ºÚÁÏÃÅ

Episode transcript

Ìý

Scott Shaw: Hello and welcome to the ºÚÁÏÃÅ Technology podcast. My name is Scott Shaw and I'm one of the regular rotating hosts on the podcast. I'm joined with by my co-host, Prem. I'll let him introduce himself.

Ìý

Prem Chandrasekaran: Hey, folks, I'm also one of the other regular podcast hosts here. Very glad to be here.

Ìý

Scott: We are lucky to be joined today by Abi Noda. Abi is one of the authors of an influential paper on developer experience metrics, and we'll let him talk about that. He's also the founder of a company called DevEx, which is, it's a unique product, I think, that assists with surveying developers and displaying and analyzing metrics from the delivery process. Again, I'll let Abi talk about that. Abi, do you want to introduce yourself and give us a little background?

Ìý

Abi Noda: Sure, yes. Thanks for having me on the show. Excited to be here. My name's Abi. I'm a programmer, researcher. I'm currently the founder and CEO of DevEx, where we help organizations measure, improve, understand productivity. I spent about seven years in this developer productivity Space ever since I first became a manager and then started a former company focused on developer analytics, which was acquired by GitHub and then at GitHub focused on this problem internally for a number of years. Spent a lot of time in this whole developer productivity problem Space. Too long, I'd say. Excited to unpack some of that today with you guys.

Ìý

Scott: Some of this came out of the work done at the Microsoft research lab with the Space framework. You've worked with Nicole Forsgren on some of this.

Ìý

Abi: There's been a story arc to all the developer productivity stuff, with Dora being, of course, something Nicole created almost a decade ago now, and then while Nicole and I both worked at GitHub, part of Microsoft at the time, the Space paper or framework was published right around when COVID had begun. Then Nicole and I still work together today, she's at Microsoft, but is also part of the research team here at DevEx.

Ìý

Nicole and I have published a couple papers recently, one of which is the DevEx framework, which I'm sure we'll talk about today, as well as a more recent study looking at the impact of developer experience, which was a joint study with GitHub and Microsoft. Nicole's been the continuing thread across all these different things. She's obviously a brilliant researcher and thinker in this Space, so it's been an honor to be able to work with her.

Ìý

Scott: I think bringing this, the rigor to research in this area, I think, has been really important. I'd like to get into that a little bit later, but let's dive right in because I get asked all the time. I visit a lot of businesses. I talk to a lot of CIOs and CTOs and engineering managers. There seems to be, have been for the last couple of years an enormous interest in measuring developer productivity. That's a term that always makes us a bit apprehensive when we start talking about developer productivity. In your opinion, why shouldn't we or should we be measuring developer productivity?

Ìý

Abi: I would argue it's always been a big area of focus. Every now and then it becomes spikes, with COVID for example, then now with GenAI. It's always been this elusive problem that we've all, I think, collectively as an industry and as business leaders, been trying to solve. It still feels unsolved, although I think we're getting closer. Why should you measure productivity? If you spend millions or tens of millions or hundreds of millions or even billions on R&D at your organization, you want to maximize that investment. How are you going to maximize that investment without getting an understanding of, A, how you're doing and, B, how to get better?

Ìý

I think that's ultimately where this problem always stems from. Whether it's a CEO asking that question or the board asking that question, or if you're a platform engineering leader or developer experience team that's trying to define your own charter and determine your investments, you need to be able to answer those questions as well. It's an important problem. As a result, it's no surprise that there's a lot of tension around it.

Ìý

Scott: I think there's a lot of misunderstanding, though, around productivity and that we use that word in a lot of different ways. We use it, economists use it in a very rigorous, specific measurement way. I feel like people are trying to take that and apply it to the field of software delivery, where we use it in a much more like we feel productive. It's a loose concept.

Ìý

Abi: Space, for example, the whole thesis of that paper was productivity is complex and it's multidimensional, and we can't just boil it down to lines of code or commits or whatever the output measure of the day is. A lot of organizations that we work with and I talk to are spending years, sometimes years, trying to come to an answer around how do we measure productivity? What even is productivity? Of course, we've published research on this and, to some extent, I think that's really advanced the discussion. At the same time, I also recognize that now there's just more perspectives and more angles to this problem, it only grows more complex. It's a fascinating problem.

Ìý

You mentioned the conundrum around how do we measure knowledge work and development work that's so different than other types of work, labor and even business functions? Peter Drucker wrote a paper in the 90s about this called, Knowledge-Worker Productivity: Our Biggest Challenge. I only recently came across that paper, but it blew my mind because all this frustration, all the dialog around whether you should or shouldn't or can or can't measure productivity, Peter Drucker was talking about this 20 years ago, this exact problem. It's encouraging in some sense validating to know that even this really is a difficult problem.

Ìý

Scott: I'm going to read that now.

Ìý

Abi: Great paper.

Ìý

Prem: I do have a question related to that. Our foundational beliefs came from Martin Fowler, definitely influenced a lot of it. He very famously wrote an article on his Bliki saying that you cannot measure productivity. What is your take on that? Is it possible to measure it? Is it not possible? Is it a good idea? Is it a bad idea?

Ìý

Abi: It's funny. Martin gave me the great privilege of publishing an article on his blog recently about measuring productivity. He seemed accepting of my perspective on it. I've often shared that article by Martin. I think he captured-- that was written 15 years ago or something, which is really funny to think. We're still having the same debates, 20, 15 years later. I agreed with everything he wrote in that article he talked about. Inevitably, every business tries to measure productivity and lines of code and function points and this doesn't work, it feels impossible. I agree with his observation of what doesn't work.

Ìý

I wouldn't agree with Martin from a standpoint that there isn't an answer to that question or that there shouldn't be. I think there needs to be and there is. I think there's some discomfort in that. It's not a simple answer and even today it's not a clear answer. I do believe there's a way forward, and that's why I've spent the last seven years of my life and we're publishing all these papers and trying to figure this out with the best researchers and best companies in the world that are trying to also solve this problem.

Ìý

Prem: What you're saying is there is light at the end of that tunnel where we can measure the right things and you actually drive the right insights is what I'm hearing from you.

Ìý

Abi: Absolutely. That's the business I'm in. I'm biased, but we'll talk more about some of the newer approaches that we're leveraging to do that. Today, I think we are closer than ever to having a clear point of view that both developers, managers, CEOs, boards, investors, that everyone can be aligned around to have a clear view on that question.

Ìý

Scott: We should probably get into what is the DevEx framework and what are the metrics that you're advocating people use.

Ìý

Abi: I want to set a little bit of context on how the whole DevEx framework and paper and research, where that originated. If you were to go back and read the Space paper, the Space framework, The Space of Developer Productivity is the title, it talks about how productivity is multidimensional. In particular, if you read through the paper, there's a lot of emphasis on the importance of self-reported and perceptual data in our understanding of developer productivity. It talks about things like developer satisfaction, or developer's perspectives and observations on how things are going.

Ìý

One of the things that happened after Space was published was everyone started asking, "Well, how do we do that? What does this mean? How do we put this into use?" The authors of the DevEx framework were myself, and both the two lead authors of Space, Nicole Forsgren, as well as Dr. Margaret-Anne Storey, who's at University of Victoria, amazing researcher, I think the most prolific, most published researcher in developer productivity.

Ìý

What we were trying to do is put forth a practical approach to this perceptual measure thing. Our research into that really took us in a number of different directions, but ultimately, took us to this concept of developer experience, which was already a little bit of a research concept, but in 2020, we actually published, I think, really the first paper that put forth a framework in conceptual model around what even is a developer experience, and what does it consist of.

Ìý

The most recent DevEx framework is really a practical approach to defining developer experience, which is something that a lot of the organizations we work with struggle with, is defining, just like productivity is difficult to get aligned around, how do we even think about developer experience? Then, of course, how do we measure it? In our framework, we provide a multidimensional framework that looks at perceptual sentiment measures, as well as objective workflow measures, combined with North Star metrics to get a complete view into developer experience.

Ìý

I should also add that measuring developer experience isn't the end all be all. Developer experience is a component of how we should think about and look at developer productivity as a whole. I think developer experience is a critical piece to it, and of course, the title of that paper was DevEx: What Actually Drives Productivity. If you're looking to improve productivity, and understand how to do that, developer experience data and measurement is key to that, but there's more to the story if you're just trying to understand the impact of the R&D organization, or the efficiency, or the quality of the work. There's, of course, other types of measurements and understanding that you need around those types of questions.

Ìý

Scott: I know that when the Space paper was first published, I'm one of the people that does the tech radar, and we discussed that, so we proposed that as a blip, but we talked about it, and we came to a conclusion we're not really sure what we're blipping here. Are we giving some advice to people? It's an interesting way to look at it. When the DevEx framework came out, that was really helpful, I think, because it put some better, more shape to it.

Ìý

Abi: Yes. That was part of the intention. It just so happened that between the time Space was published, and the DevEx framework was published, developer experience came into its own. Back in 20, I would say, what was it, '19, people weren't really talking about developer experience, a little bit they were, but by the time we published this paper a year and a half ago, it was unbelievable to see how much attention. Gardner was talking about developer experience, everyone was talking about, the Stack Overflow was talking about. GitHub is all about developer experience in AI, of course. It was good timing, and really resonated, I think, with practitioners and industry leaders.

Ìý

Scott: That brings up an interesting point that I was thinking about after I listened to your podcast, that I never put together this interest, this resurgence of interest in developer experience with the COVID-19 epidemic, and remote work, the rise of remote work. Do you think that those-- are people worried about what-- are there people actually working when they're at home? Is that what it's about?

Ìý

Abi: The COVID-19 impact was really that suddenly companies shifted from co-located, in-person teams in work, to hybrid and remote work overnight. That is when Space was published, and it was really an answer to a spike of questions around, and business is trying to understand the impact of that. Hey, everyone just went remote. Are we more productive? Are we less productive? That was the question being asked.

Ìý

I think developer experience trailed that. I think developer experience, I think there's a number of reasons why we've seen such a rise and focus and attention on developer experience. I think when I talk to companies, a lot of them, it's capped off, or finished off their continuous delivery transformations, if you will, DevOps transformation. Had moved to continuous delivery. A lot of the previous ways in which organizations were measuring, like the Dora 4 Key metrics, started to no longer really be applicable. It started to no longer give the answer to how do we get better.

Ìý

I think that's actually really what has driven the interest in developer experiences. As we've moved, as organizations have moved beyond adaption of continuous delivery, the next problem becomes, well, now that we've done this whole DevOps thing, however you want to define it, that's a whole thing, but now, how do we actually make it effective for developers?

Ìý

We've given them all these tools, and more responsibilities, and we have all these microservices, and who knows what else, how do we manage all this stuff, and make the daily work of developers streamlined? I think that's how developer experience has organically risen to be a focus area for organizations.

Ìý

Prem: Are you saying here that there are these specific metrics that we can actually look at and draw some good insights? Is that what you seem to be suggesting?

Ìý

Abi: Absolutely. As we start entering our discussion about metrics, there's a lot more to it than just that, but certainly, measuring developer experience is something we found to be incredibly valuable as a source of insight to organizations that are trying to improve productivity, or have investments in productivity platform engineering developer experience initiatives, and are trying to understand the impact of that to our lives.

Ìý

I think it's table stakes for that, and more broadly, if you're a business leader trying to understand the competitiveness and the agility of your R&D organization, gauging the developer experience needs to be part of how you get insight into that question.

Ìý

Prem: Are there good metrics, and are there bad metrics? Can you give us some examples so that our listeners understand what is good versus what is bad, if there is something like that?

Ìý

Abi: I wouldn't say that there's universally good or bad metrics. I think there's good or bad metrics for different purposes. To give you an example, if you're trying to understand and measure, or even incentivize your organization to ship faster, measuring story points would be a bad metric to incentivize that improvement and behavior for, I think, reasons that are pretty well understood.

Ìý

Now, if you're trying to understand the predictability of your organization at delivering on time or against the estimate, then story points are the metric you should be using to get visibly into that. I can't give a universal good or bad metrics. I can give context-specific point of view on what metrics work for what types of problems.

Ìý

Prem: What you're seeing here looks like the intent and the context matter quite a bit. A good metric in one context can be a bad metric in another. Is this what I'm concluding from what you just said?

Ìý

Abi: Correct. Even within a context, there's different levels of-- I have a little framework I share. I say, "Hey, are you using this data to understand, to align, or to assess?" If you're using, for example, Microsoft, they look at, pull request for developer. They heavily rely on that as an input into how they understand productivity, do research around productivity, to understand productivity, but if Microsoft were to use that to assess developers, we can all picture some of the bad things that could happen from that.

Ìý

Again, even within the context of a purpose, there's that intent, from which you mentioned, what's the intent, and what's communicated? If you don't communicate your intent clearly, there could be consequences as well.

Ìý

Prem: Right. Here is a thing that we recently did as well. There is this debate on the kinds of metrics, qualitative versus quantitative. I know that your article on martinfowler.com also talks about that, right? A couple of scenarios, in this example that you were-- this campaign that you were doing, consulting work with, they had and asked, tell us what kinds of metrics, and so on and so forth.

Ìý

Now, they didn't have a lot of quantitative information. Then we had to get started, and we said, "Okay, let's look at qualitative information, gather data through surveys and questionnaires, and so on and so forth, interviews, and so on and so forth." Now, one thing that happened was which was interesting, we got a bunch of data. It was all over the place, and then when we actually dug in to see under the covers, a lot of it didn't actually quite match up with what people actually answered.

Ìý

I was like, "Huh, I read your article as well, and I saw that there was a huge emphasis on qualitative information." I was almost like, "No, I don't know if this really works." I'd rather believe the quantitative data than qualitative information. I wasn't quite sure how to move forward with that qualitative information, especially because it seemed to give us the wrong indicators.

Ìý

Abi: Yes. That's interesting. That's not uncommon for people to have that experience. My personal recommendation, and our recommendation of our company, is to use both. You want as much data as you possibly can have. If you look at how organizations like Google, that are heavily instrumented, they have the IDE to every part of the developer environment instrumented. They have metrics on everything. They talk about how they still use self-reported data as a way of complementing that and even double-checking the metrics they collect.

Ìý

Prem, based on the experience you described, I think partly it could just boil down to the design of the survey itself because getting reliable and accurate information is really difficult without really rigorously designed questions. The process for designing one question is a lot more QA than we do for our software releases, I can tell you that. Also, the fact that you found divergent information from what people were telling you and what the systems and objective data were telling you, I see that as an opportunity.

Ìý

That's an opportunity either to realize that only one of these things could be true, either the people misinterpreted the question, the people lied, [laughs] or your objective metrics aren't telling you the full story. Whatever the reason may be, that's probably valuable to know. Again, it boils back down to it's about having the most data possible, so double-checking your qualitative data with the quantitative and vice versa is really bad. I'm sure, Prem, you've come across instances where the objective metrics are misleading as well.

Ìý

Google talks about their famous example is they were trying to measure build times. They were seeing some data that had them really concerned about developers getting blocked waiting on CI, and then once they talked to the developers, they realized, "Oh, those are actually programmatic bot builds." They weren't actually holding developers back. Again, I think the opportunity is to collect as much data as possible from people and from systems and then triangulate it and leverage it together to give you the fullest possible picture.

Ìý

Prem: Absolutely. I think, again, it boils down to the thing that you previously said, which was about intent and communicating the intent. It wasn't that people were lying, but I think they naturally became defensive because they weren't sure, or rather, they probably were speculating that this would be misused. If they self-reported bad things, then it would have adverse consequences for them, and what's the natural thing humans do? They become very defensive, right?

Ìý

Abi: Yes. That's where, in our approach, things like anonymity and the communication and the design of it, it all goes together to hopefully produce reliable as possible measurements and then cross-cultural concern. We have to translate things into multiple languages. It gets pretty deep and complex. Yes, it's not easy to do. I always tell people, doing a SQL query on your GitHub data is a lot easier than designing, getting good developer experience data, a lot, lot easier. Again, by leveraging both, I think you're able to get the insights you need as quickly as you can and in a cost-effective way.

Ìý

Scott: I'm really curious about the methodology and how the methodology might have led to the conclusions that the papers describe. I'm guilty of having sent surveys to developers many times to try and track what's going on in our business and the things we're delivering in. Social scientists have been doing this for a long time. This is a really well-established field, how you ask questions, how the questions are worded, how do you statistically interpret the results? I wonder, how much of that went into the development of these frameworks?

Ìý

Abi: Funny thing about Nicole, I think she's a real pioneer because she's a computer scientist who ventured into psychometrics, which is the social sciences. That's one of the things that's inspired me and inspired a lot of the folks in our team. At DevEx, we have folks who actually come from the social sciences, so industrial-organizational psychology, for example, is this is what they do, is develop questionnaires and assessments, psychometrics to assess, make really mission-critical assessments about people's health, mental state, performance. It's a fascinating field.

Ìý

Our approach has been heavily driven by the social sciences and best practices. The methodology of the development of our measurement instrument, or survey instrument, if you will, adheres to an industry standard that comes from the social sciences. That is a standard in any sort of psychometrics assessment. There's a rigorous standard for how that has to be developed, tested, and validated in order to be considered even a valid data source.

Ìý

Then, of course, we have had to solve challenges that are unique to R&D organizations and software developers, and there are many, let me tell you that. [laughs] Then find how to apply this approach of self-reported data and psychometrics, how do we apply that in a way that's most useful for business, and that's, of course, an ongoing journey as well. A lot I could share, but it's definitely been a really exciting opportunity to blend these fields together and be a pioneer in that.

Ìý

Scott: One of the things that's really challenging is survey fatigue, and that people are just so reluctant, even if it only takes 5 minutes to fill out a form. We get response rates when we do that kind of stuff, we get really poor response rates from well-intentioned people because they've got a lot of other stuff going on. How do you avoid that?

Ìý

Abi: Well, at ºÚÁÏÃÅ you all covered, did a little blip on our survey platform. I think you covered the response rates you saw at ºÚÁÏÃÅ. We at DevEx sustain on average over 90% participation rates on the self-reported surveys that our customers run. People don't believe us when we bring that up, but it's true. You can talk to our customers to verify that.

Ìý

Then the second question we get asked is, "How do you do that?" There's a lot of things that go into it. I could share a couple of examples of, I think, some of the strategies that we leverage. One thing we do is that the results of the survey are transparent to the entire organization down to the developers themselves. If you think about what it's like to typically take a survey, you get an archaic-looking form, probably sent to your email.

Ìý

You fill it out reluctantly because, out of the kindness of your heart, you want to be helpful, and then you never hear about it again. Maybe you see a bullet point about it in the next all-hands, but you never see the results, you never get any benefit from it, there's nothing that data does for you. Then, the next quarter, you get that same link in your email, and you decide, "I'm not going to participate in that."

Ìý

One thing we do with DevEx is that the data is open to everyone, and it's compiled in a week, so it's not like you see the results three months later, you see the results within days of submitting your feedback. Again, there's a lot of things that go into the response rates, but that's one I would call out as an example of a very stark difference between, I think, how organizations traditionally run surveys and our approach with the customers we work with.

Ìý

Scott: I think that's really important, the transparency and seeing that there is a consequence, or that there are actions that are taken in response to those. It takes a certain amount of courage, I would think, on the part of the leaders to be that transparent, vulnerable.

Ìý

Abi: Absolutely. It's funny sometimes I wonder if they don't realize it [laughs] till the data becomes available. I will say, if you had to ask me what's one thing that really surprises me about what we're seeing with the organizations that use our platform, I think that would actually be one of them. It is surprising to me that so many of these organizations are comfortable capturing the voice of the developer, and revealing it to everybody. Because developers, they voice their opinions, and they're direct about the concerns and opportunities they see.

Ìý

I'll also tell you that these leaders find that data to be incredibly valuable and insightful. There's some risk associated with that, but ultimately, I think keeping it open benefits all parties and solves, or helps solve the participation rate problem, which pretty much kills a lot of these survey programs before they even really get any traction typically.

Ìý

Scott: I see a lot of organizations that are really embracing this, because they understand the benefits that it's going to bring to them. I still encounter a lot of businesses who-- When you talk about developer experience as a critical thing to improve, they see that as maybe narcissistic on behalf of the developers. Like, why do I care about their happiness? My business is at stake here, and I wonder, what do I tell those people? Maybe you can help me out.

Ìý

Abi: This is a problem I'm really focused on personally right now, because as you both know, I've spent the last few years championing developer experience, and getting a lot of organizations to adopt and invest in developer experience. I think a lot of organizations run into the exact challenge, Scott, that you're bringing up, which is they get this data, they want to make investments, so then the rubber meets the road, and executives are asking why? This is about making developers happy, that's not our highest priority at the moment.

Ìý

We have to be profitable. We have to ship this feature, this product, get it to market. I think developer experience is a positioning problem, because, as you said, Scott, you said the leaders you talk about seems almost narcissistic. Why is developers' feelings the priority for our business? That's a problem of misunderstanding, and mispositioning on the people championing developer experience, because developer experience is not about making developers happy. Developer experience is about understanding the friction, what's slowing you down, and costing you money in your business.

Ìý

Today, I think my recommendation, and what I'm advocating for, and the guidance I give to leaders is to think about developer experience as just one pillar in how you're thinking about productivity. You can think about productivity maybe across four dimensions, speed, effectiveness, quality, impact. I see developer experience as fitting in under that effectiveness pillar. I know that's something at ThoughtWorks you all have published quite a bit about as well, so that's all it is.

Ìý

Developer experience is your best signal into effectiveness. It's not the whole pie. The developer productivity is bigger than that, but it's a critical piece, and in particular it's about understanding improving effectiveness so that you can also improve speed, quality, and impact of your R&D organization.

Ìý

Scott: I think one of the things about DORA and the DORA metrics is that they showed that those metrics are predictive of business success. I think that gets lost a lot. It's one of my pet peeves is people measuring Dora metrics without really understanding why, or what they're—

Ìý

Abi: Me too, Scott.

Ìý

Scott:Ìý—getting out of it, and dying to use that to define team performance and things like that. Anyway, is this something you've looked at? Like, how can we predict business performance based on developer experience? Is there any correlation there?

Ìý

Abi: We are about to publish something just on it. Earlier this year, we published a paper with Microsoft and GitHub, a very quick study looking at the impact of developer experience. Following that study, we've invested in a much larger study around developer experience, and predictive modeling around our measure of developer experience, we've developed DevEx, which is called the Developer Experience Index, and it is predictive. We've looked at business outcomes like profitability and growth.

Ìý

We've looked at other outcomes like rate of delivery, time waste, or regrettable attrition. I can't go into detail here, but I can tell you that one of the things we found was a very strong correlation between developer experience and time waste. To the point where you can pretty much say, "Hey, X a number of point increase in your developer experience index score translates to X percentage point decrease in time waste, which translates to X amount of dollars per engineer and recapature efficiency savings."

Ìý

That data's coming, and that data is there. I still don't think that will solve the whole problem we were talking about, Scott. I still think executives will just at face value misunderstand developer experience. Again, that's why I think developer experience needs to be spoken about and positioned as part of a larger focus on developer productivity rather than as its own singular thing.

Ìý

Prem: Wonderful. Oh, it's 2024, so thanks to ChatGPT. This is where AI finally has maybe become mainstream. Are there any AI and generative AI techniques that we can use to make this process of understanding developer productivity a lot more easier?

Ìý

Abi: Yes. At DevEx one maybe obvious application of this, we have organizations that on a given survey will collect upwards of 20,000 comments from their developers. If you think about an organization with 8,000 developers, and each developer maybe writes two or three comments that adds up to tens of thousands. We've been able to leverage GenAI as a fantastic tool for synthesizing that information, and deriving actionable insights from it quickly without manually sifting through tons of data.

Ìý

Of course, more broadly I think GenAI has a lot of potential to help with developer experience, and developer productivity. One of the examples I'd like to share is one of the most common things, opportunities, challenges we see with the organizations around developer productivity is knowledge sharing, and documentation, and technical debt. Well, GenAI has huge potential to help with quickly understanding a code base, or automating, documenting thoughts and processes of a developer.

Ìý

That's a use case in particular I'm really excited about. Of course, the code gen copilot stuff, I think, is, of course, pretty intuitive and exciting, and we do see data on the impact of that. That's very positive. At the same time I think it only tackles the code writing aspect of the developer day, and developer experience, which is really not where I think the most friction actually exists if you look at the SDLC.

Ìý

Scott: We've found the same thing, and in that developers really like copilot. They don't want to give it up, and they find it useful, but it's been much more difficult to find any impact on the overall cycle time, and there's a lot of reasons for that. I think that's another podcast. One of the things we always try to do when we wrap up the podcast is try to get some call to action, and how to help people get started. Like if you have a development organization, an engineering organization, and you want to start understanding developer experience, where's a good place to start with that?

Ìý

Abi: Sure. I'll give two calls to action. One is, if you're in the scenario you just described, Scott, you're a leader, or you're in charge of a platform organization, productivity organization, and you want data, and you want data fast, then a survey is the fastest way you're going to get holistic baselines about productivity and experience. On that Martin Fowler article that we referenced today, we did provide a starter template for a survey you could use as well as other examples of different types of questions and insights you can capture.

Ìý

I'd recommend, folks, you go-- It comes with a Google form template, you can make a copy of it, get started with it right away. The other recommendation I have is, whatever company you're at right now, if you don't have an aligned way of thinking about and measuring productivity today, then I would recommend you get that into place, because every organization I talk to at some point has to answer that question.

Ìý

If you can get ahead of that problem now, not only is that going to save you a lot of pain down the road, but it's going to give you value today to get a baseline and start tracking how your organization is changing and where the opportunities are. If you don't have an answer to how do we measure productivity at your organization, yes, I would invite folks to try to figure out that question.

Ìý

Scott: Yes. I know that's one thing that we run into a lot, is the absence of a baseline, where people want to know right away how they're doing. It's like, well, what do you have to compare against? I think we'll wrap it up here. I wanted to remind everybody that you also have a podcast, and they should go check that out. There's a lot more information and in-depth discussion these things as well as, I think, the practical experiences of a lot of businesses who are putting this into practice. Where would people go to find that podcast?

Ìý

Abi: It's called Engineering Enablement podcast. It's on Apple, Spotify. I think anywhere where you listen to podcasts, I think, it's published there. As you mentioned, we're fortunate to have folks like Nicole Forsgren, leaders from companies like Google, Airbnb, frequently come on the show to talk about how they're thinking about productivity, how they're measuring it, and all the challenges that come along with that. I would invite folks to check that out. Thanks for the shout-out.

Ìý

Scott: Great. Well, thanks a lot. I could talk about this for a lot longer, and I have lots more questions, but we need to wrap it up. Thank you so much for taking the time to do this. I really appreciate it.

Ìý

Abi: Yes. Thanks, Scott. It's a pleasure to be here.