Description
Watch Tia Xuan (Lead Software Engineer) and Spencer Ball (Senior Solutions Engineer), Disney Ad Platforms, share how they rebuilt their Salesforce release process — including the specific pain points that pushed them to change, and how they went from manual, risky releases to fast, auditable ones.
Transcript
Taylor Volk, Gearset: All right, we'll kick things off here, and thanks for those who are joining. We are recording this session, so for those joining after the call and watching this when we're not live — welcome. Excited to be kicking things off, and thanks for joining today's fireside chat.
Taylor Volk, Gearset: I'm Taylor, I work on our enterprise sales team here at Gearset, and I've been working alongside the Disney account for well over a year at this point across various teams. Excited to be joined here by Tia and Spencer at Disney Ad Platforms. I'll give them a second to introduce themselves, but the focus of today's session is really to understand your journey with Gearset — where you started, and
Taylor Volk, Gearset: how the process for DevOps has changed
Taylor Volk, Gearset: since being with Gearset, and then what's next.
Taylor Volk, Gearset: Our goal for those of you watching is really to share a practical, experience-based story, rather than what you might see on a typical demo on our website — really what release management looked like for this team before they started with Gearset, and what's ahead for them.
Taylor Volk, Gearset: We'll leave a few minutes at the end for questions — feel free to drop those into the Q&A as we go. I'll set the scene here with Tia and Spencer. Would you mind telling us a bit about your role and your team, and what you're responsible for day-to-day?
Tia Xuan: Sure, I'll start. My name is Tia Xuan, and I'm a lead software developer here
Tia Xuan: on the Ad Platforms team. I run Sales Hub, which is our CRM on the sales side for ad platforms. We go from development and intake
Tia Xuan: to running the pipeline and DevOps, and deploying
Tia Xuan: to our production instance — that's generally what I do, the end-to-end of Sales Hub.
Spencer Ball: Hi, I'm Spencer Ball. I'm a system admin working in Sales Hub. I help deploy releases to production.
Taylor Volk, Gearset: Awesome. Thanks, Spencer. Thanks, Tia.
Taylor Volk, Gearset: I'd love to set the stage — for those watching, this is true of a lot of Disney teams who've moved over from Copado to Gearset. Disney Ad Platforms was using Copado similarly right before coming onto Gearset — I'd love to get a sense of what the deployment and release process looked like
Taylor Volk, Gearset: before you started with Gearset. Tia, do you want to take that one?
Tia Xuan: Sure, so our release process was very manual before we got into Gearset. Generally,
Tia Xuan: Copado is very based on what's in the org. Sometimes there are features that are still in flight, so they'd sit there for a while before going into production.
Tia Xuan: That's generally because our features are really UI-driven, so it's harder to do dark releases with UI features.
Tia Xuan: So we let things sit in UAT for a long while, and then we have our staging environment, which is the midpoint between production and UAT.
Tia Xuan: That allows us to put all the fixes together, component by component, because when we tried to do it
Tia Xuan: a more automated way, we were always finding different errors or blocking points. So we found it easier to do that component by component. Sometimes we'd have many stories that would have to go out at a time, and that took up a lot of time to merge everything and make sure everything that
Tia Xuan: shouldn't be there, isn't there.
Tia Xuan: So that allowed us to really —
Tia Xuan: we'd spend at least a day or two trying to package everything up, then testing it in that environment, making sure we didn't miss anything before going to production, so —
Tia Xuan: it was a very lengthy process for us.
Taylor Volk, Gearset: Lengthy process, for sure. And you mentioned that when you're pulling things out of UAT component by component to make sure what you're sending to staging or production is what you intend to — how many people had to be involved to get that release out the door?
Tia Xuan: It was basically the whole team, because depending on the story, it would be worked on by different team members. I would generally do most of the releases, but I'd pull in different team members depending on which story
Tia Xuan: had problems.
Taylor Volk, Gearset: I remember when we first met, you having to play many hats leading up to release days, for sure.
Taylor Volk, Gearset: Can you talk us through a specific release that went wrong while you were using Copado, what happened, and what the knock-on impact was for the team — or specific scenarios that were often pretty difficult when using Copado?
Tia Xuan: If we had a small release, it was usually fine — we'd take a few different components, package it up in staging, and send it on its way. But anything bigger required a lot of merging, a lot of
Tia Xuan: effort to figure out what was changed and what wasn't for different components.
Tia Xuan: So that was a little more
Tia Xuan: work to do — but generally our releases were a bit bigger, just because we had a lot of things sitting in UAT, so that wasn't ideal for us.
Taylor Volk, Gearset: It sounds like you'd set up a bit of a workaround to make sure what you actually wanted to go to production would go to production — was that with your staging environment? Can you share a bit about that?
Tia Xuan: Yeah, before we had a staging environment, it went straight from UAT to production.
Tia Xuan: But since we have features still in flight that we didn't want to release, and as I mentioned, a lot of UI components that couldn't go out in a dark release, we
Tia Xuan: decided to have a staging environment where whatever's going into production sits in stage.
Tia Xuan: That gave us a place to package things up, test them, and have a package ready that would be exactly what we deploy into production. But
Tia Xuan: as I mentioned, it was time-consuming, since each component of each ticket had to be merged there separately.
Tia Xuan: Oh.
Taylor Volk, Gearset: True.
Taylor Volk, Gearset: And prior to introducing that staging environment, or even after — were there times you'd package something up for release and later realize a component was missing, or something was wrong with that release? What would happen when that occurred?
Tia Xuan: If there were ever issues with deployment, we'd have to find creative ways to either fix it right away or deploy again. That's part of what we do when we smoke test. We had
Tia Xuan: testing in the staging environment, but it wasn't complete because we have downstream integrations, and sometimes we didn't catch those since there's no integration environment in stage — only a subset
Tia Xuan: of the environments are available in stage. So
Tia Xuan: sometimes it just went directly to prod and we had to test in prod, which isn't ideal either, because then we'd have to bring in components and validate things mid-deployment — so deployments could sometimes last hours, because every time we went into production
Tia Xuan: we had to run our test classes, which takes a while.
Taylor Volk, Gearset: Etc.
Taylor Volk, Gearset: It sounds like there was a lot of manual intervention leading up to releases, and then on release day. What did that process look like, and how easy or not was it to keep your environments in sync
Taylor Volk, Gearset: after a production release?
Tia Xuan: It was non-existent — all our environments were pretty much not in sync. Whenever we had any
Tia Xuan: problems with an environment, we'd have to really analyze what went wrong with it. That part is definitely
Tia Xuan: easier now, because we just tell Spencer to create sync PRs, and those go through to all the different branches, as well as
Tia Xuan: the features, so everything is in sync.
Taylor Volk, Gearset: I'd be curious to hear your
Taylor Volk, Gearset: experience with that in a second, Spencer, when we talk about how things have changed since implementing Gearset. I'm curious about that tipping point for you both —
Taylor Volk, Gearset: what was it that made you go looking for a new way of doing things, once you realized Copado maybe wasn't giving you what you needed for release management?
Tia Xuan: A lot of it had to do with the dev time required to analyze everything that came up with merges and conflicts between environments,
Tia Xuan: and how unreliable it was for testing. Our QA team gave us a lot of regression
Tia Xuan: issues — those kinds of things kept popping up, and we'd have to go into all the different environments to figure out exactly what happened, and sometimes we just didn't know, because of the volume
Tia Xuan: of work we had to do per sprint. So it really was
Tia Xuan: a blocker for the dev cycle.
Taylor Volk, Gearset: You mentioned earlier a few days just to package up a release — what was actually happening across those couple of days?
Taylor Volk, Gearset: Was that communicating with the individual developers who'd packaged the initial change? Going back to business users to figure out what testing issues they were running into?
Tia Xuan: Generally it was with the developers, because they'd know what changes they made — sometimes those changes had to be manually done in staging, which wasn't ideal. And then sometimes
Tia Xuan: it would also mean going back
Tia Xuan: to our testers, our QA, to figure out what was done and what the expected functionality was, then seeing if it still held with the fixes we made in stage. So it was almost like doing a new
Tia Xuan: round of development.
Taylor Volk, Gearset: Yeah. Okay.
Taylor Volk, Gearset: I'd love to switch gears and understand a bit more — what parts of the SDLC, the software development lifecycle, are you using Gearset for today?
Tia Xuan: We're using the pipeline in Gearset. What I like about it is that the source of truth is Git, so anything with Git functionality — basically everything you see in terms of history and tracking — is there.
Tia Xuan: We start everything in our sandboxes. There are developer sandboxes, shared among all developers, and then our dev, QA, and UAT environments,
Tia Xuan: which move up the pipeline.
Tia Xuan: So we have that set up, and it's a pretty smooth transition all the way up
Tia Xuan: to UAT, then staging, and prod.
Taylor Volk, Gearset: Awesome. Spencer, I'd love to hear where Gearset shows up for you in your day-to-day,
Taylor Volk, Gearset: or week to week.
Spencer Ball: Gearset shows up for me deploying releases to production specifically, and with Gearset it's much easier than Copado. I pretty much press a button and it gets deployed.
Taylor Volk, Gearset: Awesome. And after a release to production — once you hit deploy — what does the upkeep or maintenance of lower environments look like for you? Is that your responsibility to keep those in sync, or something the devs handle individually?
Spencer Ball: It's kind of collaborative, but I create what are known as sync PRs after every release, which help
Spencer Ball: restore each environment to what's currently in production.
Taylor Volk, Gearset: So that's a one-click button to make sure each environment is in sync with what you just pushed to main.
Spencer Ball: Yep.
Taylor Volk, Gearset: Nice. Bit of a leading question from us — I know that makes a big difference for you.
Taylor Volk, Gearset: Tia, I'd be curious — in using Gearset to package your changes and deploy them into the pipeline, what about your day-to-day has changed the most significantly?
Tia Xuan: I think it's that it's easier for me to track changes in Gearset, because of Git — I can look at the different branches and their history to see who did what.
Tia Xuan: Easier than before. The source of truth is that, and you're forced to work off the branch rather than individual sandboxes, which makes it easier because you have one source everything comes from.
Tia Xuan: So I think for me, it made things easier just to
Tia Xuan: see where everything is and diagnose a problem — it's a lot faster than before.
Taylor Volk, Gearset: Love to hear that. One of the things we see from a lot of Disney teams, a lot of enterprise-scale teams — this is true of Salesforce generally — is lots of shared components between
Taylor Volk, Gearset: teams, with multiple people touching similar layouts, profiles, or objects — lots of Apex, which is a common component across multiple user stories.
Taylor Volk, Gearset: Has Gearset made interacting with that kind of metadata any easier for you?
Tia Xuan: Yes — one of the big things we got from Gearset is that we can do partial commits. Every single story can have its own commit, so you can pick and choose which lines you want to commit to a particular story.
Tia Xuan: The merges are handled for you throughout the pipeline, so you don't need to
Tia Xuan: go back and take out things that shouldn't be there, because every story is essentially
Tia Xuan: its own atomic set of changes,
Tia Xuan: which is
Tia Xuan: made things easier for us in terms of keeping track of Apex code.
Taylor Volk, Gearset: It sounds like that's made it easier when you're building changes — you can just commit what you've worked on within that Apex file, not the entirety of the file. Has that reduced the amount of merge conflicts you used to see in the pipeline? What's been the impact of that?
Tia Xuan: We didn't really
Tia Xuan: have merge conflicts before, because it was really all or nothing with Copado. Now we can go back and look at all the merges inside GitHub, and it's very specific to our particular feature,
Tia Xuan: as opposed to Copado, where the source of truth was somewhere in the middle between the org and what's in Git — so we weren't quite sure what was actually going to get deployed at the end.
Taylor Volk, Gearset: Okay, fair.
Taylor Volk, Gearset: We've talked about how debugging releases in Copado, and the commits and changes being deployed between environments, used to be a pretty much daily task. What have you started to do with the time you've gotten back since making the switch?
Tia Xuan: We've basically been able to take on more stories and get more work done in general.
Taylor Volk, Gearset: With your source of truth being Git now, what does that change about being able to see what actually went out?
Taylor Volk, Gearset: You mentioned earlier that in Copado, the source of truth wasn't quite Git — it might have been the environment, which made it tricky to troubleshoot. How has that changed with Gearset?
Tia Xuan: It's just a lot easier to figure things out now.
Tia Xuan: Say we put in a feature and we're not sure why it's in the org — we can go back to the branch, see its history, and see what feature actually impacted it and why.
Tia Xuan: Often a question comes up about why something is working a certain way, and as developers we can go over and see the actual flow or Apex class that got impacted. We can go back to that element and see its history to understand which feature impacted it, so going forward we understand why. It could just be a mix
Tia Xuan: of requirements that caused it, rather than us.
Tia Xuan: So that allows more flexibility in analyzing an issue.
Taylor Volk, Gearset: In the few minutes we have left, before we get to questions from the group — what's next on your roadmap for Salesforce? How do you think Gearset enables some of that, going into next year and beyond?
Tia Xuan: From our side, we're still looking at the roadmap for next year, so nothing's set yet. In terms of Gearset, the next step for us is getting a hotfix branch in place.
Tia Xuan: That would give us a direct connection to
Tia Xuan: production, so we can get hotfixes in more easily
Tia Xuan: than we do now.
Taylor Volk, Gearset: Has Gearset enabled that in a way that would have been difficult in Copado?
Tia Xuan: Yes —
Tia Xuan: we would have had to push something directly into the org, rather than doing it from
Tia Xuan: a hotfix branch and pulling it through.
Taylor Volk, Gearset: Greater flexibility, depending on the need from the business and how quickly something needs to get into production.
Taylor Volk, Gearset: You've been on Gearset for several months now. If somebody told you tomorrow you had to go back to Copado, what's the first thing you'd miss out of Gearset?
Tia Xuan: I think probably having one uniform source of truth, Git — it's well-tested and a lot easier to work with. That, and the sync PR, of course.
Taylor Volk, Gearset: Spencer, I'll ask you the same — if you had to go back to Copado tomorrow, what would be first on your list of things you'd want back?
Spencer Ball: I think sync PRs is a big one. Refreshing sandboxes every couple of weeks wasn't very fun.
Taylor Volk, Gearset: And this has let you keep environments in sync without going through the manual integration work each refresh used to take. Is that the big time saver for you, Spencer?
Spencer Ball: Yes.
Taylor Volk, Gearset: A slight change of topic — this has come up with a lot of teams I've worked with, including at Disney. On AI capabilities — a bit of a personal question for you both — when you think about your day-to-day,
Taylor Volk, Gearset: where do you see the biggest opportunity for AI to improve Salesforce DevOps as a whole?
Taylor Volk, Gearset: And are you planning anything there for DevOps on Salesforce today?
Tia Xuan: At the moment, no — we use AI mainly for our dev cycle, using Cursor.
Tia Xuan: For the development cycle, but we haven't talked about it for DevOps yet.
Taylor Volk, Gearset: No worries. That brings us to about five minutes left, so I'll open up the Q&A box. We've got at least one question — if anyone else has one, feel free to add it.
The first one: "Our releases are basically working, the team is live and stable on Copado —
Taylor Volk, Gearset: what did you actually gain that we'd notice?" I think you've shared a few things already, but anything else you'd call out for the group?
Tia Xuan: I think it's just a matter of less manual work for us, in terms of
Tia Xuan: everything — getting to the actual environment-to-environment deployment is a faster process.
Tia Xuan: Starting from dev, through QA and that whole cycle — that also counts in terms of time saved.
Taylor Volk, Gearset: And another question that's come in — what did the transition process from Copado to Gearset look like for you? How did onboarding go?
Tia Xuan: It was actually pretty simple. It took a while, since we had to set up the repos and make sure everything was connected, and train our developers to use it.
Tia Xuan: But it was relatively simple, and Gearset gave us some resources at the start, and office hours so everyone could ask questions as needed. So it wasn't super hard, and since we were already
Tia Xuan: used to working with Git
Tia Xuan: and all the branching that goes with it, it was
Tia Xuan: fairly fast to get through.
Taylor Volk, Gearset: How about from your perspective, Spencer?
Spencer Ball: From my perspective, getting set up on Gearset was really easy. We had a lot of support from Gearset representatives — weekly or twice-weekly meetings where we could ask questions,
Spencer Ball: so a lot of support and help getting set up.
Taylor Volk, Gearset: I think that rounds us out for the questions we've gotten so far.
Taylor Volk, Gearset: Any final thoughts for the group, Tia or Spencer? Thank you so much for everything you've shared.
Tia Xuan: I think if anybody wants to explore this route, they should at least get a call in to see whether it's a good fit for them or not.
Taylor Volk, Gearset: Perfect. Doing my job for me, Tia.
Taylor Volk, Gearset: It's been great chatting with you both. It's been great to see this — I got the pleasure of working with you when you were evaluating Gearset, and it's been great to see the level of maturity you've reached on the platform in such a brief period, under a year at this point.
Taylor Volk, Gearset: Definitely one of our most successful teams at Disney using pipelines.
Taylor Volk, Gearset: Really appreciate your insight — I think that's gone over well with the group based on the questions we got. We'll be able to end a minute early, so everyone can get to their next meeting, or if you're like me,
Taylor Volk, Gearset: logging off — I'm over here in the UK. Thanks so much, both of you. Appreciate it.
Tia Xuan: Thank you, bye.
