4 Tips for Managing and Scaling Your Startup’s Engineering Organization

Last week at Split’s conference, Flagship 2021, Transposit Founder and CTO Tina Huang connected virtually with Split Co-Founder and CTO Pato Echagüe for a fireside chat on Engineering Management at Scale. Their candid discussion was guided by their personal experiences and centered around how and when to scale, where to focus energy and resources, and the importance of making institutional knowledge accessible.

Their advice can be distilled into four key areas:

Stay Nimble & Small To Find Product-Market Fit#

Scaling shouldn’t always mean increasing the size of your team. In fact, both Huang and Echagüe spoke at length about the value of staying small while seeking product-market fit. Huang remarks that “it’s important to keep [your] team small otherwise it's impossible to make those sudden product adjustments.” When you increase the size of your team, you also invite additional “communication overhead.” Echagüe recalls Amazon’s famed “two-pizza team” approach to limiting communication inefficiencies and team size.

To be successful, all organizations will eventually need to shift towards what Echagüe calls the “factory mode” of operating. This development of centralized teams, structured hierarchies, and company-wide processes may result in moving slower but are necessary components of long-term company growth. However, the autonomy and agility that smaller teams offer are what allows for maximum innovation and iteration. This “innovation mode” is a key part of identifying product-market fit for early-stage ventures. It is also something that large companies like Amazon have sought to replicate in order to stay competitive and generate new ideas.

Outsource What’s Outside Your Core#

Within a small team environment, it’s impossible to tackle everything yourself. According to Huang, “it's really easy for engineers to find areas of the infrastructure that they want to make more robust or perform [better]...[but] razor focus on what you care to optimize for is super, super critical.” She goes on to share how important it is for engineering managers to shift their mentality from a focus on ‘cost-to-build’ to the holistic view of costs — to build, operate, monitor, and maintain. She believes you shouldn’t optimize for any part of your infrastructure without first finding product-market fit.

Both Echagüe and Huang caution, however, on making sure this lack of optimization does not lead to mounting technical debt. You don’t want to be forced to “stall development to go back and address things that you could have addressed earlier,” says Echagüe. This is where outsourcing comes into play. “I recommend folks who have a few core technologies — say a database or a queueing system — become an expert in those one or two and squeeze as much as you can from them. Everything else, leverage an outside provider.”

Democratize Knowledge#

For Huang, this advice is two-fold. As a leader, she sees a fundamental challenge with scaling when trying to turn “product thinking and strategy into first-order principles that [her] team can follow.” More generally, she recognizes that “specialized knowledge is still a thing even though we like to pretend it isn’t”. She distinguishes specialization through a comparison between “understanding how to write and deploy Terraform, which is ‘just code’ ” to “coding your java web app.” She then goes on to note that making this specialized knowledge accessible to anyone who engages with those systems is pivotal.

The good news is, tools can help with this. Both Split and Transposit have built into their product ways to promote knowledge sharing and eliminate gaps between specialization areas. Their goal is to provide different teams with “a common and unifying language with which to communicate and expose the right kind of controls and tools to operate.”

Choose Your Tools Wisely#

Throughout the conversation, Echagüe and Huang highlighted the significant role tools often play in scaling effectively. When evaluating whether a tool is right for you, Huang advises you keep in mind how much customization might be needed, as well as the cost to maintain that customization. “There are certain tools out there that the second you start customizing you have a maintenance nightmare...every upgrade is months of work.” She recommends teams invest in tools that limit the burden of continuous maintenance to retain customization.

Echagüe shares the importance of tools that "play well with the ecosystem and the practices that you already have” and to ensure that the tools you are using continue to serve your needs. He reminds us that teams can outgrow what a tool was originally brought in to provide. Lastly, some tools can even serve as agents of culture change. Huang underscores that exercising pragmatism and “meeting people where they are” is critical. “You can’t effect change into a culture by having them entirely uproot all their practices.” Both Transposit and Split have scaled with this in mind. Each has built into their products functionality that is easy to adopt and integrate into existing processes, rather than rendering those practices cumbersome, difficult, or completely obsolete from the outset.

For more tips from Echagüe and Huang, make sure to check out the full conversation:#