delivery that won the 2023 Google Customer Award for that category. We anticipated that a monolithic centralized logging platform would not be a scalable solution as Wayfair transitions to a microservice architecture and leverages more cloud-native services. With product teams having the freedom to choose their programming language, libraries, and infrastructure components, managing logs from all of their different systems would be a complex undertaking. Transitioning to GCL provides this functionality out of the box and streamlines the cloud-native adoption process while keeping public cloud costs low. In addition, it provides a unified view of all logs and metrics, as well as integrating with existing tools and applications. We believe that this will be more cost-effective and improve efficiency while also providing a better experience for software teams at Wayfair.

You can view Zubin’s session here.

Google Next 2023

Brenton Sellati: Director of Product, Global Supplier Technology

Why was it Important for You to Attend?

I have a product-focused role on Wayfair’s Global Supplier Technology (GST) team. In that role, I’m looking for ways we can extend our tech transformation and solve the highest ROI problem in the most efficient way possible. For this, I need to understand the developer ecosystem and gain context into what my teams are talking about, what opportunities I see with GCP, and how I can help connect the dots. That’s why it was important for me to attend this event.

What were your big takeaways from this year’s show?

Google is laser-focused on extending what they do best, developing excellent software, and making that available. They are continuing to lean in on their strengths in unstructured data, machine learning, and AI. Because of this focus, Google Cloud is best positioned to help clients quickly unlock the value of things like Generative AI.

What was your session focused on?

Wayfair has grown up with Microsoft SQL as the core database technology that runs all parts of the business. But if you’re going to run SQL in a non-Microsoft cloud, how is that going to work?

My session focused on Wayfair’s merchandising strategy and the development of Market Context, which is replacing our current system, Brand Catalog, which allows us to merchandise products and create unique experiences differentiating by store, country, and language. While this data is in GCP, it doesn’t support B2B or physical retail. It’s also locked in SQL servers.

Market Context will give us all the dimensions we need for flexibility today (store, brand, language, country, etc.) while adding B2B vs. B2C, online/offline for omnichannel, and more. And unlike with Brands Catalog, these elements can be independent of one another. It’s a massive unlock that will allow us to better support today’s footprint and our expansion into physical retail, new languages and countries, and much more.

You can view Brenton’s session here.