Scaling Servant Leadership

Gareth Wilson
Glitch: Make Better Software
10 min readNov 30, 2016


In this Tech Talk, Jude Allred, CTO here at Fog Creek, explains the methods he uses to scale Servant Leadership. He explains why relying on rapport doesn’t work and describes a number of techniques that you can employ to successfully apply servant leadership across multiple teams and departments.

About Fog Creek Tech Talks

At Fog Creek, we have weekly Tech Talks from our own staff and invited guests. These are short, informal presentations on something of interest to those involved in software development. We try to share these with you whenever we can.


  • Rapport Doesn’t Scale
  • Techniques to Scale Servant Leadership
  • Starter Engines
  • Diminishing returns in consensus
  • Decision making as a service
  • Setting up and Promoting Channels of Communication
  • Servant Leadership as a Constraint


What follows is a lightly edited transcript of the talk:

Servant leadership is, to me, a set of principles that mostly come down to you want decisions to be coming from the most informed people and that’s usually the people closest to the problem, not people overseeing the problem. Therefore, you want your leaders to be focused on removing obstacles from those people and giving those people the context so that they can make those decisions effectively. My third point is that you want the individuals to have really high autonomy so that they can be creating initiatives and acting on them, and then the other edge of the sword is you then depend on that.

It’s not enough to just give people autonomy, and then when they don’t use it, take command and do stuff anyway. That’s probably the real challenge — if you give them the autonomy, and then you depend on people executing on that autonomy and taking that initiative in order for your organization to move forward. Not every situation that can come out in management fits neatly into one pile or the other. In incredibly tactical situations or under extreme time constraints, where you have a dedicated expert and you just want everyone to follow through with that because let’s say you’re being hacked. You have one person on point countering the hacker and the other people supporting them. You can be a little bit more command and control. You don’t need to be absolutist in saying, “We must solve this with servant leadership.” That said, I would feel mighty proud if I was able to. This is what servant leadership is for me.

“You want decisions to be coming from the most informed people”

Rapport Doesn’t Scale

When I was a 2-month-old team lead at Fog Creek, I started coming up against a communication challenge that was new for me and was invalidating my previous strategies. My favorite way to communicate still, but I can’t use it all the time, is to have such a strong rapport with someone that we can trust that we’re building towards the same thing and then we can communicate at high bandwidth as engineers. We’re sitting next to each other, or at different computers, but building the same thing, and we all know we share a specific goal or a thing we’re trying to solve. If I see a way you can do your job better, I can just tell you, and that’s not abrasive. That’s pure constructive criticism and vice versa. You don’t have to coach that. You don’t have to be delicate. You can just trust that you’ve built that relationship, and then you can leverage it to communicate quickly and efficiently.

The other part of this that I like about having that really high bandwidth communication, is it gives me more social permission to sit down with someone and share a bigger blob of context. This takes a lot of time, but if there’s a challenge I’m trying to solve and I want your help with it, I would much prefer to be able to sit down with you and explain all the reasons why it’s a challenge in the first place, why people in the organization care, the people who are being affected, the things that can be made better. Just all the bits of the story except for my conclusion.

I want you to see all the context, and then one thing that could happen is that you then come to the same conclusion that I do. My confirmation bias jumps in, but if that happens, cool. We’re on the same page, and you should be able to advocate for it just as much as I can because no part of your representation of it is we should do this because ‘Jude said so’ or because ‘Jude thinks so’. It’s going to be because we look at the same set of information and you came to the same conclusion. Furthermore, you should now care about it as much as I do, and any difference there is going to come down to maybe we have different value systems, maybe we have different risk tolerances, or most frequently, maybe we just have different information, which is the most interesting one. If you came do a different conclusion than me because you have different information from me, then we can both learn. I love it if I can do that.

“Try to communicate context over conclusions”

In general, this is me practicing my take on an important tenant of servant leadership in the team lead role. I’m not giving you my decision as a lead. I did the legwork of getting you the information to make a good decision, and I want your recommendation. If I don’t agree with it, I investigate more. I don’t push. Through this and perhaps fortune or some mismatch of happenstance, I never had a strong disagreement with any of my team members.

So that was my communication technique as team lead talking to individuals on my team. I wanted that across the board. That’s what I learned. I learned to do this and it worked very well for me, but now I’m a lead. And when you’re a lead, you’re not just communicating with the folks on your team. You are communicating up to someone coordinating other leads, and you’re communicating across to the other leads.

This got me in trouble because there were, even in our small size, there were people in the organization with whom I hadn’t and never would spend enough time with to build the rapport to have the social permission, to have social high bandwidth communications. And even if I did, having that depth of communication with all the leads can be really time inefficient. Rapport doesn’t scale. At some point, if you move into management and as you get further and further from direct reports or simply as your organization gets larger and larger, there will come a point when you will never, ever actually be able to have rapport with all of the people who it is essential that you communicate with effectively, so my strategy had to change.

“Servant leadership at scale has to be indirect”

Techniques to Scale Servant Leadership

Servant leadership at scale has to be indirect. My servant leadership as an individual contributor or as a team lead could be very direct, and then at higher levels, I can’t have those direct channels. I have to do other things. Those other things manifest more as communication structures. I construct predominantly for other people, or for feedback to come forward, or for structures in the organization that makes it easier for people to take initiative or act autonomously.

Starter Engines

One technique is starter engines. One thing that is a recurring pattern that I see is one where there’s a hole that needs to be filled and you have people with autonomy and the ability to fill that hole, but they’re just standing on the edges looking in, not actually solving it. There’s a number of reasons why that can happen, and there’re a few things that you can do to overcome that. One that we’ve employed a number of times is like the thunderbolt from Zeus. This is generally Joel saying, “No, let’s just do this thing,” and then we do it.

There’s a nuance here though that’s very important. At the start of last summer, Joel helped us build a starter engine. This was a marketing plan of just development tasks for Fog Creek. The items on the plan were not really of consequence. They could have been. Indeed, many of them were capable of being of consequence, but the point then wasn’t to fulfill the items on the plan. That’s a step. The starter engine here was here’s a list of initiatives that you can do, and all of them are targeted around moving the needle of the company, and we’re optimistic that some of them will. There were 17, I think. Some were long horizon. Some were short. Some we had much more confidence in than others, but they were all experiments.

The deliverable here wasn’t executing on all of those things, although that would have been great. The deliverable was creating the habit, so that the next set of things could be done without needing someone like Joel to come in and say, “Here’s some things to do.” So now the engine is running. Another way to look at it is if you already have a habit for a thing, it’s easy to continue that habit, whereas if the habit is absent and then you want to do a thing, you have two problems. First, you need to get up the energy to do the thing. Second, you have to get up the energy to start the habit. If you’re looking at promoting an intrinsic motivator, the more that you can do build the habit, then the less of a natural objection there is to take the next step. In a grander sense, I view this as structures, which assume success.

“Some people hear servant leadership and they think that means full consensus. That is false.”

Diminishing returns in consensus

Consensus is great, but if the decision doesn’t matter, then at some point, you’ve got to say, “Well, if we’re in 80% consensus and the cost to us of being wrong is very cheap, then it can be more expensive for us to get to 100% consensus or 90% consensus than it is for us to just accept that we’re wrong.” Some people hear servant leadership and they think that means full consensus. That is false. If you can get it, great. If you’re operating with zero consensus on anything, then that’s a weird thing, but being aware of the diminishing return curve in consensus is pretty important because that last little bit can be most of your communication, and if an early assumption in the pipeline fails or new information comes in, then it can just be a complete and total waste.

Decision making as a service

Decision making as a service is a fun little servant leadership thing. Another thing that I’ve seen prevent autonomous initiative taking people from putting forward a decision is that they’re just nervous that they’re wrong. Maybe they’re only 90% sure of something and maybe you’ve been with us for 6 months and you’re afraid that we’re going to make a big company decision based on your opinion and then it’s going to be wrong and you’re going to get fired. You wouldn’t be, but you can still have that nervousness, and also you don’t want to feel bad if it doesn’t go right. But, if it’s a decision that needs to be made, and you’re the absolutely the most informed person on that subject and you have the best recommendation on that, then we should go with your decision.

So a thing you can do in that situation as a servant leader. If you’re a team lead, although you can do this for your peers as well, is to let the decision come from them but then back it up and take the blame yourself. Then I can champion it, and I’ll give credit where due at every opportunity, but if it goes wrong, I’ll catch it. I’m being the manager. I’m being the leader. This is on me, and that’s good and that’s fine because we’re doing this together. Sometimes we’re going to be wrong, and that’s okay. If we can alleviate the fear of failure from the people putting forward creative, inventive ideas, then you’ll get more creative, inventive ideas. Another type of decision making as a service is when a decision just freaking doesn’t matter. “It’s blue, moving on”. I’ve seen this ruffle feathers, but I’ve also seen this save a tremendous amount of time.

Setting up and Promoting Channels of Communication

The theme of all these points, I’m talking about service leadership, but I’m also really talking about communication. The thread of all of these is setting up and promoting channels of communication that help people engage, and it’s all iterative.

“I’ll give credit where due at every opportunity, but if it goes wrong, I’ll catch it”

Servant Leadership as a Constraint

One of the big distillations I have for myself is the view of servant leadership as a constraint. It’s very easy to come to conclusions and then put them forth or give strong directions and commands and whatnot. But, it’s less effective most of the time. The only time I think it could be effective is if I happen to be the most informed person on the subject, but that in itself is a failing because the people working on the subject should be the most informed about the subject.

For me, I view servant leadership predominantly as a constraint upon myself, which pushes me to be more effective by not giving up and falling back to just giving commands. I don’t give orders. I’m instead trying to suss out the best decisions or the best directions and bring them to life in the organization, and I’ve talked about some of the techniques.

But remember, it’s a spectrum. If you want to lean more into servant leadership, then just pick a part of the spectrum and do so. As someone in management, I would say try to communicate context over conclusions, and to scale that — build structures that promote it.