Buffer is a growing software startup that began with employees scattered around the world, then opened up a San Francisco office, and ultimately decided to revert back to an entirely distributed/remote setup. Even after creating a well-functioning mixed office+distributed team experience, they still decided the office benefits were not worth the cost.
In the mid-2000s, I also was part of a company that shifted from remote, to partially remote. The Electric Sheep Company was a leader in virtual worlds in the mid-2000s with about 85 people at its peak. After operating with an entirely distributed team, we found ourselves with a growing critical mass in New York City. We decided to open an office, but we had less success with a mixed environment than Buffer. Even though we still used virtual tools, we quickly saw a split between information-haves, who had the benefit of informal “watercooler” conversations, and information have-nots. Pulling the entire team together three times a year was not enough to compensate.
So what is the right way to work for your company, and how do you decide?
In this post, we’re going to cover pros and cons to 4 different models of work. While it is possible to get mixed models working, as Buffer states, our conclusion is that entirely co-located, or entirely remote, is to be preferred.
Martin Fowler broke our four basic models of work in compelling Oct 2015 piece, Remote versus Co-located Work.
- single site (everyone together)
- multi-site (two or more locations with co-located teams)
- satellite workers (most of the team is together, but a few people are remote)
- remote-first (everyone works from a separate location)
Here’s our take on the four approaches:
In an ideal world, this is the best approach. People form relationships and bonds that improve motivation and collaboration. Communication is fast and informal. The team can focus less on documentation and more on action. It is easier to keep everyone pointed in the same direction (i.e. understanding the why, not just the what, behind your business). This model requires less duplication of management, is more resilient to leadership change. Lastly, it is easier to catch personnel issues.
The downside to the single-site model is recruiting and retention. If you are outside of the major tech hubs, the talent pools are so small that your growth options are really limited. However, within the major hubs, there is fierce competition across employers (from startups to consulting firms to larger companies) for the same talent. This will loosen if there is a tech downturn, but it has certainly been a brutal recruiting environment for the last 5 years. It not only takes months of effort and expense to hire someone, but it can be harder to keep them as well. Your decision about single-site might hinge on the depth of your talent network.
I should add that if you are startup that is still trying to find its way, it is almost impossible to work in any model other than single-site. Things are way too chaotic, and if you can’t sit down with your teammates face to face, you will move way too slowly and with too much confusion.
This is where you have two or more sites with major teams. Common variations are 1. different offices for different major regions that have their own requirements (North America versus Europe, for example), or 2. a mix of “onshore” and “offshore” labor for cost savings and/or time zone coverage. Multi-site has the advantage of critical mass at each location, giving you some of the efficiencies of single-site. However, these efficiencies are best reaped if work can be split across location, rather than having everyone pooled onto one task. At Neo, a product innovation consulting boutique, we ended up staffing client projects with mixed teams across our New York and San Francisco offices, simply to solve some of the us-versus-them feelings that brewed between those cities. It helped with morale, but impacted productivity even with simple things like a 3 hour time difference.
There are downsides. The locations can easily slip into an us-versus-them attitude, which usually can only be fixed by bringing people together periodically and taking on the expense of letting people travel between the offices. The other downside is you have to double up leadership, and it can be complex and expensive to find leaders who share the right values, strategic alignment, and capabilities.
This is the hardest model to get working. Once a group of people are co-located, it is pretty much impossible to *stop* the quick, informal communication from happening because it feels both more productive and is easier. But that inherently means that satellite workers are cut off from a huge amount of the knowledge sharing and alignment that happens. People become isolated very quickly.
Buffer overcame this because they had a significant number of satellite workers, and they made a very concerted effort that, “There are no advantages for people who come into the office, no disadvantages to staying home to get your work done.”
This model can work is if you have a likable, independent, and experienced teammate who is willing to periodically come to the mother ship. It takes independence and maturity to be on the outside and *not* let the feeling of being an outsider eat at morale and thus productivity. It also takes being likeable for the members of the co-located team to make the extra effort needed. But the challenges are steep. Martin Fowler even wrote, “I hardly ever hear of this model without the satellite person getting increasingly detached.”
Pushing the team to bring most communication onto digital tools, whether chat, longer-form interactions like Jotto, video-conferencing, shared documents, etc. will make this easier.
Remote-first is emerging as an increasingly viable model, given the quality and cost of digital tools. You can hire great talent. Being outside of the major cities, talented individuals are often cheaper and easier to retain. Frankly it is better to have a fully distributed team of A players than a co-located team with B players, who will demotivate the As and ultimately cause them to leave.
Because everyone is in the same boat, you don’t see an us-versus-them attitude forming. You *can* see misunderstandings form. Fierce debate, while healthy, should be handled over video or phone. You need to bring everyone together into one physical place a few times a year, but the cost savings of not having office rent should cover this expense.
A big bonus here, as just mentioned, is cost. The money spent on real estate can go towards many things — another pair of hands, better salaries or benefits, higher marketing budget, better team retreats, you name it.
Communication is equalized — there are no information have and have-nots. Because there is no proverbial watercooler, everyone is forced onto communication platforms to collaborate. Everyone understands the need to document and share decisions and the drivers behind those decisions.
One thing about remote-first, that could be viewed as a positive or negative depending on who you are: it requires careful attention to individual and group accountability. You can’t let accountability be too informal or loose, but rather need clear, transparent, and concrete documentation.
One of the biggest challenges with remote-first teams is mentoring junior workers. It is not impossible, but you need a junior who is highly motivated and has solid professional maturity, as well as mentors willing to go the extra mile. We agree with Martin Fowler’s statement: “Be wary about having junior people join remote-first teams, certainly don’t try it until the remote-first team is working smoothly and then only add junior people very slowly.”
As we noted above, remote-first can be easier in many ways than the mixed models, but it will probably never be as efficient and effective as having everyone physically together. It all depends on your trade-offs.
Each of these models has pros and cons, and your choice will depend on context. Each model requires effort to get culture right and to build shared identity and shared understanding within and across teams. Hopefully this post can help you think about the model that works for your business.
We call Jotto the “shared understanding company” because these are the problems we want to help solve. Please check us out at www.jottohq.com.