A Bulwark in Volatile Times: More Efficient Engineering Teams
By Vivian Guo | December 7, 2022
Insights from the 2022 ICONIQ Growth Engineering Report and Private Engineering Summit
Alphabet CEO Sundar Pichai sent shockwaves through the tech industry earlier this year when he declared it was time for a “Simplicity Sprint” to find 20% more productivity at the tech behemoth to “get better results faster.” When macroeconomic trends drive even the most successful companies to examine their profits and practices, engineering leaders at organizations of all sizes should ask themselves difficult questions about the productivity of their teams, where they can improve, and how quickly they can deliver efficiency gains.
Unfortunately, that’s easier said than done. Most technical organizations don’t have effective systems in place to measure and track the output of their developers, and doing so has been made even more difficult by hybrid work. Lacking hard data, teams and leaders can find themselves misaligned on issues of performance, an issue Microsoft CEO Satya Nadella has dubbed “productivity paranoia.”
“As software tools have made it easier to track and optimize performance across many roles, business leaders have adopted data-centric management philosophies and set KPIs as their teams’ True North. However, despite all those advances, engineering productivity has continued to be a black box,” says Aditya Agarwal, Former CTO at Dropbox and ICONIQ Growth Partner. “But now that market sentiment has shifted away from growth-at-all-costs strategies and toward more efficient models, technical leaders are being challenged to find new ways to measure developer productivity and more wisely allocate their increasingly far-flung resources.”
At ICONIQ Growth, we noticed a dearth in information on this topic that could benefit technical teams, despite a growing need for such data from C-level executives at our portfolio companies and others in our network. Two years ago, our analytics team began collecting research and delivering data and quantitative insights about engineering organizations. We are pleased to share this year’s report on how engineering organizations have changed in 2022 amid market volatility and a shift to hybrid work.
Related, on October 27, 2022, our Technical Advisory Board (TAB) hosted a private engineering summit in San Francisco for more than 70 technical CEOs, CTOs, and CPOs to discuss the most urgent topics facing technical leaders today, including impacts from the shift to hybrid work and how to build antifragile engineering organizations in volatile times. Members of our TAB include Aditya Agarwal, former CTO at Dropbox, Matt Eccleston, former VP Growth at Dropbox; Mike Abbott, VP Engineering at Apple Cloud Services; Mike Curtis, former VP Engineering at Airbnb; Neha Narkhede, co-founder and former CTO at Confluent; Keith Adams, former Chief Architect at Slack; and Anantha Kancherla, Head of AI Infrastructure at Meta. Attendees included technical leaders from BetterUp, Chime, dbt, Drata, and Monte Carlo, among other market-defining companies.
Below are highlights from the event and our report on “Engineering in a Hybrid World.”
In order to achieve a productive and efficient engineering organization, you should consider using a productivity framework that is built for the long-term and can be understood by your non-technical leaders and colleagues.
As an engineering organization grows, different types of questions and challenges emerge around the investments in time and people that the organization is making. It’s critical to have a framework that enables the company to think about productivity and prioritize engineering investment in a way that makes sense for engineering internally but is also understandable by the rest of the business.
In 2022, engineering organizations allocated ~15% of total capacity to “keep the lights on” (KTLO) activities and ~85% to elective investments across new capabilities, quality improvements, and internal productivity . Multiple attendees at our summit attested to throwing out old frameworks and moving over to a simpler structure like the one above. “It was like a light bulb went off for our CEO,” said one Summit participant.
Developer productivity is no longer a taboo topic. Instead of struggling to track everything, our Technical Advisory Board recommends tracking three to four metrics that allow you to get started quickly and build a reporting muscle over time.
While specific KPIs will vary across companies, we generally recommend engineering teams track a few metrics to understand revenue / FTE cost, release time, and developer velocity on a trended basis. There can even be value in tracking unproductive metrics like time spent in meetings or lack of contiguous two hour “flow blocks.”
The roles and structure needed for engineering teams will differ massively across orgs. However, we commonly see the below ratios.
A common challenge shared among attendees is that while “one manager to eight engineers” is the rule of thumb, first-time managers often do not want to manage such large teams or let go of independent contributor (IC) work. As such, it’s imperative to design the right paths for engineers to progress in your organization (whether in a senior IC role or people manager).
In times of uncertainty, building resilience and insulation from market ups and downs comes down to your people and culture.
“As leaders, when we’re in this highly volatile environment, one impulse is to retreat into safety toward a more stable environment, but I encourage you to think that maybe that’s not a reasonable goal. Instead we should focus on creating a culture and ethos that embraces the change.” — Aditya Agarwal at the ICONIQ Growth Engineering Summit.
Wartime leaders who have grit and empathy and can help embed this attitude into the overall company culture are essential in building resilient organizations. Companies can also create stability by ensuring people feel like the work they’re doing is aligned with the organization’s goals and that they are being recognized and rewarded for such work. Simple processes like streamlined OKRs that are reported on during each meeting and tracked over time can be a great way to build organization-wide visibility.
As your organization scales, it becomes important to collect data to understand job satisfaction and whether people think they’re doing work that aligns with organizational goals.
Rather than relying on subjective and irregular feedback from one-off conversations, developer satisfaction surveys allow you to see how things are changing among hundreds of engineering employees. At one engineering organization, leaders found that developer satisfaction wasn’t tied to things they expected (e.g., code base) and, as a result, helped eliminate all ambiguity thereby allowing leadership to set the roadmap appropriately. Developer satisfaction surveys can also be a helpful way to track productivity and efficiency to determine the situation is indeed getting worse or better longitudinally.
Remote engineering organizations tend to see stronger organization health. However, managing hyper-distributed teams also presents a unique challenge.
Over the past two years, distributed workforces have fundamentally changed how engineering teams collaborate with each other and altered the key processes and tools needed to enable software development. For engineering organizations with distributed teams, this presents a unique challenge of maintaining connection and collaboration across geographic barriers.
Companies with a primary remote work arrangement tend to have a higher percentage of racially diverse employees as well as a lower annual attrition rate than in-office companies. However, our summit attendees expressed a shared challenge of managing hyper-distributed teams. While some companies have developed processes and policies to help navigate the difficulties—such as guardrails on scrum team size and structure, autonomous engineering pods, a focus on async communication, or testing during the hiring process on candidates’ remote work capabilities—many companies are still figuring out the best way to organize across time zones, how to manage teams across engineering, product, and design, and when in-person meetings are truly needed.
For deeper analysis and more data on these critical issues facing engineering organizations, including more detail on engineering benchmarks on spend, headcount, and tooling, please view our full report here: “Engineering in a Hybrid World"
Notes & Disclosures:
1 From the Engineering in a Hybrid World study summarizing engineering data collected from a survey completed by certain ICONIQ Growth portfolio companies and others in the ICONIQ network in September 2022