Skip to main content
Paula Farrugia
  • Knowledge Sharing
  • Communication
  • Transparency

Single Point of Failure: Five Habits for Sharing Knowledge in a Remote Team

8 min read


What I believe about knowledge sharing in remote teams, and the habits I put in place so my team can keep moving without depending on any one person.

My team grew from 3 developers and a tech lead to 12 developers, split into three teams of four, and we kept getting blocked for the same reason.

The problem: all the knowledge sat with one person

A team that depends on one person's knowledge stops whenever that person does. In my team, most of what we knew lived in our tech lead's head. Projects weren't documented, so he had to onboard every new joiner, unblock everyone and mentor them too. He also had to review every high-risk change.

When he wasn't available, the team couldn't progress. He was forced to check Slack on his time off, and other teams were blocked waiting for us.

How we worked was the problem, and he wasn't to blame. Because every answer came from him, developers rarely talked to each other, nobody challenged the answers they got, and few people took part in discussions. There was little sense of ownership or motivation.

Documentation that new joiners could read would have helped a lot. But the bigger change had to be in how the whole team shared what it knew every day, including what people did with the information they got from the tech lead.

The fix: sharing knowledge with purpose

This is harder in a remote team, because remote work takes away the casual encounters of an office: the quick question across a desk, the idea shared over coffee, the colleagues' discussion you overhear and jump into. Nothing replaces them automatically. If information is going to reach the people who need it, someone has to choose to share it.

That's why I ask my teams to over-communicate. By that I mean making knowledge visible to everyone in two ways, rather than documenting everything or messaging constantly.

Documentation is selective. Something is worth documenting only if someone else will need it in the future and it's necessary for the team to work. If people can easily figure it out, or it explains itself, it stays out. Over-documenting can work against the team too. If someone asks a question that others will likely ask later, it's worth documenting. If not, a reply is enough.

Messages are for transparency. They keep the team informed, and they let people contribute and give their opinion at the right time.

The aim is a team where people are less dependent on one another. People still work together, but no single person blocks the rest. A team is stronger when knowledge is shared widely, and that only happens when everyone puts effort into sharing it. To make that happen, I put five habits in place.

The five habits

1. Default to public

Unless something is private or sensitive, it goes in a public channel. A question asked in a DM gets answered once, for one person. Asked in a channel, it's answered for everyone who has the same question, now or later. Ideas follow the same rule: share them! An idea posted in a general channel gives the team a chance to brainstorm and build on it, just like sharing thoughts over coffee in the same office.

In practice, that means almost no DMs. When I have a question for one person, I still ask it in the public channel and tag them, so the rest of the team has visibility.

2. Keep one source of truth

All documentation lives in one place. This is the most important piece of the whole approach, because it's what lets people be self-reliant. Asking people to search before they ask is only fair if the team actually writes things down, and keeps them where people can find them.

In practice, the structure matters more than the tool. Tools change, but a clear structure is what lets people find things, and a good structure can move to any tool. Here's the top-level structure that worked for us:

Team documentation/
├── Useful team links
├── Onboarding and project setup
├── Tools
├── Team processes
├── General documentation (all apps)
├── App documentation/
│   └── one section per app
└── Investigations (all spike documentation)

Slack discussions link back to the documentation, so people get used to going to the source rather than the thread. When someone asks me a question that was already documented, I answer it and link to where they could have found it. That shows them where to look next time, and over time people start defaulting to searching first.

3. Write down what someone will need again

Processes and features get documented: how they work and how they're built. So do decisions that affect people who weren't in the room, because someone who wasn't there might see a better solution, or a limitation nobody in the room thought of.

In practice, where something goes depends on what it is:

  • What a feature is and how it's built: it goes in the team documentation, including any important decisions taken during implementation that the team might need to refer back to.
  • A meeting changes something already agreed on, for example because of a limitation: the people in the meeting post a short FYI on Slack with a summary or the action items. Others can contribute without having to be in the meeting, and the reviewer and QA who pick the work up later get the context.

4. Write for someone who wasn't there

A message or document that only its author can understand defeats the purpose. The reader probably doesn't have the same context, so explain more rather than assume everyone knows the same things.

In practice, before a document is published, it gets one more read from the perspective of someone with no prior knowledge. Any unexplained "what's" get explained, and terms get simpler. I've also built a skill to help with this, which is on GitHub.

5. Own your loose ends

Getting a message to the people it's meant for is the sender's responsibility. Posting it isn't the same as it being seen. Requests to third parties work the same way: the request stays with the person who made it, and so does making sure it gets attention in time.

In practice, that means tagging the right people, using the right channel and following up when there's no reply. When all I need is an acknowledgement, I ask for a reaction to my message. Everyone leaves regular status updates on their tasks, as a comment on the ticket or a short note to the team on Slack, and always before a holiday or when they hit a blocker. That way, if someone is sick or on holiday, the team can pick up where they left off without being blocked.

Making it happen

Start with the problem, not the habits. When I introduced these habits, I spent most of the presentation on the problem and the why. Everyone had faced the problem themselves, so once we focused on what we were trying to solve, the team could lock in on it. Nobody disagreed, because they understood the benefit.

Only then come the process, the expectations and what everyone gets out of it. From there, two things matter most: making sure the habits are followed consistently, and having different developers take different parts of the same feature, so the knowledge never sits with one person and keeps moving. Because everyone saw the benefit, the team started calling each other out when a process wasn't followed, and enforced it themselves without needing me to step in.

It wasn't free. Documentation was extra load for about a month, until we had written down what was missing. After that, documenting in small increments was quick. Some of the shyer people still sent DMs. I encouraged everyone who got them not to reply there but to redirect the person to the main channels, until they got used to asking in public. The tech lead was fine with the change, since it took a lot of load off him. The public channels didn't get noisy: they had what the team needed to be efficient.

What changed

Shared knowledge proves itself when a team is most stretched: when new people join, and when people are away. For my team, the first clear sign came at the next round of onboarding. We onboarded four developers at the same time, and it went smoothly. Each team member could be paired with a new joiner and onboard them on all topics, instead of everything going through one person.

The second came during a high-stakes release. In the weeks before it and during the release itself, different people were off at different times. One developer was off for a week, and when he came back another developer was on leave, without a single working day of overlap. Our PO was off for two weeks during the release. Nothing was impacted.

The day-to-day changed too. I have visibility of all topics without micromanaging, and I can unblock the team without them having to ask. I also wait less for answers: when I post a question, anyone on the team can answer it, so I don't have to wait for one particular person. The team challenges each other and adds to each other's answers. People give their opinion and have a sense of ownership. And the team keeps functioning in different situations, without depending on any one person.

The goal includes you

The real test is whether the team can work without anyone, including its manager. That's why I practise every habit in this article first, where the team can see it.

As a lead, the most valuable thing I can do isn't to be the person the team can't work without. It's to make sure nobody is.


Further reading: LeadDev's guide to Slack etiquette

Need to solve the same problem on your team?