Plato

Login to Plato


This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

Don't have an account? 

Achieving scale as an engineering leader — from technology to teams to culture

13 October, 2020

Scaling means different things to different people. But in any case, achieving scale is impossible without a good framework. In this article, Vijaya Kaza (Airbnb’s Head of Engineering for Trust & Safety) shares tips so you can successfully scale your team. She sat down with Christian McCarrick, a senior engineering leader at Facebook, to lay out a framework that she uses (thanks to the team at Lohika for connecting us with her!).

To Vijaya, scaling is about creating multiplier effects. In fact, it means figuring out how to invest in the most highly-leveraged areas so you get the biggest outcomes, ideally step-function changes.

The Pyramid of Scale

She uses a framework called the Pyramid of Scale (https://i.imgur.com/FgYnk9i.png):

  • The bottom layer refers to foundational capabilities (e.g. technologies like automation or methodologies like agile or DevOps).
  • The middle layer focuses on scaling teams. This doesn’t always mean growing the size of your team, but instead, this can refer to getting more out of your existing teams and making it easier to operate.
  • The top layer is culture — you can have the best tools and people, but if you don’t have the right culture, then teams aren’t rowing in the same direction.

Change management is hard, and simply knowing which solution to adopt isn’t enough. Leaders have to bring their teams, peers, and executives along for the journey. Vijaya often spends her time on getting buy-in because it’s so fundamentally important. You want to avoid a monolithic, top-down approach when implementing something new, and instead, make it a collaborative effort that gets inputs from different levels. The goal is to make any change feel like it’s the whole group’s idea.

As for scaling your foundational capabilities, Vijaya suggests:

  • Automate test frameworks and unit tests; automate the end-to-end cloud deployment pipeline and use tools like Spinnaker
  • You want tools that let you not just deploy but roll back changes if you need.
  • Data is a core piece, so leverage machine learning and automate as much business decision- making as possible
  • Not every decision can be automated, but create frameworks that make decision-making easier

As for scaling teams, Vijaya advises:

  • Have a clear “operating system” — there’s often confusion around who does what, and whether there’ll be a top-down or bottoms-up approach to decision making
  • As a leader — explain why something is being done, what needs to happen, and then empower teams to figure out the “how”
  • Always ask whether building a technology will support the company’s core mission (when deciding to build vs. buy). Building every single thing in-house is not scalable, and buying comes in the form of buying off the shelf, leveraging open source, and outsourcing.
  • You need to ruthlessly prioritize — know what you want to do as a team AND what you won’t do.

Ruthless prioritization

To unpack the concept of ruthless prioritization — it comes down to having the discipline to draw a line somewhere and being comfortable that certain items are falling below the line (and not worth doing for the next quarter or whenever your timeframe is). People want to do everything, but when you do everything, you don't do anything really well. Also, Vijaya recommends being comprehensive when you prioritize — it’s easy to only focus on customer-facing features, but take into account technical debt, all of the maintenance, all of the forward-looking innovation, or any type of capability like that. Put all of those initiatives through the same vetting process, and have strategic conversations about why you are (and are not) doing something this quarter. In fact, our host Christian McCarrick referenced something that Sherly Sandberg tells her team — to paraphrase: it's not about deciding between good projects versus the bad; it's about deciding which of the good things you're not going to do and which things you want to make great.

As for how to prioritize, Vijaya recommends:

  • Before you rank anything — have a clear understanding of the stakeholders who are requesting features, tools, etc. (the Product team, Customer Success team, Sales/Marketing, etc.)
  • Create a stack rank list of every single initiative, feature, etc.
  • Continually rank things to a point where it’s a muscle that you get better and better at

Above everything, culture is critical. If your team members are not aligned, and they aren’t motivated and happy, then nothing else matters. How do you achieve that? First, incentivize ownership for people. You ultimately want every individual to feel like they're an owner of the company, not just an employee. Next, focus on customers — in fact, obsess over whether your customers are happy. Some customers are external and some are internal, but nevertheless, they are always a priority. Finally, democratize communication so there’s uniform and smooth flow of information and consistent messaging across all the teams in your company.

Here are some things that Vijaya does to increase the flow of information:

  • Post a thought a day on Slack (even just a paragraph) to your team — it’s helpful especially when you receive feedback on something or your team is confused about something, so this is a chance to clarify, inspire, and inform.
  • Host small round table meetings — invite a group of 5 to 8 people so they can open up and discuss a specific topic (because people aren’t always so keen to share when it’s a much larger forum).

As your teams scale, make sure to track progress qualitatively and quantitatively. Ultimately, you want to make sure you’re getting a positive ROI and the unit cost goes down.

Embracing change

Finally, Vijaya advises that engineering leaders need to embrace change. Chasing the latest shiny object isn’t the goal, but rather leaders should deliberately strive to stay current with new technologies and frameworks. As a principle, she recommends that engineering teams stay no more than one generation behind (because every other change can become difficult if they start stacking up). She recommends that engineering leaders do a few things:

  • Participate in public Slack channels
  • Read blogs, listen to podcasts, attend conferences
  • Talk to other leaders, especially tech leaders in other verticals

We hope this gives you some tactical tips on how to achieve scale. Thanks for reading!

You're a great engineer.
Become a great engineering leader.

Plato (platohq.com) is the world's biggest mentorship platform for engineering managers & product managers. We've curated a community of mentors who are the tech industry's best engineering & product leaders from companies like Facebook, Lyft, Slack, Airbnb, Gusto, and more.