Startups: How to Build a Global, Remote Workforce from Day 1
Recently, my portfolio CEOs, current investors, and friends in the startup community have asked me about my thoughts in building a company remotely. There is a lot of material on the principles or the thought process behind why a remote workforce (especially a remote engineering workforce) makes sense or “is the future”, but this is less helpful from a practitioner's perspective. As a practitioner, you want immediately applicable tactics, not high-level paradigms. Prior to Dataframe (my current company, where I am CEO), I built an international Enterprise Data Science company with offices in Boston, SF, Berlin, Zurich, Tokyo, and Paris — all set up in a relatively short timeframe of fewer than 4 years. Through my experience, I have learned about what things work and what things don’t work in remote work.
I will focus on managing a team of technical “builders” — folks with competency in engineering and product design. I will use the term “engineer” to signify a person in this category.
This time, it’s different. Really.
I have observed that tips and tricks from the “OG” remote companies (think Basecamp, Automattic, InVision, GitLab, etc) do not apply well to companies starting out today because the world looks vastly different.
- The cost basis for US-based (even non-SV) engineers is much higher than 5–10 years ago.
- There is an abundance of venture capital for early-stage companies. Cash expectation for early-stage companies, especially as a competent engineer, is high and can actually be met in most cases.
- Markets are more competitive than ever (goodbye to raising a $250K seed round and getting to $1M ARR). Once there are signs of PMF for a particular market segment, competitors form within a 1–2 year timeframe. The PMF cadence is being continuously de-risked and there are more resources than ever for entrepreneurs.
- Product expectation is higher than ever from users (you need usability AND amazing design AND an amazing self-service motion).
- Big tech companies are growing faster than ever and compensating their employees more lucratively than ever. It’s hard to convince technology workers to leave their jobs.
- Video conferencing+hardware is dramatically better than 5 years ago. Zoom+commodity hardware really changed the game here. I could not have a proper international conversation on Google Hangouts in 2013 without my laptop overheating.
- In general, collaboration tools are abundant and cheap. This enables high leverage of an individual to do many peoples’ worth of work (and at a startup, you will do this).
Remote Strategy from Day 1
Bottom line is that as a founder, it’s fiscally irresponsible and impractical NOT to have a remote strategy. You need to be open to hiring engineers from multiple cities or even in other countries on day 1, not on year 5.
So how do you hire and manage a remote workforce in the “new normal” world?
Tactical Tips to adapt to the new normal
- More ownership to technical co-founders/early founding team members. Generally, it’s harder to convince competent engineers to leave their jobs, take a salary cut, and be in a less certain environment. You need to motivate a technical co-founder or early engineer with ownership: financial, and operational. Set up a higher ESOP pool, and give more trust to your team members.
- Have an outcomes-based culture (actually). Define what a unit of work is. Because it’s hard to monitor each others’ progress synchronously, you need to explicitly define what constitutes success (writing code, writing documentation, making designs). In the past, I have had trouble with employees spending time on “doing research” or “thinking”. While thinking is of course helpful and necessary, it should not count as a unit of work.
- Hire more senior, more competent, and fewer people. In the new normal, you have even less time to train less experienced employees. The communication costs and managerial overhead, especially in a remote environment (especially with time zone differences) will kill you. Pay up for the best, and hire fewer people.
- Document everything. Have a system of record for work. We use Notion for general notes and project management (any work gets documented here by default; every single meeting has an agenda and meeting notes). We had several iterations for the system of record for software engineering — we started with Asana and landed on Clubhouse. For design, we started with Sketch+Zeplin but ported over to Figma+Notion. I also personally use Threads to document thoughts that are more long-form that needs to live outside of Slack.
- Operate in 1 official timezone, but have at most 2 main timezones. This is one is very important. Oftentimes, you need to have synchronous check-ins (all hands, engineering meetings, etc) over Zoom. When talking about time, it’s helpful to have one “official timezone”, usually where the center of mass of the company is at (ours is San Francisco). When making a tradeoff in accommodating busy schedules (and you will), you have to prioritize the official timezone (e.g. 11PM PT meetings are in general discouraged). Secondly, you need to make a proactive choice in picking two timezones that make sense for you. Battle-tested insight: you can choose PT and ET; ET and CET; PT and KST. But rarely does PT, ET & CET work (e.g. SF, New York, London). As a point of reference, you see Western European startups have offices in Boston or New York but not on the west coast. You see Asian startups have offices in Seattle or San Francisco but not at the same time with New York or Boston. As you grow, you may have silo-able teams, in which case you can expand into many more timezones, but as a small team, it’s impractical.
- Be hyper-responsive; screen for hyper-responsive people. Some people need external stimulation to get work done. These people won’t thrive in a remote setting. A good proxy for proactivity is response time to messages and emails. More importantly, you need to screen for follow-through: communicating about getting something done, and getting it done on time (this is harder than it sounds).
- Structure your time for artificial breaks and brainstorms. Sometimes you need to artificially schedule time for brainstorming sessions. Explicitly separate brainstorming meetings with structured, get-shit-done meetings. You can’t operate just on tasks, you need time for creativity and brainstorming. If you are very organized, I believe you can re-create watercooler-sessions this way. Admittedly, this is very hard.
- Have a high bar for performance in a remote setting. The world has not adapted to remote work, and the vast majority of people are not ready for remote work. The vast majority of people also should not work at startups. Hence, the intersection set of people who are capable of working at remote startups is a small one. If someone is not performing, they need to go ASAP. When you have gathered enough evidence for the person’s non-performance, it will already be too late. Get good at reading between the lines and getting early signals (again, good proxies are response times and follow-through rate).
- Understand that remote work is still sub-optimal. Newer processes, tools, and experimentation will make remote work a lot more enjoyable in the future. But it still is suboptimal in a lot of ways. If you need social stimulation to be productive (like me), you will need to find alternate ways to find the stimulus (in progress). Prepare to be amazed by increased productivity/efficiency in certain domains (cost-effectiveness of employees, no commute time, freedom to work whenever). But also be prepared to be disillusioned by the downsides (lack of social interaction, decreased creative productivity, no boundaries between work and life).
Joseph Moon is a data scientist, entrepreneur, and investor. He is the co-founder and CEO at Dataframe.ai, a modern DataOps platform helping data teams solve problems together. Interested in learning more? Get in touch at email@example.com or follow me at @yosupmoon on Twitter.