Startup Life/Opinion/ How to scale global remote engineering teams Scaling teams internationally in the age of hybrid and remote working presents both risks and opportunities \Startup Life UK visa endorsement fees will now cost founders £2k — down from up to 15x that By Zosia Wanat 24 March 2023 Startup Life/Opinion/ How to scale global remote engineering teams Scaling teams internationally in the age of hybrid and remote working presents both risks and opportunities By Nicolas Baranowski Friday 17 February 2023 By Nicolas Baranowski Friday 17 February 2023 Startups are increasingly looking beyond their front doors for international engineering talent amid the rise of remote working and hybrid teams. Wage arbitrage is another draw. However, managing remote teams comes with its own challenges, such as simultaneously managing multiple time zones, office locations, cultures and skill sets. I’ve seen companies get it right — and seen it go wrong. Here are some considerations any startup should have in mind to scale remote engineering teams successfully. Stay close to your engineering team No matter what configuration you choose (full remote versus hybrid, micro-offices versus no offices, contractors versus employees), always try to stay as close as possible to your engineering team. Have one of your cofounders based where most of your engineering team is, or at least fly there once a month. We’re two cofounders at Frame, and one of us is based in Georgia with our engineering team. This allows us to speed up communication and product development velocity, and facilitates the creation of a strong company culture early on. Avoid agencies In the past, many companies outsourced product management work to remote agencies. Agencies would allow them to get started quickly by providing experienced engineers and a professional product development process. But today, hiring, training and managing remote engineers has become more and more accessible. “That said, I would still recommend agencies — but for specific skills like branding, security, or SEO” On top of the extra markup you’d pay (varying between an extra 50-100%), working with an agency can cause multiple challenges. To list a few: lack of internal knowledge building, risk of the agency hiking up fees during development, weak company culture and less goodwill from VCs when it’s time to fundraise. That said, I would still recommend agencies — but for specific skills like branding, security or SEO. Understand compensation To be competitive, try to standardise your offering as much as possible and know the market you operate in. Developers are in high demand, so don’t hesitate to be open on numbers (like cash and stock options) early in the discussion, so you’re not disappointed later on. Typically, developers fall into one of four markets when it comes to compensation. From the most expensive to the least: startup hubs in the US, startup hubs in the EU, non-hubs in the EU and US, and non-hubs outside the EU and US. For a defined skill (like a React developer) and level of seniority, you will typically see a 20% drop in cash compensation and a 10% drop in equity cash compensation as you move from the first category to the fourth. Deel can be helpful to benchmark your offer compared to the market you want to hire in. Be organised from day one Run a tight hiring process. At Frame, we have four stages of selection: a Linkedin screening, a CEO interview, a lead backend interview and a frontend interview. Since we’re quite early-stage and the CEO is technical, we try to shield the programming hours of our engineers as much as possible. Hence, the CEO interview comes before the lead engineer’s interviews — but this will inverse as we scale. We use our own app to manage the hiring process end-to-end, enriching common notes on candidates at each interview, which typically last 30 minutes each. Candidates are asked about their skill balance, level of seniority (less than two years is considered a junior profile, two to four years is medium, and more than five years is senior), and favourite technical languages and frameworks. Lastly, we ask the candidate to showcase something they worked on recently. Bonus points are given if the candidate can show a personal project. When it’s time to make an offer and sign the contract, be 100% digital: For template job offers, we recommend Pitch For compensation simulation, we recommend Google Sheets For contracts, we recommend DocuSign For a global payment system, we recommend Deel For stock options, we recommend Carta Align on team standards, tools and processes Aligning early on a set of operating standards, processes and tools is key. Agreeing on things like working hours, internal communication tools, product management tools, in-office days, product management frameworks and recurring meetings like retros, check-ins and sprint planning sessions will get you off a great start. Don’t think of these standards as set in stone, but as a common base that you can continuously improve over time. We recommend Ramp to streamline the payment of software-related costs. “Team members are expected to be online and in a professional environment during the two blocks of four hours of work” At Frame, we all agreed to work in the office Monday to Wednesday from 9am to 6pm with a break from 1pm to 2pm, and the rest of the week is remote. Team members are expected to be online and in a professional environment during the two blocks of four hours of work. I recommend WeWork to offer a professional and globally accessible office to your remote teams. Implement an effective production system. We recommend a by-the-book Scrum methodology with two-week sprints. Then, try to come up with a point system (like the Fibonacci suite with a Planning Poker) to load your sprints according to the actual capacity of your team, taking into consideration developers who are off for the next sprint. Have clean data and reporting systems Your engineers should be focused on delivery. We recommend Linear to manage your product backlog. Having a clean backlog will allow you to extract almost real-time completion analytics — you can use Google Data Studio to build analytics from a Google Sheet. When it comes to measuring productivity, there are two schools of thought. You can either develop a quantitative velocity measurement system (for example, an average story points per developer per week) or keep a goal-focused approach (like shipping a sizeable feature each month). A weekly demo is generally a good control mechanism to ensure the entire team delivers regularly. Fix mismatches fast While it’s relatively straightforward to test hard skills, testing soft skills and mindset is much more challenging. A weekly demo is your best control mechanism to test your engineer’s constant productivity. When it doesn’t work out, the demo is the first place you’ll start to see things breaking down. Typically, it will be due to a combination of hard and soft skills mismatches. While fixing hard skills is sometimes possible, fixing a misfit with your company culture will be almost impossible. Keep in mind that product development costs will typically represent 70% or more of your costs. So, when it doesn’t work out, I’d recommend terminating the relationship quickly — especially in these challenging times where every dollar counts. Nicolas Baranowski is the founder and CEO of Frame. Related Articles Founders flee to Estonia’s digital paradise By Christina zur Nedden Click here to read more How Surple puts the office on power saving mode Sponsored by Google For Startups Click here to read more How to nail the hiring process as an early-stage startup By Miriam Partington in Berlin Click here to read more What it takes to scale up European tech By Marie Mawad in Paris Click here to read more Most Read 1 \Startup Life UK government to reform ‘equity for visas’ residency application system 2 \Fintech Is Revolut really worth $33bn right now? 3 \Startup Life Techstars unexpectedly pulls out of Sweden mid-programme 4 \Deeptech The other funding gap: it’s not just unicorns that are leaving Europe 5 \Deeptech ‘There’s going to be a bloodbath’ — is generative AI a bubble?