The New Stack Podcast

What’s the Future of Feature Management?

Episode Summary

Feature management isn’t a new idea but lately it’s a trend that’s picked up speed. Analysts like Forrester and Gartner have cited adoption of the practice as being, respectively, “hot” and “the dominant approach to experimentation in software engineering.” A study released in November found that 60% of 1,000 software and IT professionals surveyed started using feature flags only in the past year, according to the report sponsored by LaunchDarkly, the feature management platform and conducted by Wakefield Research. At the heart of feature management are feature flags, which give organizations the ability to turn features on and off, without having to re-deploy an entire app. Feature flags allow organizations test new features, and control things like access to premium versions of a customer-facing service. An overall feature management practice that includes feature flags allows organizations “to release progressively any new feature to any segment of users, any environment, any cohort of customers in a controlled manner that really reduces the risk of each release,” said Ravi Tharisayi, senior director of product marketing at LaunchDarkly, in this episode of The New Stack Makers podcast. Tharisayi talked to The New Stack’s features editor, Heather Joslyn, about the future of feature management, on the eve of the company’s latest Trajectory user conference. This episode of Makers was sponsored by LaunchDarkly.

Episode Notes

Feature management isn’t a new idea but lately it’s a trend that’s picked up speed. Analysts like Forrester and Gartner have cited adoption of the practice as being, respectively, “hot” and “the dominant approach to experimentation in software engineering.”

 

A study released in November found that 60% of 1,000 software and IT professionals surveyed started using feature flags only in the past year, according to the report sponsored by LaunchDarkly, the feature management platform and conducted by Wakefield Research.

 

At the heart of feature management are feature flags, which give organizations the ability to turn features on and off, without having to re-deploy an entire app. Feature flags allow organizations test new features, and control things like access to premium versions of a customer-facing service.

 

An overall feature management practice that includes feature flags allows organizations “to release progressively any new feature to any segment of users, any environment, any cohort of customers in a controlled manner that really reduces the risk of each release,” said Ravi Tharisayi, senior director of product marketing at LaunchDarkly, in this episode of The New Stack Makers podcast.

 

Tharisayi talked to The New Stack’s features editor, Heather Joslyn, about the future of feature management, on the eve of the company’s latest Trajectory user conference. This episode of Makers was sponsored by LaunchDarkly.

Streamlining Management, Saving Money

The participants in the new survey worked at companies of at least 200 employees, and nearly all of them that use feature flags — 98%— said they believe they save their organizations money and demonstrate a return on investment.

 

Furthermore, 70% said that their company views feature management as either a mission-critical or a high-priority investment.

 

Fielding the annual survey, Tharisayi said, has offered a window into how organizations are using feature flags. Fifty-five percent of customers in the 2022 survey said they use feature flags as long-term operational controls — for API rate limiting, for instance, to prioritize certain API calls in high-traffic situations.

 

The second most common use, the survey found — cited by 47% of users — was for entitlements, “managing access to different types of plans, premium plans versus other plans, for example,” Tharisayi said.

 

“This is really a powerful capability because of this ability to allow product managers or other personas to manage who has access to certain features to certain plans, without having to have developers be involved,” he said. “Previously, that required a lot of developer involvement.”

Experimentation, Metrics, Cultural Shifts

LaunchDarkly, Tharisayi said, has been investing in and improving its platform’s experimentation and measurement capabilities: “At the core of that is this notion that experimentation can be a lot more successful when it's tightly integrated to the developer workflow.”

 

As an example, he pointed to CCP Games, makers of the gaming platform EVE Online, which serves millions of players.

 

“They were recently thinking through how to evolve their recommendation engine, because they wanted this engine to recommend actions for their gamers that will hopefully increase their ultimate North Star metric,” its tracking of how much time gamers spend with their games.

 

By using LaunchDarkly’s platform, CCP was able to run A/B tests and increase gamers’ session lengths and engagement. ”So that's the kind of capability that we think is going to be an increasing priority,” Tharisayi said.

 

As feature management matures and standardizes, he said, he pointed to the adoption of DevOps as a model and cautionary tale.

 

”When it comes to cultural shifts, like DevOps or feature management that require teams to work in a different way, oftentimes there can be early success with a small team,” Tharisayi said “But then there can be some cultural and process barriers as you're trying to standardize to the team level and multi-team level, before figuring out the kinks in deploying it at an organization-wide level.”

 

He added, “that's one of the trends that we observed a little bit in this survey, is that there are some cultural elements to getting success at scale, with something like feature management and the opportunity as an industry to support organizations as they're making that quest to standardize a practice like this, like any other cultural practice.”

 

Check out the full episode for more on the survey and on what’s next for feature management.

Episode Transcription

Alex Williams  0:08  

You're listening to the new stack makers, a podcast made for people who develop, deploy and manage at scale software. For more conversations that articles go to the new stack dot I O. All right now on with the show.

 

Colleen Coll  0:30  

Unleashed developer productivity for the software powered world by fundamentally changing how you deliver software to your customers with LaunchDarkly ease feature management platform, empower developers can empower the business to release new features faster and more efficiently than ever.

 

Heather Joslyn  0:49  

Hello, and welcome to another episode of the new stack makers Podcast. I'm Heather Jocelyn, your host Features Editor of the new stack. Today we're going to be talking about the state of feature management. We'll be joined today by Ravi thirsty, who is Senior Director of Product Marketing for LaunchDarkly, a feature management platform. Hi, Ravi. Hi, Heather.

 

Unknown Speaker  1:06  

It's great to be here. I'm a big fan of the show. So it's really exciting to join you and talk about this.

 

Heather Joslyn  1:12  

Excellent. Thank you for listening and for being a fan. And for joining us, we'll be talking about the next generation of feature management tools. These tools are powered by feature flags, and they let developers release their code into production without fully releasing them the code to end users. What's the state of feature management tools and practices, we'll hear Robbie's ideas about what the current status is and what the future holds. And we'll talk about the findings of a new study by LaunchDarkly. And Wakefield research about the state of feature management is a taste of what you're going to hear about the survey found that nearly 70% of respondents believe that feature management is either high priority or mission critical. And among organizations that deploy several times a week or more 77%, say their leaders view feature management is top priority. Before we begin, we'd like to thank LaunchDarkly for sponsoring this episode of makers. And there's a lot of talk a lot to talk about. So let's get started Robbie, to make sure that we're our listeners understand the parameters of what we're talking about what is feature management and what are feature flags.

 

Unknown Speaker  2:11  

Yeah, obviously a great place to start. And these are definitely interrelated concepts. I'll start with feature flags. Because these are concepts has been around for quite some time and different forms of fashion. Some organizations will use what's called config files to manage application elements. But at the core feature flags are levers of control in your application. And they allow development teams to turn off features on or off without having to redeploy the entire application, which can really be a powerful capability. And so feature management is really an extension of that it provides a scalable framework for teams and really, organizations to adopt feature flags at scale as a standardized practice, allowing those organizations to as part of their standard development process to release progressively any new feature to any segment of users, any environment, any cohort of customers in a controlled manner, that really reduces the risk of each release, and gives the organization a lot more control over the process of releasing features to end users. So it's a really powerful capability. That term feature management was actually coined by our CEO, Edith Harbaugh. But now it's been recognized by the major analyst firms as a growing DevOps trend. And we're excited to be at the center of it.

 

Heather Joslyn  3:45  

Just to give our readers an example of a use case, like that would be for example, if you had like a premium version of a service versus a basic version of a service, like that would be one when use case perhaps for feature flag to be able to turn it on and off.

 

Unknown Speaker  4:00  

Absolutely. So in that kind of scenario, the ability to beta that new capability with a certain cohort of customers to gather feedback from just that cohort of customers, before releasing it more broadly, being able to test and see that it's working correctly, and then progressively releasing it more so. Yeah, that's the concept.

 

Heather Joslyn  4:24  

What role does a feature management and progressive delivery play in a DevOps organization?

 

Unknown Speaker  4:31  

Sure, I mean, DevOps has been such a core element to modern development. When I think of DevOps, I think of so many basic ideals, ideals, like collaboration between teams, this idea of learning and driving faster feedback from especially from customers. And then you know, things like measurement, how to understand whether the things that you're shipping are actually making an impact and feature management really, I feel enabled a lot of those ideals in so many different ways, at its core is this notion of the ability to release new capabilities, new features into production more frequently. And obviously, we're going to talk about the survey. But one of the findings from that survey was that indeed, feature management for many of our respondents, 40% indicated that feature management was considered a foundational element to continuous delivery, as you can imagine, like this ability to put a feature behind a feature flag just makes you a lot more confident to release to production, because you know that if something goes wrong, you can roll it back, you can control the blast radius, by controlling who gets to see that new feature. So it really just enables a lot more frequent release into production. So I think that's, that's definitely one big element. The other element is around delegation. And, you know, what's really powerful is that once a feature flag has been added to your application, and you're leveraging a feature management system, suddenly, you're really empowering the rest of the organization to potentially have the ability to control that release to end users. So just in the scenario that we were, you were just mentioning, maybe a premium offering, you know, in that kind of scenario, the development team can add the feature flag, but then the product manager can leverage a feature management system, and manage who gets access to that premium service, without the development team having to be involved. And for many organizations, that's a real powerful capability, it really again, enables a level of collaboration across teams that really, I think, aligns with those DevOps core values.

 

Heather Joslyn  6:52  

And sounds like the velocity would be impacted, too, because you wouldn't have to go through another product manager to just make a decision as opposed to going through developers.

 

Unknown Speaker  7:01  

Absolutely, absolutely. And, you know, again, I think that kind of reduces the cognitive load for developers, you know, sort of adding the feature flag, moving on to the next area, the product manager is able to manage some of those entitlement elements in the release process. So yeah, I mean, it really streamlines the ability to do development at scale in a more efficient way. Last thing I want to mention, is just around measurement, because I think what's really powerful is the fact that with feature flags, you have the ability to attach metrics to those flags, and gather data relative to those features that are that are being lost. So again, going back to our example of that premium service, potentially adding performance metrics, like the latency related to that service to that new feature, or the session time length related to adopting that service. So really, given your unique ability and granular ability to understand the impact the business impact of the feature that you're developing.

 

Heather Joslyn  8:11  

Let's talk about the survey, which we mentioned briefly in the intro, and you've referred to Wakefield research surveyed 1000 software and IT professionals that companies have at least 200 employees. So these are larger companies, or midsize or larger companies, what what to you is the most important or surprising finding?

 

Unknown Speaker  8:28  

Yeah, I'll just mention as well, I mean, this is the second time that we've done this data feature management survey, we're excited about the size of the sample, I mean, 1000 professionals taking part in this survey. And I should mention as well, that these were folks that have either recently adopted feature flags, or are planning to adopt feature flags, so are using feature flags already or planning to so that's sort of the cohort of respondents that we were looking at what the survey, there was a lot of interesting findings, one that I'll start with was just the number of recent adopters of feature flags. So we saw that 60% of respondents indicated that they had started their feature flag journey just in the past year, and 20%, within the past six months. And obviously, we're as a LaunchDarkly, we're sort of at the center of this. We see a lot of customers getting started, you know, it was interesting to see that so many were were very early in their adoption. And then related to that was the pretty incredible results that these organizations were seeing relative to some of the their peers who had started previously, we saw that these early adopters were seeing faster deployment frequency as well as better. Meantime to resolution on incidents. It was interesting to see that these early adopters are not only adopting quickly but seeing results quickly. That was really interesting and I think, you know, aligns with some what we're seeing in terms of just the growth in this the adoption of feature management as a practice.

 

Heather Joslyn  10:08  

And just to surface one of those findings, the serie read that 14% Of those, those recent adopters are resolving incidents within our compared to 6%, who had been using feature flags for longer than a year. What do you think that is that the newcomers are seeing such good performance or just resolving these incidents so much faster?

 

Unknown Speaker  10:29  

Yeah, when looking at those results, one of the things that our team thought about was this overall discussion around developer experience, and you know how to help developers be more successful earlier and faster. And the kinds of enablement areas that developers need to be successful faster. And, you know, when we look at feature management, we think there are there are a lot of those elements, there has been a lot of learnings from, from previous experiences. There's more commercial vendors on the scene that are providing best practices and ideas for how to get started, a lot of organizations sharing their experiences. So we think that might be one hypothesis for why those early adopters are seeing better successes. They're kind of having the opportunity to build upon the learnings from previous adopters, as well as the growing library of best practices that are really starting to emerge that they can build upon.

 

Heather Joslyn  11:29  

There. In the survey, clear majorities of those who use feature management say they've seen improved processes as a result, for example, 98% of those who use feature flags said that toggles save their company money and delivers demonstrable ROI. What are some of the other more common specific ways in which they're seeing benefits from from using feature management?

 

Unknown Speaker  11:53  

Yeah, for folks that leverage feature management, there's use cases that often come to mind things like dark launching, for example, putting something behind a feature flag, similar to what we were just talking about earlier with it with a premium service. Other examples are things like trunk based development. So this idea of putting something behind a flag, even though it's incomplete, because you know that it won't be seen by users, and that you can emerge more often more confidently. What's been super interesting, not only in this survey, but as well, in our previous survey is the incidents that we see from respondents indicating they're they're leveraging long lived feature flags, I want to talk a little bit about that there's a couple of different areas that we see organizations leveraging more long lived feature flags. An example of this is sort of long term operational controls, leveraging feature flags. So you might hear things like API rate limiting, but basically, the concept here is leveraging feature flags to manage kind of overall system elements of the application. An example of this when I first joined LaunchDarkly, a friend of mine was leveraging LaunchDarkly for a mobile ordering. And was talking about how when their app was used at certain trade shows that there would be this massive spike in traffic that that they needed to figure out how to manage. And so they talked about how they leveraged feature flags as a way to manage those high peak traffic situations. And the feature flags would basically prioritize traffic that was sort of high priority that they wanted to ensure was being responded to quickly. And then deprioritize other traffic so that not every customer was feeling slow traffic that they they were able to sort of prioritize the traffic in the most optimal way. And that's an example of leveraging feature flags, and sort of this API rate limiting use case where you're able to manage the system ensure sort of graceful degradation by leveraging feature flags in that really dynamic way. We saw that 55% of respondents are leveraging feature flags, this type of way in this type of notion. And then another example of this is entitlements, managing access to different types of plans, premium plans versus other plans, for example, and again, going back to our example that we started with, but again, this is really a powerful capability because of this ability to allow product managers or other personas to manage who has access to certain features to certain plans, without having to have developers be involved previously, that required a lot of developer involvement. And we're seeing organizations leverage feature flags to streamline that. In our survey, 47% of respondents indicated that they're using feature flags for entitlements. So it's really interesting and part of what's fun being in this area is we're oftentimes learning from our customers and terms of the kinds of things that they're doing that the kinds of use cases. And so we're looking forward to continuing this survey to continue to learn more as organizations find new and interesting ways to get value from it.

 

Heather Joslyn  15:13  

The study found that the largest plurality of respondents think they will need roughly an additional $100,000 in their budgets in the coming year for feature management. And it also revealed that basically, the closer they are their sponsor to the C suite, the less likely they are to believe they'll get that that funding, what in your view are the arguments that should be made to the people in organizations that control that that funding, you know, the C suite or or upper management to be able to fund future management?

 

Unknown Speaker  15:42  

Yeah, we did dive quite a bit into this. And part of what's interesting, we took the survey towards the end of September, early October timeframe. So the macro environment certainly was baked in to some of what our respondents were indicating there. I think, in spite of that, what was interesting was that we're in this environment where a lot of organizations are looking at their budgets, looking where they can streamline their budgets. But we did see from our survey that at least, feature management appeared to be an area where many organizations were looking to invest in more we saw that 80% expected their budgets to grow in the coming year, one of the things that we think is an opportunity for organizations to think about in relation to investment in something like feature management, is sort of reframing some of the conversation we were talking earlier about DevOps. And certainly things like DevOps have created a lot of urgency around speed and sort of this quest to get to market faster to be more competitive faster. And from our standpoint, we've seen that that quest for speed has delivered a lot of value in the sense of more iterative development, faster feedback, more agility. But you also do think that there are still problems, even as organizations make that quest for speed. And a couple that that we think are really important is managing the risk related to speed, that this notion that going faster, can also be risky, and how to think appropriately about counterbalancing, the quest for speed with ways of managing risk. And then the second one is about measurement. And coming back to this notion of measurement, you know, a lot of what we see from engineering teams, the measures are oftentimes proxies for actual business value, you know, sort of metrics around deploys. And lead time, we consider these sort of proxies for ultimate business value, but but not ultimately the the measure of business value. And that there is still this quest to understand that the business value of of engineering efforts, and that's where feature management comes in a way to both manage risks related to speed as well as a way to attach measurements, conduct experiments, and really understand the business value of engineering efforts. So we think those two are really great ways to think about investment and ways for organizations to earn more investment support and feature management. The

 

Heather Joslyn  18:26  

point you make about how developers measure things aren't necessarily how tell you anything about business. I mean, the number of deploys doesn't tell you how customers are reacting to what you're giving them. One thing that report did was sort of indicate that there's a maturity model for feature management emerging, what are some of the characteristics of that model? What what constitutes a mature or maturing feature management?

 

Unknown Speaker  18:47  

Yeah, this was an area we were interested to explore something we're seeing with our, with our customers. One of the dimensions we tried to explore in this survey was just around standardization. And, you know, sort of going from ad hoc adoption of feature management to sort of team level adoption to multi team level and organization wide standardization. It was an interesting time to explore that we know there's a lot of discussion around things like platform engineering, and sort of how to standardize certain practices. One of the interesting things that we saw in exploring that was, we saw that organizations that were earlier in their adoption of feature management saw some quick success, there was sort of this maybe low in terms of success, you know, sort of a dip in terms of success, you know, in sort of the middle stages of standardization, and then a return to greater success at full standardization across the organization. So a sort of this J curve of success, that actually we've seen in other reports, we thought especially of the accelerate state of DevOps report from 2021, which showed a similar sort of finding, you know, I think the hypothesis is that when it comes to cultural shifts, like DevOps or feature management that require teams to work in a different way, oftentimes there can be early success with a with a small team. But then there can be some cultural and process barriers, as you're trying to standardize to the team level and multi team level before you sort of figure out the kinks in in deploying it at an organization wide level. For me, you know, one thing that I thought about was early in my career when I was an engineer, and our organization was adopting agile, one of our adjacent teams had adopted agile very successfully, and was real evangelists of that way of working. But when when it was passed to our team, you know, definitely there was a lot of skepticism. At first, it was a different way of working, there were a lot of bumps in that road to get success. But ultimately, we did. And I think that's one of the trends that we, you know, observed a little bit in this survey is that there are some cultural elements to getting success at scale, with something like feature management and the opportunity as an industry to support organizations as they're, as they're making that quest to standardize a practice like this, like any other cultural practice,

 

Heather Joslyn  21:28  

as we all know, like, every technology problem is a person people problem. It's at its core. How does LaunchDarkly handle feature management,

 

Unknown Speaker  21:37  

there's a few areas that we really think are important to be successful as a provider in the space of feature management, I'll talk about a few of them. One of them is this notion of supporting developers. And that's not because we think, as we've talked about, it's not because we think that the value is only for developers. But in this particular practice, developers are the ones that are adding feature flags to the code. Our feeling is that if that processes as easy as possible, and the cultural barriers, barriers, and such are as low as possible to adopt feature flags, as much as possible that the rest of the company benefits from that, as we've talked about, a personas like product managers, then have visibility and ability to to participate in the process. So you know, this, this notion of ensuring that it's easy for developers is really important. We offer the widest array of SDKs. In First Class API's, we make it as easy as possible for developers. The second I'll talk about is just around governance, which is not actually, you know, maybe the sexiest topic to talk about. But we do think it when it comes to something like this, and involves changing functionality in production for end users. Finding this right balance of giving autonomy to development teams, with the right guard rails is is such an important factor. For us. We've built a very comprehensive workflows capability that automates a lot of the release process, but also provides guardrails in the form of approvals and integrations with tools like ServiceNow, to give organizations confidence that they can deploy and adopted at scale. One more that I'll mention is just around just the architecture for solutions like this. You know, when you think about the fact that these feature flags can be adopted, in a server in a mobile device, client side, development is becoming more and more popular, they can be processed anywhere in the world. And so an architecture that is supporting feature flags, and the delivery of feature flags, with security with reliability, consistency, I mean, it's a real undertaking to build a service like this and, and LaunchDarkly is really invested in a global delivery network that gives our customers confidence that they can deploy a feature flag, in any device in any situation can be successful without,

 

Heather Joslyn  24:15  

is there anything new on the horizon, and LaunchDarkly that you'd like people to know about or something coming up that you think will be you can talk about, that you think people might be interested to hear about.

 

Unknown Speaker  24:26  

We've talked a little bit about measurement and, you know, this notion of rethinking measurement, but also staying core to some of the ideals of DevOps. We've invested a lot in our experimentation capabilities, and recently, a lot some major improvements to our experimentation capabilities. And at the core of that is, you know, this notion that experimentation can be a lot more successful when it's tightly integrated to the developer workflow, and we've been really successful in support Going developers. And now with our experimentation capabilities, we really think we're going to be able to empower the entire development team, from developers, to product managers, to their executives to really see the impact of of the features that they're building to optimize those features. An example that I'd like to talk about is one of our customers, CCP Games. They're the maker of EVE Online. It's a massively multiplayer gaming platform with millions of gamers on there. They were recently, you know, thinking through how to evolve their recommendation engine, because they wanted this engine to recommend actions for their gamers that will hopefully increase their ultimate Northstar metric, which was not something like deploys it was session time, like how much time are the gamers engaged in the game. And so using our our experimentation capabilities, they were able to AB test in different permutations, and really make an impact in terms of the session length that gamers were having with the game and the enjoyment that they were having with the game. So that's the kind of capability that we think is going to be an increasing priority. And we're continuing to invest in more and more capabilities that give our customers the ability to measure their development efforts.

 

Heather Joslyn  26:20  

Excellent. Thank you very much. I guess that wraps it up for this episode of the new stack makers. We want to thank Ravi Turzai, for joining us today. And we also want to thank LaunchDarkly for sponsoring this conversation. I'm Heather Jocelyn from the new stack on behalf of makers. See you next time.

 

Colleen Coll  26:39  

Unleashed developer productivity for the software powered world by fundamentally changing how you deliver software to your customers with LaunchDarkly ease feature management platform empowered developers can empower the business to release new features faster and more efficiently than ever.

 

Alex Williams  26:57  

Thanks for listening. If you'd like to show, please rate and review us on Apple podcast Spotify, or wherever you get your podcasts. That's one of the best ways you can help us grow this community and we really appreciate your feedback. You can find the full video version of this episode on YouTube. Search for the new stack and don't forget to subscribe so you never miss any new videos. Thanks for joining us and see you soon.

 

Transcribed by https://otter.ai