Skip to content
492: Defining value within your team episode artwork

Feb 3, 2026 32 min

492: Defining value within your team

from The Bike Shed Hosted by thoughtbot

Sally and Aji assess some common metrics for success when working a project and how they may not always provide the clearest picture of how things are going. Together they discuss how to communicate effectively with stakeholders who are less technical to fully appreciate certain decisions and choices being made on a project, as well as the different metrics you can use to better reflect success and setbacks on a project. — Your hosts for this episode have been thoughtbot’s own Sally Hall and Aji Slater. If you would like to support the show, head over to our GitHub page, or check out our website. Got a question or comment about the show? Why not write to our hosts: [email protected] This has been a thoughtbot podcast. Stay up to date by following us on social media - YouTube - LinkedIn - Mastodon - BlueSky © 2026 thoughtbot, inc.

Episode details come from the podcast's RSS feed.

Embed this episode

Ready to play

492: Defining value within your team

0:00 32:52
of matches

Transcript Auto-generated, may contain errors

Hello and welcome to another episode of the Bike Shed, a weekly podcast from your friends at Thoughtbot about developing great software. I'm Sally Hall. And I'm Audrey Slater. And together we're here to share a bit of what we've learned along the way. So, Audrey, what's new in your world?

Well, I think I need to report back on something that at the beginning of the season, I think it was I had started maybe trying to make a switch over to VS Code. I used them in the terminal for years and years and years, but I had gotten sort of fed up with the amount of configuration of the tool that it sort of required. And so, try it out on VS Code. I had given a couple of running starts at VS Code over the years, mostly because I wanted to smooth over hurdles of robust friction for pairing, right? For folks that don't use them and don't use my them specifically.

But this is all come crashing down. I'm back in the terminal. It lasted longer than usual, but I finally just hunkered down and fixed the configuration thing that was driving me up the wall because there's just something about the closeness of the terminal that I really both find very satisfying and also just gotten so ingrained into my workflow that it made me think too much about how to use the tool rather than just getting text to come from brain, right? I solved that problem by using VS Code's built-in terminal. Yeah, that's what I thought I could do, but the ways that I like to interact with the file system and get our all terminal-based programs that kind of have a VIMFERS key binding methodology that are just so easy for me to use at this point that it was too big of a hurdle.

And I knew I could get there, but the years worth of quickness that I've built up would again take years to get there. And there are also just some things in GUI programs like Ruby Mine or VS Code that just can't be easily keyboard navigated to and are just built for the mouse like that's what those programs are. And it's not my preferred way of interacting with my system as I'm working. So it's an amazing program. It keeps getting better every time I dip back into it.

And I recommend no one start using VIMF these days. If it seems like something that your brain would jive with then sure try it out, but VS Code is an amazing editor. Maybe I'm just too stuck in my ways. Yeah, I don't use VIMF for this reason. I don't go camping.

I have a perfectly good bed at home. Yeah, I started with Adam was my first, well, like a sublime technically, but Adam and one of the first things I did that got any kind of eyeballs in an open source sort of way. It was like two or three dozen eyeballs, but still was an Adam theme that made everything like blocky and looked like the terminal. So every unit of UI was a cursor size block and everything looked like that. And it got downloaded, I don't know, a hundred times or something like that.

But I was surprised because I thought this is just me. You know, when else is going to want this, apparently, at least 99 other people want it. Yeah, the internet has shown me none of the things that are weird about me are unique. They're all still weird, but they're not unique. I love that.

Everything has a set filter. There's a whole slew of people that are really into whatever niche thing you can find. And I absolutely love them. Yeah, my first experiences with coding were in college. We used Emacs and I didn't really know that there were other things available.

And so the first time I saw like a GUI, a pretty fine IDE, I was like, where has this been all my life? I'm never going back. Right. So what's new in your world? I've had the past few weeks been doing something that I think is fun, although another context is horrible, which is revisiting code I wrote a couple years ago.

So I'm having the opportunity to do some like improvements and fixes and new features on an app we started I think two years ago. And this is like the second time we've done some updates to it, which, you know, is one of those trusted frustrating situations where sometimes you can be like, who wrote this? Oh my God, it was me. But it's also nice because when we finished, which I use air quotes for business, there's still a lot of things that I knew weren't exactly the way I wanted them. Mostly performance.

It's a really complicated data model and a lot of different like filter views and reports and all of this stuff. I knew that there was just like a flaming dumpster of N plus one queries in the depths. But you know, for a variety of reasons that were like, you know, people in time are what makes software hard, right? We could all write perfect software if there were no users and we had infinite time. I didn't get to address those nearly two degree I wanted to the first place.

So I've been able to go through and use a couple of different tools and find all of those N plus one queries and then find other little things that were going slower than ideally they would and have been making those improvements. And it's, you know, some of those, it's like, well, I know that this is a problem. I know that this is an N plus one query and that I can see that it's, you know, taking longer than it should. But is that the problem? Like, is that what's actually making it difficult for the users?

Right. And so it feels really good to like, you know, do a bunch of little tweaks here and there and have it add up to a much faster experience for the user because sometimes you do a bunch of little tweaks here and there and then it's the same or got a bit worse. Eco loading is not always a good idea. So I've been doing that and it's been really, really satisfying. Nice.

How many of these things would you say are things that you had kind of identified or sort of felt that there was something going on there that you could come back to and how many of them are new explorations or seeing your code with a new set of eyes? It's about half and half. Okay. I knew I was walking away from a dumpster fire and plus one queries, sadly, which I don't know that maybe sounds worse than it was, but it felt like that. Yeah.

It's an app that like, I think I've been the main person working on it, but there have been a few periods where somebody else like jumped in for a couple of weeks and did some quick improvements or bug fixes or feature upgrades. So there were also some fun discoveries like model callback that when implemented, it wasn't clear that that model was being updated and saved like 10,000 times in a background job and that background job was wrapped in a transaction. And the callback was refreshing a materialized view. So it was like 10,000 times refresh this materialized view that has like 10,000 rows in it. But also, I haven't dug, I wanted to dig more into this.

Also, this is all wrapped in a transaction. So like something somewhere is keeping track of all of these. That sounds an awful lot like 10,000 squared. So that was a fun one to dig into and find. And then very satisfying, discovered what was going on was able to fix it and have that go like 400% faster or something like that.

That's a good number, 400% faster. It was. Yeah. That's the kind of thing that gets stakeholders attention. They don't know what an N plus one query is, but they do know that this job takes, it's 400% faster than it used to be.

Right. Right. Absolutely. There's a lot there about, you know, sort of knowledge transfer and these kind of quick projects that come in and out and have small lifetime of different engineers working on them and just kind of by the nature of the business or any number of different things kind of lead to this situation that you're kind of by necessity losing a bunch of tribal knowledge regularly. And yeah, I could see how there would be plenty of foot counts for that sort of thing.

Yeah, there's only so much you can put in the read me and there's only so much you'll read in the read me. So how has that been working with you say that your stakeholders obviously noticed this 400% difference, but how has it been kind of revisiting those and seeing the changes with their like, well, first of all, let me ask, is the person you're working with, are they technical? So they don't necessarily know what an influence. They are technically, I guess, non-technical, they're not a developer. I don't think they've been a developer in the past that we're working as a project manager.

They do know a little bit about SQL. They can, you know, poke around in a database, more technical than the average of the street project manager, but not a developer. Sure. So it's interesting to communicate with them about like what they see or notice and need and what changed and then how do we evaluate whether or not that actually changed. Right.

I can tell them that we have, you know, eliminated 80% of N with less one queries, but you don't know what that means. Right. So that's the metric. Yeah. Yeah.

And so it seems like there may have been some easy things to look at in this like page load time or time that it finishes. What are some of the metrics that you like to lean on when you've got metrics or, I don't know, concepts that you lean on when you're working with someone that doesn't speak engineer? I think you really have to go back to like, what is this person spending money to do, right? Because they don't hire a developer or a consulting company because everything's exactly the way they want it to be. You know, so I think it's important to sort of, I don't know, I like to keep that in mind.

Like, why did they come to us? What are they actually helping to get out of this engagement? What will feel like value and success to them? Right. And so, you know, everyone's reviving up, digging into the things that truly are just a helper experience or things where it's like, this could become a problem, but it hasn't yet.

And I'm going to make sure it's not a problem before a problem, which is super valuable and great. And often your stakeholders don't care, you know, unless they've been in a situation before where they've seen that thing become a problem, they're like, okay, yeah, you're just, you're the men in black who just saved us from aliens and I'm just going to keep going on my day. And that's really just not that. That's something that we have to do a lot, right? And it is not only just communicate the things that we've done to non-technical stakeholders, but also help to evaluate and prioritize what work is going to successfully address needs and value propositions, et cetera, et cetera, that aren't necessarily directly linked to code architecture or something like that.

Right. Like, I know the number of hours I worked on something is not a valuable metric for how much value it has, right? You can fix a really complicated problem quickly if you're lucky and you can spend a whole week on something that doesn't really matter if you're not lucky. Yeah. So AI has taught us that the lines of code is not a valuable metric.

Well, I think we all knew that already. Yeah, I think the main metric for evaluating things like this is problems solved. Okay. Or problems prevented, right? So ultimately, it doesn't really matter how fast the page loads alone, right?

If the page loads really fast, but it doesn't do the things they needed to do, who cares? It doesn't really matter what beautiful architecture I've concocted in the back end. What matters is like the person who uses this app in their day-to-day job, are they able to do it more quickly? Or are they able to do it with fewer errors? Even are they able to just have a more pleasant experience while they're doing it?

I like that metric of problem solved. It's that feels like it could be just as nebulous as some other kind of concepts or whatever, but that actually, I feel like it can be quantified in a way, right? Like, that's something that people can think about. Like, I used to not be able to do this or this used to error every time I did that. And as those things go away, you can point back to that and say, oh, yeah, we fixed this together.

You said you had this issue. And we worked on why that was an issue. Sometimes you don't even know why it's a bug until you've had someone who knows what they're trying to accomplish. Explain it to you. I think there's another metric which is money, right?

If we're building something where the whole purpose is to get people to sign up and give us money for it, then that's a pretty straightforward metric. Have these changes I made caused more people to sign up and give us money or more people to continue giving us money rather than canceling or whatever. That's not least favorite metric. Yeah, I think it all comes back to what is the value that we're creating. And you're right, the value is not the lines of code that we have produced.

It's the systems, the functionality, the things that people are able to do that they weren't able to do before something that we built. And keeping that front of mind is really important. And I think it's very easy to fall into doing something very engineering because it's really satisfying, engineering-like. That might not deliver the kind of value that you need in that particular moment. Something that I think can be tricky is it can be tempting to say that the value is sort of the number of features or the number of new features or the number of things it can do.

But a lot of times that's not actually valuable. It's satisfying to the developer. It's probably satisfying to the product person who dreamed up the feature. But you're not actually solving any problems with that feature or nobody uses it or it costs more money than it makes. No matter how beautiful and usable and well engineered that feature is, it didn't actually add any value.

Customer money that it makes didn't actually add any value. Why are we talking about AI again? Dang. That loops back to problem solved, doesn't it? There have been plenty of features that have been developed that seemed great from the product team side of things.

I'm including myself in that. I'm not saying. I'm not saying that either. That don't really get used by the users. And it turns out, the things that were sort of anticipatory solutions without having the problem first aren't really as strong as they could have been otherwise.

Which I think is probably the core of why startups are both exciting and terrifying. You're trying to solve a problem. And hopefully the problem actually does exist. People care about it and what you're doing fixes it. But those are three big ifs.

You miss one of those. You're not going anywhere. Yeah. That's the value of the pivot though. It turns out this isn't actually a problem anyone has.

How can we take what we've already built in the right way? Yeah, but this other thing that we accidentally solved, kinda selling. Right, exactly. A little bit of what we're sort of dancing around is not only how do we talk about what we build and what we do with less technical folks, but also how do we hope that they will evaluate our work too, right? Mm-hmm, mm-hmm.

Which is such a hard problem. Like I'm a founder, I have this great idea and all of this domain knowledge, and I'm gonna hire all these engineers to build this dream I have. How on earth do they, I know that they're doing a good job. Yeah, and there are certainly cases where there could be a lot of work that goes in that isn't yet visible, isn't understandable from the other end that it was the amount of work that it was, right? And that's when it's easy to point to metrics like lines of code and things like that, when that's not actually going to give you a lot of like quality data to know if things are on track or not.

Yeah, this doesn't mean you think of something, I think Jared has talked about this, possibly when he was talking to you about the president's doctor, but the idea of like sometimes you just wanna do something, you know, it probably makes the most sense to you to spend the first couple of weeks setting up infrastructure, right and getting things integrated and doing like all of the like boring basic setup stuff that you know you're gonna need, you know you're gonna need a database eventually, I'm gonna set it up now, right? But sometimes for the client or the stakeholders, what actually makes the most sense is, I'm gonna build a quick and dirty UI that dumps things into and out of a Google sheet because that's value for you. And you can see what we're actually building. And then I'm gonna go back and turn the Google sheet into a real backend or turn the fake login into a real off system or.

Yeah, exactly. Especially when you're still dealing with those three or however many big ifs, right? Like is this actually needed? Is this actually feeling the need? I think it is, then it's more value to do the quick and simple thing that you're gonna iterate on or throw out later.

Yeah, and that, you know, sometimes there is, the value is throwing things away, right? Like it feels so bad to completely remove something you worked hard on from an app. But sometimes that's like the most valuable thing you can do is like, oh, we've proven that this doesn't actually do anything. And so now it's just this big fat pile of code that we have to maintain and deploy and run. All code is a liability.

I forget who said that once. Maybe that's a, we were talking about best piece of advice that we, other thought about it myself were talking about that was the best piece of advice that you've gotten. And I feel like something that I keep coming back to is all code is a liability. And when there's more code, there's just more possibility for things going wrong and changes that are required to do that all important pivot. And so starting really and starting really small is vital.

And ruthlessly editing out things you don't need. Ruthlessly editing, yes, very good. Man, it can be so tempting to comment something out or rapid in some sort of conditional, just in case. Especially, you know, and I feel like we've talked about this before, but especially version control just to lead it. You can go find it later.

This actually is making me think, I feel like we've been talking about the developer perspective on this and the project manager or product owner stakeholder perspective on this. But I keep thinking about one of my favorite groups are lovely designers. Because I think one of the most valuable, well, there's so much value I think about working with the designer. But one of the things I love about working with the designers I thought about is the way they can test and validate things so quickly. And prevent us from building those things that actually did not give any value.

Right. You know, a lot of times the hidden value in something like that is, you know, there's the value of building the right thing. So a lot of value in not building the wrong thing. Not building the thing that one wants. Which feels like a lazy developer's way of taking credit.

Like I wrote no code. How valuable are I? That's not what I mean. Yeah, you can't just flip that metric on its head and say like how few lines of code did you write and then do a more valuable? I wish, but that's not going to look.

I don't know. That's also a thing that's really hard to communicate the value up with like a non-technical stakeholder or like a first-time founder or somebody who's never built software before. The idea of like, I know you feel like you just spent all of this money in order to eliminate all of these things you thought you were going to build. It's like a waste of time and money. But really you've just saved all of the time and money you were going to spend building it, supporting it before you discovered that it was useless.

Right. Just because you can't measure an increase in productivity or revenue or whatever. If everybody who works for you is so much happier using the new tool than the old tool, did you create value there? I would say yes. How do you put that into the communicable terms?

Would you say that it reduces turnover in your staff then maybe? Yeah, it could increase retention. Yeah, there you go. Increase employee engagement. That's the word I hate people use.

So I think what it really boils down to is always like, it depends. There is no one metric, right? And you've got to talk with people, talk with users, talk with stakeholders. It's that communication and that shared understanding that is the real value. It's the friends we met along the way.

It is, as it turns out, once again. Yeah, it's what motivated this person to ask you for help? What do they think they're coming for you to fix? And what are they actually coming to you to fix? Right, because sometimes that's something that you have to kind of explore together and find out what's at the heart of the need, especially when there are folks that might come first with a solution.

And it is kind of an implicit or unspoken problem that you need to understand what the real problem is before you can create a lasting and really meaningful solution. Yeah, like are we putting a band-aid on a symptom? Or are we actually addressing the root of the problem? Because it's a lot easier to see the symptoms when they're terrible. And when you don't deal with the code, right?

So you're not a super technical or a, I don't love the categorization of non-technical, but that's a whole different conversation. If your person is not working in the code, it can be easy to see the symptoms, but maybe difficult to see the root causes or the connection between the symptoms. And so, if we treat all the symptoms, but don't think about the root cause, they're probably short term going to feel like you added value. But I like to think that really digging into it and treating the root cause instead is going to be much more satisfying for everyone. Now my brain's coming up with this extended metaphor about medicine.

Right, right, hit me with that 11 metaphor. I go to the doctor, I have a fever and my phages. So I get Tylenol and foot cream or whatever. And I'm fine, excellent. Yay, I go home.

I take the Tylenol and the foot cream every day. I don't have a fever, my foot doesn't hitches. Itch, you can tell from this example that I obviously have a deep medical background. Right, but then eventually my left thumb starts twitching. And that's weird.

And we don't actually have a good way to address that as a symptom. There's no like anti-twitch powder you can put on your thumb. And we dig into it and it turns out, my liver has turned to blue. And the symptoms of that are fever, foot itching, and thumb twitching. So now we can treat the liver turning blue and all these other things go away.

And then my right eye being slightly bigger than my left eye, I didn't even realize it was a symptom and now that's better. Right, so if you come to us and you say, I want a loading spinner on this page cause it takes forever and people keep refreshing it. And they think, you know, cause when something's slow, sometimes you refresh it. And then that just, you know, and Q is a bunch of, or it just makes it take longer, right? I don't know what the deal is, but it makes it take longer.

And so you put a loading spinner on there and then people know not to refresh. And it still takes as long as it took before, but it's not taking longer for all the refreshes. Right? And then they're coming and saying, you know, this thing was marked closed, but then it got modified anyway. And you're not supposed to be able to do that.

And so I don't know, you put some callbacks or something to wrap around and like, prevent something that is marked closed as being modified. And then you run out of database, your database like gets full or gets, you know, CPU use gets really high and everything kind of just crashes down. And you realize like, oh, every time we load this page and accuse a background job, and if you load it again, it gives a second one without even checking to see it's already been acute. And then this is the developer version of like 99 bottles. It's like 99 bugs on the board.

You take one down. Now there's 101 bugs on the board. Yeah, yeah, yeah. The interesting thing as we're kind of getting into this, talking about the sort of root thing that once again, it depends, rears its ugly head. And there are times when the value that's required or the bigger value delivery is going to be the symptom patch in the moment, right?

Because it's going to solve. Sometimes you just need to tell them. Exactly, yeah. The blue liver example was making me think of an application that I had made a new version of, right? It's something that a particular agency was using for 15 years, whatever needed a new version, right?

And so we wrote a brand new thing. And as the folks in the field using the software were coming to us, they were describing the things that they do in their work. And it was all of it was taking time and all. It was like using a footprint because they were describing all of these things that were workarounds for the old software that had been layered on over 15 years. And even the folks that we were talking to never knew, some of them never knew the job without the workarounds.

And so that was the job, right? And so that was actually a lot of work to get into their day to day and understand what their business was, understand what they were trying to accomplish and how software can make that happen more easily, more straightforward. And by digging down to that root cause, which in this case wasn't the cause of a bug, it was what is the thing you were actually trying to do, not the thing that you've been forced to do with the tool you have. And I think every time we got to something and really understood what it was and got around those workarounds and gave them something that just went from A to B instead of A to D to F to X to 23 to B, that was really transformative.

And that was the most fun experience in that project for me is being able to just lift all of that workarounds off. Yeah, and probably what you built ultimately was a lot less complex than what they asked you to build, I'm guessing. Oh, 100% nearly across the board, right? Yeah, so if you're just looking at, you know, the features delivered or the lines of code or the sophistication of the code, that would look like you didn't really do anything. But in reality, you made everything work better, you made all those people more happy.

Yeah, and so much of that work was not writing code. It was, to me getting with the people sometimes shadowing them in their work. And a lot of this work that you were talking about that designers do, there's so many user interviews and all of that kind of really deep understanding of the business of the value of what was needed to get done. A lot of the times it was a fight to be able to ask those questions, it was like, they just wanted the same thing but faster. And we could have given them that, but we were lucky enough to have one enough trust in our relationship with this team by solving problems and kind of building that momentum as we went, that they afforded us the time to do that investigation as we went along.

And it's understandable if their problem is this takes forever, that they would be very hesitant to like, oh, let's have these people who are doing this job takes forever, take more time out of their day to talk to our developers about the job that takes forever. That's really hard to ask. It is, it is, yeah. And it was a very, very stressful job that they were doing as well. So I was like, yeah, looking back on it through that lens, they were all saints to build the best way that they did.

You're busy as people on whom all of your value depends. Can I just talk to them for like an hour to have? Can I pull them off the floor for 90 minutes? Just stop everything. I'm just gonna chat.

But it turned out to be the right thing to do. Yeah, I think that's what makes consulting so much harder than a job where you're sort of just in a ticket factory. Mm-hmm. Is that it's not just like the input is they ask me to build something and the output is that I built something. It's the input is that they ask me to build something, but the real input is that they have a problem.

Yeah, and those asks for what to build are, it's a personal interpretation of what their problem is and what the solution might be. And also very often we'll say that sometimes what they've come to us asking to build is what we end up needing to build, right? Right, I mean, that happens a lot. That's not a rare situation. No, but it's less interesting to talk about it.

I think that's what makes consulting really fun for me as well is that it is, it's a little bit like we were talking about that every aspect of humanity or everything that someone does has a whole niche subculture behind it, like literally everything. And I love about consulting that I get to go to so many different businesses and get to spend time having this deeper understanding of what it is that they do so that I can be the best problem solver for them. And I get to explore all these different little worlds that I wouldn't get to see if I was just working for one company. And so that's something that is really kind of satisfying for me about being a consultant.

Absolutely. I feel exactly the same. Also sometimes you feel like a detective. All the five being a detective without crime. Yes.

Yes. It's probably a different episode. I'm not known. Do we wrap up? Yeah, sounds good.

Frequently asked questions

How long is this episode of The Bike Shed?

This episode is 32 minutes long.

When was this The Bike Shed episode published?

This episode was published on February 3, 2026.

Is there a transcript available for this episode?

Yes, a full transcript is available for this episode. You can read the complete transcript on the episode page.

Can I download this The Bike Shed episode?

Yes. Use the download control on the episode player to save the publisher-provided media file.