Subscribe
Share
Share
Embed
The WorkOps Podcast is your weekly conversation with HR leaders and People Ops practitioners doing the real work.
In every episode we dig into one story. A process that went sideways, a system that just didn't work, and what someone actually did about it. Packed with practical lessons you'll want to bring back to your team. Whether you're supporting 500 employees or 5,000, this is how the best People leaders are building for what comes next.
David Hanrahan (00:00)
I got to the point at Zendesk where I was just so frustrated with it. I actually told the team to stop doing exit surveys.
Jeet (00:07)
Well.
David Hanrahan (00:07)
And
we just said, let's just stop doing them because the signal on them was so broken and so that an executives disagreed with this. I said, okay, well, we have them, you have access to them. can you interpret anything from this? I wouldn't say it was the best decision, but it was something that I did almost kind of force something for us to rethink how we should approach
Jeet (00:16)
Hmm.
Yeah.
right. Hey everyone. Welcome to the WorkOps podcast. I'm Jeet and today I'm joined by David Hanrahan, who is the SVP of People Success at SolarWinds. And every episode we dig into one story, a process or a system that went sideways and that just didn't work and what someone actually did about that. So David, very excited to have you on the podcast today. Thank you for joining us. Before we jump into things, can you tell us a little bit about yourself? How did you
choose HR.
David Hanrahan (01:29)
Yeah, so hey, thanks for having me If I go back to the origin story, I was my undergrad with psychology and I picked it within about five minutes of being on campus. Someone said, what's your major? And I just declared it because that was the only thing I could think of. And so for the next four years, I accidentally got into this because of an orientation mistake I made.
You know, along the way, I love human psychology, abnormal psychology, industrial psychology was something that I just sort of ⁓ happened into as a class. And I just didn't realize that there was a psychology of the workforce and there was practitioners here and there's both like a research element to it. But then there's an application of it that, you know, in many companies at the time, it was called human resources. And I went to a grad program originally interested in on the industrial side, like unions and union organizing and. ⁓
collective bargaining and these types of things. And so my first company was a big oil company, but I moved into software companies and which is a little bit different, know, it's fast moving, it's innovation. now SolarWinds, call it people success. And I've been at SolarWinds since January, 2025, so about a year and a half now.
person company offices around the globe on base in Austin. So anyways excited to be here
Jeet (02:48)
Very excited to have you here. yeah, let's talk about the title a little bit before we dive into the meat of the topic. So people success. Why people success? It reminds me of customer success in some ways.
David Hanrahan (03:01)
Yeah.
I think that there should be a connection back to what we're trying to accomplish within our people function, how all the teams and ultimately the success of our employees ultimately translates to the company success. So I think human resources is an older term. It's a little bit clunky term, but I think as we operate at SolarWinds at our core in our team, people success, everything that we're trying to do, we're a service organization.
to the rest of the company, much like customer success as a service organization to customers, we're trying to enable all our Solarians, our Solarians employees, to be successful, not only as individuals, but in teams. so thus, if our employees are successful, we are successful in our department.
Jeet (03:50)
Love that view. It's not internal department facing, but it's what your end users, taking a product management term from there, how you want it to succeed. That's great to hear. So let's talk about a process or a system that didn't work for your end users. The employees doesn't necessarily have to be at SolarWinds at a previous unnamed company. If you want to leave it at that, let's... Alrighty, go ahead. Which company was it?
David Hanrahan (04:10)
Yeah. I'll name the company. I'll go ahead and name the company.
So yeah, so the kind of a process or workstream that wasn't working when I was at Zendesk many years ago, we would do exit interviews or exit surveys. so I've had a love-hate relationship with this workstream, exit surveys.
And when I started there, I kind of found that we had a very fragmented approach to collecting. You know, it's an important moment. Someone's moving on. And ⁓ in many cases, you know, we have regretted attrition. We have someone who's moving on that like, what happened? Why did this person move on? Maybe the manager knows and that employee knows, but we want to learn from it.
And so we would do exit surveys and in some parts of the world at Zendesk, they would do it via email. They would have someone write, you know, write back, hey, here's why I'm leaving. And we would just copy and paste that and put it somewhere. It would go to a, you know, a shared folder somewhere. Elsewhere we had ⁓ a workflow. We had an automated task coming out of the HRIS system. It was very fragmented. And so when I just wanted to look at, you know, I'd see attrition and it's very natural.
thing for a people leader or even executive, like, I want to look at the exit surveys. We didn't have AI. didn't have like sentiment analysis. didn't have a way I could sort of like quickly take all the text in one place and analyze it in one way. So I'd have to read through these things. And I just found immediately they were all over the place. We'd have exit codes and like we were forced to choose an exit code. This person's leaving for career development reasons or compensation or, and you could, you could immediately look at the reason codes, the percentage.
of our past year's attrition, why people were leaving from the exit codes and that percentage and the overall voluntary or regretted attrition, you could kind of infer an understanding or a belief as to, so mainly people are leaving for career development reasons. And so that means they were going to another job that they wanted or a step up in their progression. But then you look at the exit surveys themselves, the fragmented ones I could get my hands on and see something is not...
aligning here. And in some cases, it seems like I wasn't getting the full story or, hey, why is this exit survey very long and this one
who I know that employee was really important. We barely got anything from them. And as you just start to pick it apart, there's all these very natural reasons for why it wasn't working, which is there's an element there to the person who's leaving of how candid do I really wanna be? And some of your best employees were feeling as though like...
Jeet (06:38)
Hmm.
David Hanrahan (06:53)
The only thing I want to do here is leave on a good note. So I'm going to give you some very canned, non-controversial reasons. And the employee who is very disgruntled.
Jeet (06:58)
Yeah.
David Hanrahan (07:05)
he's gonna tee off on you and give you all these reasons. And then you might mistakenly infer from the handful of employees who are like most vocal that like, hey, that's the big issue. This is the big issue that we should attribute to the majority of the reasons why people are leaving. And that there's a broken logic to that.
I I got to the point at Zendesk where I was just so frustrated with it. I actually told the team to stop doing exit surveys.
Jeet (07:29)
Well.
David Hanrahan (07:30)
And
we just said, let's just stop doing them because the signal on them was so broken and so bad that an executives disagreed with this. I said, okay, well, we have them, you have access to them. Do you know, can you interpret anything from this? And that was, you know, I wouldn't I wouldn't say it was the best decision, but it was something that I did almost kind of force something for us to rethink how we should approach
Jeet (07:39)
Hmm.
Yeah.
David Hanrahan (07:54)
So we pivoted on it eventually into a more consistent approach, a more system-based approach, a little bit of automation to it. But it is still a process that we run of all the different processes, know, and people ops and work streams. It's still a process I think in many ways is broken.
⁓ and debatable as to whether you're getting, you can get the right signal from an exit survey versus your engagement surveys, which are going to be much more of, I'm here and I'm going to, I'm not leaving, but I'm going to answer the survey, which can give you forward looking signal on when, when, when and why people might leave, which might be more, more valuable than your exit surveys. But, but it was exit surveys was the broken process. And, you know, we, we attempted to fix it eventually.
Jeet (08:14)
Hmm.
Yeah.
I guess you've to watch out for who the author is and the bias, whether they've had a good experience or a negative experience, that's going to heavily weigh into what they put in, right? And then what they go ahead and put into Glassdoor as well after they've left. That's another piece.
David Hanrahan (08:54)
Yeah, yeah, 100%. Another broken element, think, Glassdoor. But yeah, there's a lot of biases that can play out that can cause you to pivot in the wrong direction.
Jeet (08:58)
Yes.
Yeah.
Do you think that it is just something that cannot be fundamentally automated to the full extent or is, people just not focusing on it?
David Hanrahan (09:20)
⁓ It's a good question. There's a psychological element that happens in many companies that when the person has left, we've already told ourselves a story. There's already a story unfolding.
Jeet (09:31)
Yeah.
David Hanrahan (09:33)
The individual sat down with their manager and said, hey, I've got some bad news or, hey, I've got some news for you. And then that person told the manager a story and then the manager relayed that story to someone else. then HR learned about it. And then another story, like a myth, a fable emerged as to why this person was leaving. In some cases, we say this person is leaving and they're really good. But through the the retelling of their reasons, then we tell ourselves, well, it's the right thing that they move. It's a good thing. You know, we don't regret it, you know.
but hey, a couple of weeks ago you said they were the highest performing employee, but now it's okay. And the leader wants it to not look bad on them. And so it becomes retold. Can you automate this in a way that creates better signal, is more valuable? I'm kind of, I'm not certain on it. I think you can adapt and change it.
Jeet (10:27)
Yeah.
David Hanrahan (10:32)
in a way that maybe is more valuable and what you do with the data and the experience of it and the experience of the departing employee. Because if you're going to ask me to do this thing.
What do I get out of it? But also does it help make the team better? Do we know internally others have left? I never heard about anything from their exit surveys in the team. Never heard about anything that these other employees who have left has anything happened, you know, and do we hear about anything there?
Jeet (10:55)
Hmm.
Yeah.
David Hanrahan (11:03)
the outputs of these things and for exit surveys, it's one of those work streams where for many companies, many people leaders listening, it does go into this black box of sorts. goes into this, like there's a lot of data sitting somewhere and it's almost as if like it's stigma. It's like, we don't wanna open this up because it's kind of all the skeletons in the closet if you will. And so I think there is a...
It's a complicated answer in terms of, you can probably make it easier. You can probably make it better. You can probably make it experientially something that feels more, maybe more transparent. ⁓ But it's kind of TBD as to whether you can actually get the right signal from someone who's already made their decision to move on and then use it for good. Let's see.
Jeet (11:40)
Yeah.
Yeah,
I guess there's three vectors from what you're saying. One is the employee experience because they are going to carry that on with them when they leave and they're going to talk about the company that they left. So there's that experience, which is really important, but perhaps a little bit hard to measure. And to your point, it's kind of too late by that point because they've already had that in their memory. Then you've got the team experience. And I've been in companies previously where you suddenly see a seat that's empty and you're like, what happened here? Why is that seat suddenly empty?
⁓ And then you have the leaders experience where they need the data to be able to prevent it from happening in future where it's regrettable. And I guess with AI these days, there is that fine balance even more. So if like, what do we keep automated or leverage with AI and what do we keep human? Because an exit is a sensitive period for all of those three functions.
Where are you kind of seeing that change now with AI? Are you starting to see, okay, well, let's use it for the data analysis. Maybe let's allow employees to chat to an AI so they can freely express their thoughts and we can analyze it after that point. Are you seeing anything cool being done on that side of things?
David Hanrahan (12:56)
I'm definitely seeing a shift in terms of the sentiment of interacting with AI for what I call more guidance and judgment. And there's a comfort level of if I can interact with an AI, whether it's an exit survey or a tough question, I want to ask, you know, call it the manager bot, you know, or where I don't, I'm not sure if I even need to go to HR quite yet. I want some guidance on this as an employee or as a manager.
Jeet (13:15)
Hmm.
David Hanrahan (13:25)
I think there was maybe a thought within people in management circles that people are not going to want to interact with AI for...
complicated sensitive matters. I think that's actually not the case. I think it's the opposite. I think that you're going to be more comfortable interacting with AI. We use an enterprise AI application internally. And when you think about all our corporate productivity tools, as well as people tools, I use this enterprise AI as a sort of advice, sort of synthesis of information.
Jeet (13:39)
you
Yeah.
Hmm.
David Hanrahan (14:07)
understanding what's working, what's not working in the company. That is effectively asking it sensitive questions for guidance and support. So I think that is a shift. ⁓ We're adopting ⁓ an enterprise ⁓ AI agent in our ⁓ HRIS that goes live in a couple of weeks, ⁓ which you can ask it who's the president ⁓ of the United States, but you can also ask it, you can ask it things that is knowledge that exists outside of the company, but also knowledge inside the company.
Jeet (14:24)
Nice.
David Hanrahan (14:37)
you can ask it to transact. so there is that shift. think employees are getting much more comfortable interacting with AI where they previously said, I need to go to a human on this task.
Jeet (14:49)
Hmm.
Yeah, it's interesting, isn't it? And I totally agree with you. We're seeing it as well where employees are actually okay to ask those sensitive questions themselves, as long as they are confident that their sensitive questions are private and secure and not shown to the people that it shouldn't be shown to. And interestingly, we've heard stories where actually the change management starts from folks in the people team.
because a lot of folks in the people team have joined because they want to be helpful. They want someone to do their best work. And sometimes it's really easy to say, I'll just quickly answer that question because I know it. So I'm just going to take me two minutes. It's cool. And it's really hard to break out of that muscle memory of doing that instead of saying, let's give the AI a shot. Have you seen that too? And how have you kind of, or maybe you haven't, but if you have, how have you equipped your team now or previously to
adopt and drive that behavior change so that they then encourage employees to also then use the tools that you've invested in.
David Hanrahan (15:52)
It's very timely because one of our first big initiatives is going live soon. And I was just talking with our people tech lead about we need to see once we launch this whether people are using it and how they're using it. But if I go to other enterprise AI applications,
Jeet (16:04)
Hmm.
David Hanrahan (16:08)
I have seen for one in particular that I would say is not squarely a people tool. It lives in the IT stack, but it's the most popular application. People are clamoring for getting a seat on this. And there's a usefulness of it ultimately that kind of drives adoption.
There's a usefulness but also trust. On the people think that we had an experiment with contractors. meaning like one of our, like when you look at tickets, one of the top tickets that we had that came into the people team was needing assistance with a contractor item, onboarding a contractor, onboarding a contractor, getting a rec approved or a change or what have you.
Jeet (16:55)
Hmm.
David Hanrahan (16:57)
And we experimented with an agent that was just about answering questions on contractors. And we wound up shelving the pilot.
because we realized it was too narrow and it wound up being something that I'd asked the question, but I would need a clarification from a human. would just wind up going to the human anyways. And it wasn't really redefining the workflow per se. It was answering narrow questions that was interesting as a concept, but it would go into another topic. It would go into a payroll topic or it'd go into an advice aspect or would say, you need to talk to a human now, ultimately. And I think as companies,
Jeet (17:22)
Hmm.
David Hanrahan (17:36)
you know, sort of remap their workflows and where you want to have an agent. You also have to understand where a human still needs to be the judgment or discretion layer of like, I'm going to get advice and I still need to talk to the human and sort of mapping that out. so that it is useful, number one, but also I trust it. And you know where the human needs to play, he needs to continue to play an important judgment role. That was our first foray into this. We kind of realized it was too narrow.
Jeet (17:51)
Hmm.
David Hanrahan (18:04)
was going into other topics. We didn't really rethink the workflow per se, and so we had to go back to the drawing board.
Jeet (18:12)
I'm
glad your example wasn't about an exit survey, but it was about contractors. that's good. At least it sounds like the exit survey process is working better at SolarWinds, which is nice to hear. Yeah, good. And did you guys build the agent internally using internal tools? The one that you mentioned around?
David Hanrahan (18:23)
Yeah, yeah, yeah, I've learned. I've learned from some mistakes.
We
built the contractor agent using an internal tool. ⁓ Demos, just demos of applications that are out there, ⁓ agents and such, ⁓ informed us quite a bit of what good looks like and the innovation that's happening in the market around agents was teaching us how to think about this. But we built it internally and... ⁓
Jeet (18:35)
Nice.
David Hanrahan (18:58)
And now ⁓ we're likely to continue to building agents off the enterprise layer, but also experimenting with agents that live within specific applications.
Jeet (19:08)
Excellent. it doesn't actually sound like a build or buy. It seems like you guys have bought and now you're maximizing the value out of the existing investments and unlocking agents within those systems. Is that kind of the right way of how you guys are thinking about it?
David Hanrahan (19:23)
That's right. That's right. don't think there's a right answer on build versus buy, but more of what are we trying to accomplish? was something, I think our CEO said this on a recent conversation we were talking about spend on AI and the tools that we're adopting. And the CEO said, well, why are we doing this? Like, what's the purpose of this? What's the end game? Do employees understand it?
Jeet (19:30)
Hmm.
David Hanrahan (19:48)
and understanding and being aligned on it in terms of what we're trying to accomplish here. What is, is this a productivity? Is this a bandwidth? Is this sort of upleveling that function to be able to take on higher order challenges that we think they're more value add for the humans?
We learned, we had to course correct on this more recently, that one of our particular AI initiatives, we just didn't really have the why, the thesis aligned at the top level. And we had to get back to that and then get better at communicating it to the company who are excited about it. But ⁓ we need to be able to explain internally, because you have AI for your roadmap, for your customers. We have products, we have services where AI is externally for the customers. But then the internal use case, we just were moving too fast and we didn't even think about the why and the end game.
Jeet (20:27)
Right.
David Hanrahan (20:35)
to go back to basics.
Jeet (20:37)
I love that using words like learned and course correct, it shows that kind of iterative mindset. There's always a tension that we hear of like, hey, if we are going to experiment with this, it's going to take away time from a team that is already overly stretched because that's what typically we see people teams are at. How did you balance that? How did you manage expectations with your team, with others within the business of, we're going to have to maybe de-prioritize some existing stuff to be able to free up that bandwidth
to learn and to course correct and potentially fail at times.
David Hanrahan (21:11)
Yeah, well, I'll take the business partner team. I've talked with our business partner team about being the front end of strategic HR. And that's a buzzword that gets thrown around.
But what do I mean by that? a business partner, our business partners are incredibly busy. They're stretched thin. They have a lot of ⁓ tough stuff they deal with. They have a lot of manual requests. They have things that hit them that take their time. And if I talk with them, I realize I need their help on something that is about maybe coaching a leader or an org design or something very complicated and higher order. is a strategic HR here is us leaning in on this, you the go to market business, the objectives that they have. And there's talent implications here. There's an org.
design
and it's like, okay, well, David, where's the time, where's the time going to come from for me to be able to engage on this and actually engage proactively as opposed to just being pulled into a meeting and not being ready for it. Cause I have all these other things to do. I'm working on transferring this employee across the globe and there's all these different complicated work streams with payroll and immigration, all these different things. There's, you know, there's all these requests I have to get back to very simple questions from managers. That's taking, taking all my day. I can look at their calendars and they're in back to back meetings. And, um,
I've talked to my business partner team about this is one main way we're going to adopt tools, we're going adopt agents internally that can take off those black and white.
⁓ questions you get, but also work streams that can be automated can reduce the amount of strain that you have to something where you're going to have the most value and impact your advice, your judgment at the end. And so I've talked to them about this is the way, this is how we're going to achieve being strategic HR and strategic HR business partners by reducing that workload. And that's the thesis as opposed to I'm trying
to reduce the team or I'm trying to like cut costs or whatever it's but but being clear on that I think is also an important trust element for the business partner team to engage and and and say like hey I'm comfortable relinquishing some of these questions these tickets these types of I'm comfortable relinquishing these if I understand what you're what you what you want from me after that
Jeet (23:06)
Nice.
Love it, love it. And ⁓ Wendaz, taking us back to being in the conversation, when are you going to start to build an agent for exit surveys at SolarWinds?
David Hanrahan (23:33)
It's a good question. We have this somewhat automated right now through a system that we use, a particular platform. And what I'm a little bit more interested in is flipping the exit survey workstream on its head.
Jeet (23:41)
Great.
David Hanrahan (23:48)
which is more about getting the early signal and inferring an exit less from, need you to fill out a survey to tell me why you're leaving, to more of we have understood through real dialogue that is happening and issues in the team in real time to attribute a signal that you can then, as a manager, can correct. I would rather reduce and eliminate the need to fill out a survey, by the way, in engagement surveys too.
Jeet (23:52)
Hmm.
David Hanrahan (24:17)
I would rather eliminate the need to ask questions and more of I understand contextually why this person is leaving because of the conversations that have already been taking place. Now that is a big brother privacy, you know, it's a challenge because if I know that someone is listening,
and is already understood, this is predictive attrition, right? Like we can understand why people are leaving because there's dialogue that's happening. My hope for exit surveys is ultimately that they do go away because there's a system that we've reduced bias on, that we've created like more validity to ⁓ the contextual understanding that will tell us this person left because of these reasons. That's a big bet.
Jeet (24:44)
Hmm.
David Hanrahan (25:04)
But if we have Teams, Slack, emails, Zoom, transcripts, there's dialogue that's happening, the reasons why someone's leaving is right there. And I shouldn't have to ask in a survey, like, you know, in their last day.
Jeet (25:11)
Yeah.
Yeah, there's so much risk, rich context there. And of course, as you mentioned, you've got to balance out how much is meant to be private versus how much can be shared openly for the benefit of the person and for the business too. We know some folks who have built their kind of own retention risk formula and they're kind of 85 to 90 % been correct so far. they...
they're kind of missing that last step of that context that you're mentioning of the conversations that's happening. But it feels like there's good foundations to build on. And I really like the approach of no more surveys, please don't ask me because I'm going to give you a very, ⁓ very specific tailored answer as opposed to probably the honest truth, right?
David Hanrahan (25:53)
Yeah.
That number you cited, 85 to 99%, I just came across that same stat, which is data models can predict attrition, data models can understand it with that level of accuracy if it connects to the right sources and it has the right understanding. And I think for employees, back to the glass door example, there is a value if we can democratize this information, we can make it transparent and clear.
Jeet (26:07)
Hmm.
Yes.
David Hanrahan (26:23)
that like, if I'm going to go to Glassdoor and put you on blast anyways, why don't we actually just create that transparency internally of like, Hey, here's why we're losing people in this team. ⁓ and, the manager is, you know, like we're going to lift, lift the veil, so to speak, which is an uncomfortable moment for them. But if we can get to a level of validity and transparency internally, I'm less, I'm less apt to just go tee off on you in Glassdoor because I can see the virtuous circle. There's a dialogue happening. It's accurate. That's the reason why.
Jeet (26:33)
Yeah.
David Hanrahan (26:54)
And I'm not really as interested to go put you on Glassdoor two weeks later, two months later. And that's a belief I have that employees want to know the problems in the team. They want to know what's happening here. And the manager is a little bit challenged to their role in it. How much do I want to impart back? I want to tell you a story. I'm a little bit.
biased myself, I have a discomfort here and what I want to share. If we can get, if we can move past that, I think there's a value to employees in understanding what's going on in the team that I will be much more apt to speak candidly in whatever forums the tool is listening to. I'll be more candid to speak accurately. And if I see there's value in this for me as well.
Jeet (27:37)
Yeah.
Very good. think we're just, we're just calling for the death knell of surveys and maybe even a glass door on this podcast. Yeah. Well, David, totally, totally. Well, well, before we wrap up, David, any, any final thoughts that you want to leave the listeners with who are in the middle of a process that they want to optimize, particularly in the, in the world of AI that we are in now.
David Hanrahan (27:45)
Sure, sure, yeah. Yeah, they'll love me for that.
Well, I wrote something yesterday. I think in our roles, we're going to find ourselves having to rethink everything with agent at HR and embracing that and starting in one particular area versus being overwhelmed with like everything has to flip on its head tomorrow night.
you know, having one area that you start with where you move to agent H.R. or automation or adopting A.I. that we think can have the highest impact and discussing it, aligning on it with your team, understanding how this helps also contribute to the business and thinking first as a business leader and second ⁓ as a people leader, thinking of yourself a little bit more as a technologist than a psychologist, you know, in this this era that we're moving in is some advice that I've
Jeet (28:43)
Hmm.
David Hanrahan (28:51)
embraced recently and might be helpful for others.
Jeet (28:54)
What a way to finish. David, thank you so much for joining us on the WorkOps podcast. And to everyone listening, we'll catch you on the next one. And if anybody wanted to connect with David, we're gonna share his LinkedIn profile in the show notes. So you can find them there. Thank you, David. Thanks everybody.