Scaling Decision-Making at a Fintech Startup
James Trunk, VP of Engineering, Griffin
👋 Hello! This is Suzan. I write about leadership and interview leaders from tech companies like Etsy, Github, Google, Netflix and Spotify who share their insights on how they overcame a challenge.
Subscribe to get the edition each week. If you enjoy the content, please like this post -- it helps me focus on the content you prefer.
Alright, on the interview with James Trunk, VP of Engineering of Fintech Griffin. Enjoy!
Scaling decision-making is critical for startups to grow in size and organizational complexity. Unless we consciously decide to adjust how we make decisions as we shift into the next phase, it can become cumbersome or ineffective. James Trunk, VP of Engineering at Griffin talked with me about how they scaled decision-making at the Fintech startup. I think you’ll learn so much from our conversation.
We talked about:
How a 1:1 was the catalyst to shift decision-making
Why a self-reflection practice is so powerful
How values and principles can clarify decision-making
Why context is so important when it comes to creating organizational frameworks
The framework they use today and how it’s changed him as a leader
James’ advice for other organizations
Can you introduce yourself and give a bit of background?
I’m the VP of Engineering at Griffin. It's a people process-related role shepherding everybody in the engineering team. I've been doing that since early 2021. My background is in doing similar leadership roles at other startups. I have worked in bigger companies, but a bit further back in my career. Originally I was a software developer, and then I went the software developer team leadership and the CTO/head of route so that was my journey to Griffin.
What was the shape of Griffin when you joined?
We were significantly smaller than we are today. I think I'm right in saying that the engineering team had seven developers, one product manager, and one designer when I joined the company. We were split into two teams. Now we're more than 30 across the whole of engineering and product split across seven teams so quite a growth journey so far.
Is there anything else about how the structure of the team has changed that might inform our conversation about decision-making?
When I joined there was no product leadership so I essentially had two hats to wear – both VP of product and of engineering. But that changed over time as we grew and we hired people at that level. That meant I was making decisions that affected the process for both engineering and product. We’re also a remote-first company. I'm based in Stockholm, Sweden. We have engineers based all over Europe while the majority of the team is UK based. That affects how we collaborate, how we brainstorm and it affects decision-making.
Today we’ll be talking about how you all consciously created a decision-making framework. Before that, can you tell me how decision-making worked before this process?
The word that comes to mind is unilateral. Alan, our CTO and co-founder is extremely strong from a technology engineering perspective, you know, genius level. When I got to Griffin, a lot of the technology, tooling, and how-we-work decisions went through Alan. I also tended to be more unilateral in terms of things that were related to my role. So a lot of decisions bubbled up to me and Alan. It's not like no one else ever made decisions – we had a very friendly, open and transparent culture from before I joined the company. But within engineering, decisions tended to bubble up to Alan and me.
It sounds like centralized decision-making while considering other people’s viewpoints.
We definitely did. Before I joined, we used ADRs (architectural decision records), which are a method of collaboratively keeping track of design decisions, directly in the codebase. One of the things we established after I joined was to enhance the decision process with RFCs (requests for comments).
The basic idea of an RFC is that you write a document that describes a problem as clearly as you can and then suggest solutions to the problem. That was how we made changes that were technical in nature. Process changes tended to come through our retros. There's always been a retro culture at Griffin. After a retro, you experiment by taking away action points and making changes to the process, or to the platform with ADRs and then through RFCs. That was the initial proto version of how we did decision-making.
What were the indications that the team had outgrown the decision-making process?
I do this daily reflection practice. I keep a work diary, or a journal if you're American. I have a note from a day where I had a one-to-one chat with one of our engineers and he almost made a throwaway comment about being in an RFC decision meeting and feeling a bit frustrated with it. He’d been in a meeting where it’d gone around in circles. There weren't clear goals for the meeting to get to the decision. People were disagreeing and everyone came out of it feeling a bit frustrated. So we were brainstorming in our 1:1 what we could do about it. We decided to take it away as an action point to look into best practices and understand what other companies are doing to see if we could find anything we could steal to tweak our RFC process. So that comment turned out to be the catalyst to set the rest of our decision journey into action.
Because I did a daily reflection practice, I wrote it as a challenge. That’s when I realized that there’s also a problem with Alan and I tending to make a lot of decisions unilaterally. I thought maybe we could try and solve two problems at once. So how do we make decision meetings more effective and also diversify and scale decision-making so that we don’t have a bottleneck?
I love that you have a reflection process and that it allowed you to connect two pieces of decision-making to tweak.
Yeah, I don’t think it would have happened if I didn’t do that reflection. I guess most people's days are like mine where it's chaos and information is coming left and right. If you take the time to stop and think through what's happened and what it means, sometimes you can catch things.
I'm curious about your reflection practice. You write in your diary every day, right?
Every day, yes.
I'm going to get really detailed. Morning, night, or does it vary?
I’ve been doing it for more than a decade now. I keep notes during the day. I used to have the energy at the end of each day to write them up so I wouldn't consider my working day finished until my work diary reflection was complete. In more recent years, I found that my job is so intense that I get quite tired that at the end of the day, I just can't do that. So what I do now is combine it with my weekly reflection. So it becomes a combination of daily and weekly.
I do my reflection practice every morning. I do morning pages where you just write three pages of whatever’s inside your head. It helps me get clear on what I’m thinking about. I like the creative ideas that come out of my brain. I think if I were in an organization, I would do something more structured.
A decade is a long time. How did that get started? What made you start a regular self-reflection practice?
I watched a TED talk about the benefits of journaling. It was a woman who had done research and she gave some examples about Oprah and a few other people. I thought, “Oh, I'll give that a try.” I didn't think it would stick, but it did. I found it really useful. I used some basic format that they recommended where you keep track of the different decisions that you've made and the progress that you've made. Progress is good from a motivational perspective and decisions from a reflection perspective. Then I’d look at an insight moment – any idea that someone had shared with me that I wanted to remember – that goes in. And then challenges. What have people shared with me today? What am I worried about? You think a bit about the future and what's coming up. So that challenge is part of the daily reflection I find incredibly useful.
Thank you for sharing that. I think any sort of reflection practice is essential for leaders. Whatever that looks like.
I agree. I agree.
I do my reflection practice every morning. I do morning pages where you just write three pages of whatever’s inside your head. It helps me get clear on what I’m thinking about. I like the creative ideas that come out of my brain. I think if I were in an organization, I would do something more structured.
That's awesome. I get that. I'm quite a structured person. So it fits how my brain works. But I get that. I think you're completely right that people should figure out the way that works for them and how their brain works.
I recently added a section to my reflection – How am I feeling? What are my emotions? I added it because I felt like I wasn’t good at connecting with how I was feeling and I wasn’t using those signals to help me think about challenges.
In the beginning, my vocabulary was very limited. I just knew, well, I'm a bit happy or feeling okay. I wasn't very good at describing how I felt. Over time, because you practice, you get better at figuring out, well, how do I feel? What's the minutiae of how I'm feeling? And where might that be coming from? Also super valuable.
I love how you added feelings into a structured process. Facts are great but feelings can add just as much helpful information.
Sometimes I still struggle, if I'm being honest. I know there's a word that describes how I’m feeling, but I can't quite find it and then I'll just be lazy and pick the broader term. But I'm getting better at least.
How has adding the section about feelings changed things for you?
I feel like there's often a connection between the challenge and how I'm feeling. Not always, and it's not always so direct. Sometimes you only see it when you look back on patterns. I think you've written about this before that part of being a leader is keeping on top of yourself, keeping yourself healthy, and understanding where your mindset and your mental health are.
If I notice a few days in a row where I'm saying things around being tired or frustrated I ask myself what can I do about this? So I noticed it earlier and tried to get myself out of it. Whereas when I didn’t track that, it's not information. You're not getting any kind of feedback, right? So you can't adjust. So that's helped.
I recently added a section to my reflection – How am I feeling? What are my emotions? I added it because I felt like I wasn’t good at connecting with how I was feeling and I wasn’t using those signals to help me think about challenges.
That’s a great practice. It sounds like the thing that allowed you to slow down and recognize that there was something that wasn’t working with the decision-making process.
Yeah. Exactly. Because I'm interested in decision-making, I've read books about it, I've read books about cognitive biases. It wasn't a new topic for me, but I didn't realize how other people figured this out. We wondered, are there clear decision-making processes in other organizations and what do they look like? Dominic and I went looking. He found that the IETF, the engineering task force, had this really interesting idea that sparked us both to push toward this democratic decision-making process with rough consensus. It was an interesting idea. I hadn't heard about that before. They’re voting on big changes to internet standards. So how do you have a process that can work in that kind of situation when you have a room of really smart engineers? It's very similar to what we're doing with our technical decisions.
Rough consensus is the idea that you don't need total consensus. You don't need everyone to agree that option B is the best option out of options A, B, and C. But you do need to get rid of all of the fatal flaws with all of the options. So if you say, well, option A has a fatal flaw, then you know you can't do that. And if option C has a fatal flaw, you know you can't. But if you're left with, say, two. If there are really strong arguments one way or the other, then maybe you can convince enough people that, okay, we have rough consensus around B, somebody makes the decision, we go.
That way you're not stuck in analysis paralysis, which was a definite fear I had about consensus-based decision-making. It's kind of famous for being slow because you've got to get everyone on the same page. So we connected that idea of rough consensus with one of the values we have at Griffin, which is to disagree and commit. It is the best idea we've come up with together. It has the strongest arguments behind it. Let's do it, let's all try it. And if it doesn't work, we can always try something else in the future.
How did you all begin to shift the decision-making process? What was your goal when you began to do that?
We have this idea called 5% Feedback at Griffin. The basic idea is we want a feedback culture. We want to give both redirecting and reinforcing feedback. Reinforcing is you did a great job. Here’s how it was good. Redirecting is I was expecting this behavior or that outcome and I didn’t see it. How were you thinking? Those are the two main kinds of feedback we’re aiming to give each other.
When you do it 5%, very early on, when it's just a few bullet points, it feels so much easier psychologically to give that feedback. It’s how we try and improve early feedback. The other benefit is that you’re getting buy-in on change early so it’s not a fully finished, packaged-up idea. You get really useful feedback. It's easier for the team to give feedback. They feel involved and engaged.
Dominic was part of the whole inception of the idea. We pulled in other people and got it to a point where enough people were saying, “Yeah, this looks like worth a try.” Then I wrote a page in Notion. It was a quick decision flow chart to show how it would work. Then I brought it to people and said, “What do you think about this? Can we give it a try?” We sold it as an experiment and decided to try to it at the next decision meeting. When you sell it as an experiment, people are more willing to give it a try. If it doesn’t work, we’ll try something else. So that helped.
We tried it and it worked pretty well. So we kept developing it from there. At Griffin, no process is set in stone. We’re always learning. It's a process to add to it, tweak it, reflect on it, and keep changing it. So that's how the ball got rolling.
When you do it 5%, very early on, when it's just a few bullet points, it feels so much easier psychologically to give that feedback. It’s how we try and improve early feedback. The other benefit is that you’re getting buy-in on change early so it’s not a fully finished, packaged-up idea. You get really useful feedback. It's easier for the team to give feedback. They feel involved and engaged.
Did you have goals for shifting the decision-making process?
Solving those two problems that I identified that day was the goal. I didn't know if we could, or if this would do it. You never know when you set off on a journey, how far you'll go, right? Looking back, it’s accurate to describe what I was trying to achieve was how do we be more effective? So we have fewer of those decision meetings that people feel frustrated at, that take a long time, that aren't effective.
We need to be fast, but we need to make good decisions. How can we have enough processes to guide us? How do we reduce the friction for doing the right thing and increase the friction for doing the wrong thing?
The other thing you’ve talked about is democratizing decision-making. Was that one of your goals at the outset or was that more organic?
Yeah, that’s an interesting question. It wasn’t really a goal. The best solution we could come up with seemed to solve both. We weren't aiming for democracy. We were trying to have effective decision-making and reduce unilateral decision-making. So it was the goal but it was the solution. The democratic part, the rough consensus idea from the IETF, was part of it. But we took more parts from other places to Frankenstein it together.
A big part of the democratic bit turned out to be the anonymous part of the voting. When we initially did decision meetings, there was nothing anonymous about it. We're in a meeting, you could see (how people vote). But when you're remote first and you have asynchronous chat like Slack, like we have, you could take advantage of the fact that you can do an anonymous poll.
We found that people didn’t see how Alan or I voted because it was anonymous, right? It created a more psychologically safe environment than if Alan shared his opinion openly because no one wanted to stand up against the genius CTO. So it was a way of protecting against that. It’s the same with me. It’s like democracies, everybody gets a vote. If it's a problem that affects the whole of engineering, then anyone can write an RFC telling everyone to look at this comment, and we’re going to vote. There's nothing in the process that stops that. That's very expensive. So what tends to happen is the problems tend to have a smaller impact zone than everyone in engineering. The people who are impacted are automatically the ones who vote. It almost becomes more like local politics. Instead of a general election, it's a local election.
Sometimes we do big things that a lot more people are interested in, and it just naturally gets bigger. I'm going to comment because it's all transparent. There's nothing hidden. But normally it's, “Ah, I see that RFC, but that doesn't affect what I'm working on, so I'm just going to carry on with what I'm doing.” So it just kind of naturally scales to the size that makes sense and the people that should be. involved in both the idea generation and the voting are all included.
We need to be fast, but we need to make good decisions. How can we have enough processes to guide us? How do we reduce the friction for doing the right thing and increase the friction for doing the wrong thing?
What are the core principles you lean on?
The Griffin core values are a big part of why I joined the company. I'm a strong believer in principles, virtues, and values. It's a nice way to guide your decision-making. I’ve been trying to figure out how to make more effective decisions for a long time. I landed on this decision principle idea that you mentioned. The three that had the most impact were: long-term over short, reflection over perfection, and feedback over guesswork.
Let's take reflection over perfection as an example. I can strive for perfection in myself too much. I need to spot when I'm doing that and realize, “That's not the answer.” The answer is to reflect and get a little bit better. If I can get 1% better every day, that's better than kicking myself because I didn't do that thing perfectly, or I communicated poorly. It's a way of helping myself spot the problem or the potential problem.
The insight behind long-term over short-term is that I don’t think humans are always good at long-term thinking. It’s so easy to get trapped in these sorts of short-term loops. That helps me remember that it’s the long-term positive outcomes we’re going for, even if they come at a short-term cost. With democratic decision-making, the short-term costs – you've got to research it, figure it out, you've got to write it – there's a cost to coming up with it, there's a cost to selling it. You could argue from some perspective that there's a sort of power cost for Alan and myself or maybe other leaders that if you're giving away this power, that could be considered a cost. But when you think about the long-term, you realize that isn't. Those costs are so small in comparison to if we can make our decision-making more effective, if we can make higher quality decisions, if we can scale our decision-making as we continue to grow, this works. The value of (scaling decision-making) obliterates any sort of short-term cost thinking.
And the last one is feedback over guesswork. The whole decision-making process is almost a feedback loop in itself. It's a way of getting more information than you would have had access to before, there’s less guesswork. One of the things we do is we use this kind of nerdy thing called decision matrices – basically, a spreadsheet where you list all the criteria that help you make the right decision. But just by doing that, it's a way of opening everybody's eyes and getting us all on the same page about what problem we're solving. And what are some ideas we have about how to solve that together? Imagine driving a car. You would never drive a car with your eyes closed because you need that feedback. Now I wouldn't make decisions without having this kind of structure in place because I feel like I'm missing information here. I'm missing an opportunity to get other people's ideas and to come up with maybe a better solution.
There’s so much good in there. How does reflection over perfection look at the organizational level? Or are these all personal?
These are my personal decision principles. These are not common things across Griffin. We do have other values and behaviors that overlap. We share a core value around kindness and how we treat each other. If I'm expecting you to be perfect, then I'm putting unrealistic expectations on you. If I can say, “Oh, that didn’t turn out great. Should we think about how we could do that better next time?”, then we learn from our mistakes together. It creates a nicer atmosphere, a nicer environment where people want to be because they feel like, well, I'm growing here. Rather than these really strict expectations that I'm never going to fulfill.
The Griffin core values are a big part of why I joined the company. I'm a strong believer in principles, virtues, and values. It's a nice way to guide your decision-making. I’ve been trying to figure out how to make more effective decisions for a long time. I landed on this decision principle idea that you mentioned. The three that had the most impact were: long-term over short, reflection over perfection, and feedback over guesswork.
I love that. Are there any philosophical principles that underpin the model as an organization?
No, that is a good idea though. I will think more about that. I like underlying things with principles because this process will change over time. Capturing what we’re trying to achieve, why we’re doing something and our values can help guide (when people leave or leadership changes). In the same way, my decision principles guide me, hopefully, towards my values, we could maybe do a similar thing. So you've had a good idea that I'm gonna think more about it.
One of the things I do in my workshops is help leaders identify what their operating principles are and ensure that they’re communicating them and that there are values and principles for their team. If you decide to do it, I’d love to know.
Can give an overview of what the decision-making process looks like?
It’s six steps. I can whiz through them. The first step is to identify a problem that's big enough to require a process like this. Not all problems go through the process – it has to be big enough. Concisely and succinctly as you can describe the problem – that's a core of the step that you don't start with a solution, you start with the problem. It's the problem first and getting everyone's buy-in on that. And then we start. Step three is collaboratively working on different solutions. Normally it works that the first person who writes the RFC has a go at their preferred solution or maybe several. Others will dive in to add, suggest, and change. It’s a collaborative process because that's how you get more ideas, better ideas, and hopefully to a better decision. I talked about the decision matrix. That's an important step. That's the final artifact out of that work that hopefully gives you a clear document that tracks how we are thinking at the time. And it's sometimes so obvious just from that, you don't even need to have the anonymous vote. It's just that one, it ticks all of our boxes. Those aren't as good. It's just done and everyone's on the same page. Sometimes it's a bit less clear, then we do the anonymous polling. If we're not at consensus, we do the decision meeting and try and get to rough consensus. We randomly assign someone to drive the meeting so that it's not me or Alan again and people learn how to do that. Then you can get to rough consensus, disagree, and commit. And that's it. It sounds like a lot of steps, but it works well in our context. As you said, every context is different for our context it seems to be working well. I'm sure we'll tweak it in the future.
I know you’ve written about it so we’ll be sure to include a link. One brilliant thing is the criteria. I find so many times that we’re trying to make a decision but we haven’t had a conversation about the criteria we’re using or what we’re optimizing for. We go in circles.
Yeah, exactly. What’s important to us? With Engineering decisions, that's so key. Sometimes that's the discussion that unlocks the agreement and the consensus in the group because suddenly people realize, “Oh, you were thinking about that criteria. I didn't even think about that trade-off. Oh, you're right,” Suddenly the whole thing gets unlocked. If you don't have that shared discussion around those important things for the decision, you lose all of that. You can just waste time. Uh, like you're saying, just going around the houses.
Right. We go 10 rounds on a thing without agreement. I think it's part of why we get into analysis paralysis, we get stuck, or friction and conflict arise. I wanted to point that out because it's quite brilliant and it seems obvious, but you'd be surprised by how often people get stuck because they haven’t had the conversation.
That wasn't our idea, by the way. That was a stolen one as well. Everything here sounds like it's not new. It's just us taking good ideas other people have had and matching them together.
Yeah, I think that that's how things work. Collaging a few things together can be effective. I worry when people take the whole thing and make it ours because their culture and incentives could be different. I think it’s best when we do research about what else is out there and think about what’s right for us.
Completely agree. Completely agree.
There’s a feeling of autonomy and involvement in the decisions that we didn't have before. And there's clarity because the process is laid out. When you join, you can see it. Well, this is how we do it. This is why we do it.
What changed in the team and the culture, or even for you as a leader, as a result of this new process? What do you notice is different now about not the decisions that you make, but sort of the gooey people side of it, the org side of it?
I do 1:1s regularly with all the engineers or skip levels. I wonder if because this is my baby people won’t give me 100% honest feedback about it. There's a risk that I'm having a slightly more rosy view of this, and I'm slightly biased. So I just wanna start with that. Putting all that aside, the feeling I get is that people are enjoying it.
There’s a feeling of autonomy and involvement in the decisions that we didn't have before. And there's clarity because the process is laid out. When you join, you can see it. Well, this is how we do it. This is why we do it.
Nobody's come to me and says, James, this is awful. I think it could be that sometimes (the process) can feel a little heavy for some decisions, like if you make the wrong choice early on and you start this kind of process, but it turned out that, oh, it was really obvious we should have just done A, that people feel like, ah, was that the best way to have spent our time? But I think that's just part of the learning as a group of people – how do we use this tool? When you get a new tool, you need to learn when to use it and how to use it.
In terms of how it's impacted me or how I view work, it's been a bit of a load off. And not because I think that Alan and I were making bad decisions before. I think we were generally making good decisions.
It feels like now we're unlocking the power of all of these really smart people that we've hired and we're giving them a chance to come up with ideas and vote on what they think is the best one. I feel more relaxed about the decisions that we make. It feels like there's a higher chance that we're going to be making better decisions. And I like that. I like making as good a decision as we can, given the information that we have. My shoulders relax and I feel calm about it.
I love that because leaders hold so much responsibility on their shoulders. It can feel heavy. Democratizing decision making, I get that it would take a weight off your shoulders, that it's not only on you to make those decisions.
Yeah. I can see how it can almost be seen as a dereliction of duty in a way because we almost expect leaders to make decisions and now suddenly... we're making slightly fewer, right? There are some decisions that we're not making in the way that we did before. And from one perspective, I guess that's true. But I think the trade-off is worth it. There are still times when me or Alan will come in and that does have to happen sometimes. So it's not like we've completely said, we're never gonna make any decisions again, not at all. It's more like, there are certain decisions that we think are made better with this tool and this process. So let's give that a try.
I don’t see it as a dereliction of duty at all. I think it depends on how you define what it means to be a leader and what is our role. I think we could also argue that our job is to make sure we get the right information and consider all the factors.
Yeah.
It feels like now we're unlocking the power of all of these really smart people that we've hired and we're giving them a chance to come up with ideas and vote on what they think is the best one. I feel more relaxed about the decisions that we make. It feels like there's a higher chance that we're going to be making better decisions. And I like that. I like making as good a decision as we can, given the information that we have. My shoulders relax and I feel calm about it.
What advice might you have for other scaling startups who need to scale their decision-making?
I think you summarized it perfectly earlier – don't just copy-paste someone else. I think we have a good idea here, but I wouldn't necessarily recommend that people just copy-paste it. Or at least if you do that, do that with the idea that you'll try that, but you'll tweak it to match your context as quickly as possible.
The other way that they could do it is to do more of a first principles approach like we did look around and see. Is there a better way than this? Is there some other way? Maybe the democratic part of this isn't as valuable as we think for some reason. If you figure out an even more effective way that leads to better outcomes, I would love to hear about it because we will steal it and include it in our process. I think your idea about not just copy-pasting is key. I think you put on something really insightful there and we didn't copy-paste this from somewhere else and I don't think other people should copy-paste it from us.
Thank you for your support of this substack and the leaders featured! You can read all of the past interviews and learn more about how Constellary supports companies and leaders here.
If this piece resonated with you, please let me know and give the heart button below a tap.