Truss The Process

In 2016, software infrastructure consulting firm Truss made salaries transparent across the entire company. Salaries were revealed internally for all employees, from the executives on down. In this episode, Truss CEO Everett Harper and COO Jen Leech talk about why and how they approached their salary transparency project, and how they’ve adapted this system as the company has grown.

For more details, check out Jen’s write-ups of the project on the Truss company blog: “Why We Made Salary Transparent” and “How We Made Salaries Transparent.”

• Truss website | LinkedIn | GitHub | Twitter - 00:30
• Everett Harper’s bio | @everettharper | Jen Leech’s bio | @jennifermleech - 1:20
• Buffer’s policy on open salaries - 3:27
• “The True Story of the Gender Pay Gap,” a 2016 Freakonomics Radio episode - 3:47
• “Demystifying the Gender Pay Gap: Evidence from Glassdoor Salary Data” - 4:52
• Dreyfus model of skill acquisition (Wikipedia) - 6:43

Transcript

(00:00:00) Broken By Design by Clip Art plays.

Wailin: (00:00:02) Welcome to REWORK, a podcast by Basecamp about the better way to work and run your business. I’m Wailin Wong.

Shaun: (00:00:07) And I’m Shaun Hildner. Wailin, how much money do you make?

Wailin: (00:00:12) [Laughs] Ooh, I’m having such a stress reaction.

Shaun: (00:00:16) You know, I don’t really like talking about it, either. It’s one of those topics that’s really important, but really hard to talk about in the workplace.

Wailin: (00:00:23) I was recently reading around on this topic, and came across a couple of older posts on the blog of a company called Truss. They’re a software engineering and infrastructure consulting firm that’s fully remote. And they’d written in detail about how the company took on a big project around salary transparency. This is back in 2016.

Shaun: (00:00:43) Truss wanted everyone at the company to know what their fellow employees got paid, even the executives. They published a spreadsheet internally that showed names, job levels, and salaries. But that was sort of the end result of a year-long process.

Wailin: (00:00:56) It turns out that the team at Truss put the most time and energy into something more foundational, which was setting up a crystal clear system for assessing everyone’s job performance and skill level. It was only after doing this work that they could be really confident that employees were getting equal pay for equal work.

Shaun: (00:01:12) Today on the show, we talked to Truss’s CEO Everett Harper and COO Jen Leech. Both of them are also on the founding team.

Wailin: (00:01:30) I’m gonna have to ask you to go back a few years now, probably five to six years. Do you remember either a conversation or a series of conversations or a moment that started you down this path?

Jennifer: (00:01:44) My recollection is that when we first started hiring employees, we had some conversations around what we wanted to be building and what kind of organization we wanted it to be and how we cultivated that. And one of the things that is shared amongst the three of us is that we wanted the company to be diverse and inclusive. We saw transparency as a key part of having an inclusive workspace. Having the structure to support a fair work environment was key. So that was something we talked about pretty much as soon as we started hiring. And then within about a year of that is when we began to roll the program.

Everett: (00:02:23) I’d say the additional thing for me was seeing how companies, and not just tech companies, but many companies, had differences between men and women, people of color, and whites. The evidence and the data was all there. We didn’t want to be doing that. We also saw that there were well-meaning companies who intended to do the right thing. But because they didn’t take the next step, wound up with the same kind of problem. So we knew we needed to have a little extra rigor, not only in terms of who we recruit and how we recruit, but how do we make sure that people are getting paid the same for the same work.

Wailin: (00:03:05) When you looked around did you feel like there were not a lot of other examples out there of how to undertake this kind of work? Did you feel like you were figuring it out on your own without a lot of kind of external resources?

Everett: (00:03:18) I remember looking around for a couple months to try to figure out, were there other companies that were doing and it was scant. Buffer was probably the most well-known one, but that’s just one company.

Wailin: (00:03:32) So the team dug in. And as they started their analysis, they looked into some statistics and assumptions around pay equity in the US. Here’s one that gets repeated a lot. Women make 70 cents for every dollar paid to men.

Jennifer: (00:03:44) That was super interesting, because that 70 cents and a $1 statistic was the one that was in our heads as well. And so one of the things that I was trying to do when I started digging into that is I wanted to understand, okay, well our goal here is we want to have fair salaries. So I want to understand how big the problem is and what the shape of the problem is, in order to try to rectify that.

(00:04:06) I discovered that it was reflecting the entire United States workforce as a whole. And that included things like if 50% of the women in the workforce were only working part time, that would show as a X percentage reduction in their salary, according to those statistics. So they were literally comparing like part time work to full time work as reference points. Job choice was another one, that people might choose to work in professions that were traditionally paid less that that was impacting, in this particular case gender disparities in pay.

(00:04:40) We wanted to know, people doing the same job, same number of hours per week, literally the same title, are they being paid the same and being paid fairly? And that data was hard to find. Then, subsequently, studies were released that showed specifically what we were looking for which was in our industry, for the specific roles that we were hiring for, for working the same number of hours. Was there a discrepancy? And the answer was, yes, there was.

(00:05:03) So one of the ultimate questions that came out of that was, do people have the opportunity to do the same work. If you could actually, from a societal perspective, begin to correct for that, then that would actually have, by far, the most dramatic impact on pay quality. Within our company, we did form policies around this. Around things like we’re a fully distributed company and we have been since inception. And part of that is supporting people’s flexibility in work, flexibility in location and the time of day that they work. Because we see that as part of being a fully inclusive company.

Wailin: (00:05:41) The team at Truss wanted to figure out what everyone should be paid and make adjustments if there are disparities. But like we mentioned at the top of the episode, they couldn’t get to this point without first defining skill levels that would be specific to their company. Enter the leveling matrix.

(00:05:57) One of the things that is kind of like the centerpiece of this work is the leveling matrix. And can you explain what that is for listeners who might not have heard that term before?

Jennifer: (00:06:07) Leveling matrix is a rubric for establishing the skill level of an employee in their discipline. And one of the really key things about that, for us, was that we wanted to explicitly be very clear about what we were measuring. What constitutes success in this role for us, which may not necessarily be what it is for someone generic in the industry. Before we created a leveling system, we already had a reference point, a shared reference point that we were using to assign levels. We just hadn’t really talked about it or articulated it. We were using something called the Dreyfus model of skills acquisition, which is a five level model of mastery. And then we would assign people a level within that and then say, where do we think within that level? Somebody is 50th percentile towards the top versus the bottom. And then we would use a reference point from the market to assign salaries based on that.

Wailin: (00:07:03) How did you go about putting together the leveling matrix? It seems like it involved a lot of conversations with employees, lot of self-reflection. Can you talk about some of the steps that went into it?

Jennifer: (00:07:13) We had about probably two or three tries at this that were failed before we yielded an approach that was working for us. The first approach that failed was trying to start from our values as a company, and then work that into things that we wanted to see in our leveling matrix. Which, in theory should work because those things are things that we’re trying to incentivize within the organization. But we found that it led to things that were too abstract or too subject to interpretation.

(00:07:44) And then we had a couple of iterations that ended up landing finally on concrete descriptions of behaviors. And really thinking of a specific person that we think is kind of a platonic example of this particular level for this particular role, what are the behaviors that they exhibit? Then we had a small committee, a group of three people that we’d formed to work on this together, and rated each other within the group, I guess, this matrix. And then that ended up yielding, where are the gaps? What stuff’s wrong? What’s unclear? And then we took that iteration, we handed it to everybody in the organization and said, Hey, can you please rate yourself against this and tell us what you think?

Wailin: (00:08:25) Were there certain criteria or parameters specified in the leveling matrix that jumped out as being particularly resonant with your employees, like maybe some things that you normally wouldn’t think would be something that would be used to evaluate performance, but ended up being important and meaningful?

Everett: (00:08:44) I can jump in on one of those, which I think is a little bit more overarching, which is the ability to communicate, in written and in verbal form. For a technical company, a software company, often that isn’t the first thing, especially back then when we were starting, was not the first thing that people would select on. But we knew that, A, collaboration and coordination is really important. B, in a remote-first organization, the ability to sort of speak across clients as well is important. And three, because being remote, documentation is really important. So being able to communicate asynchronously, synchronously in various forms, became a very important piece that that really suffused through the entire leveling and rubric process.

Wailin: (00:09:36) Was it interesting to do this exercise as senior executives and founders?

Jennifer: (00:09:41) I’m smiling because this is something that we actually started running into as we scaled the organization. At that time, when we first created the rubric, we were, I think, only about 20 people. So the distinctions between levels were… it was a bit clearer because since we were a smaller organization, there were only a few people that were very senior. And so it didn’t require much detailed breakdown or nuanced breakdown.

(00:10:07) As we’ve scaled and gotten larger, we’re about 120 people, and we have a larger leadership team. We have to have more clarity around levels within leadership. And that’s actually been the challenge that arose. At first we just had level five, people can deal with complex systems. And that’s just everyone, past a certain level. And now we’re needing to take that and break it down into various degrees of complexity, all of whom are capable of running complex systems. But, to what scale?

Everett: (00:10:41) I think the other really interesting thing with regard to sort of leadership and rubrics is, when making decisions, someone coming up through a particular practice or discipline, whether it’s design or infrastructure, whatever, gets really good at those complexities and conflicts within that discipline. When you switch to a leadership or management function, at a company level, all of a sudden, the matrix changes. And being able to coordinate around larger systems, larger bodies of work, communication. One of the things I remember, for us as founders. All of a sudden, we realized we were always behind on every single communication. And we were like, wait a minute, why is this happening? And then we realized, when you grow an organization, the number of nodes of communication, multiply geometrically. So we had to spend even more time with things like communication across different disciplines in order to catch up simply to the number. That isn’t something that we put in our rubric and didn’t anticipate. And so now we have to add that in.

Wailin: (00:11:49) How would you describe the tenor of the conversations you had with employees as you were figuring out the leveling matrix or matrices? And what kind of existing trust or empathy did you feel like you needed to have in place before you could really have those conversations in a meaningful quality way?

Jennifer: (00:12:10) Even in our committee where we were creating the leveling matrix, we were having very frank conversations around I rate you as a level three against this matrix. We were rating each other and sharing what we came up with. We did have a trusting relationship in that group and that was super helpful. And then in terms of rolling it out to the rest of the organization, the process of doing it from a perspective of here’s a draft and we’re soliciting input was really super important for people to feel as though they were helping to create an artifact, rather than being subjected to the artifact.

Everett: (00:12:48) Overall, I think all the exercises we did, including starting out with a survey saying, hey, we’re thinking about doing this? Do you have strong feelings one way or the other? Is starting out and modeling that your input is valuable. Luckily, most people said, yeah, I’m down. And then every time we hit a communication beat, we said, we’re gonna get back to you in March. When we did that, even if we didn’t have much answers, by establishing that cadence, It grew trust along the process.

(00:13:21) I also think we, as Jen said, we involved a bunch of different people, it wasn’t all just executives, it was across the range of the company. So a junior person could have impact.

Jennifer: (00:13:33) That also had this intended, knock on effect of when we actually rolled out the rubric, people were much more comfortable with where they landed, because they felt like they had a say in it. There were some people who felt like they were correctly leveled. Like, oh, yeah, that makes sense. That’s where I thought I’d be and then there were some other people who were like, I had no idea this was my job. And those were extremely useful conversations to have and coming at those conversations from the perspective of, well, I’m glad that we had this talk. We’re on the same page now. And we can use this to drive in the same direction.

Everett: (00:14:12) I think the other thing is that we started with the why of this, like, what’s the purpose? Like, hey, this is a business issue and we want to have a good diverse company that’s across as many states as want to work here. And we know salary transparency is a part of that. That’s why we’re doing this in order to make sure that we don’t run into this issue. So whenever there was some confusion, we could always come back to that purpose. And I think that built trust is like, yeah, I’m signing up for that.

Wailin: (00:14:42) Was explaining the purpose of the project also helpful in talking with people who maybe felt really anxious or scared about it when they heard about it the first time? If there were people like that.

Everett: (00:14:54) I think we addressed some of the anxiety relatively early and it would kind of pop up and down. I think it can be an exercise in developing trust, not one that’s exposing rancor or division. If it’s done with some intention and connected to a business model. I mean, we want a diverse company. We didn’t want it to, we didn’t want to recruit just in the Bay Area, because there’s not as much diversity as exists in the rest of the US. So, I think probably what has more impact is after we rolled it out. So once we rolled it out, and then we put it into our recruiting and hiring processes, in our career path processes. As people encountered those, particularly folks who are trying to say is this Truss organization for me? If you put things out front, like our values, and salary transparency, putting that out early, it attracts the right people, and it repels the wrong people. Some people shouldn’t be working at Truss, and would be a bad match for everybody. So the earlier we’re able to say, Here’s why we’re doing this, what we’re doing, then people can make really good decisions before we actually engage each other at all.

Wailin: (00:16:06) In the cases where in the conversations, you discovered that there was kind of a big, surprising mismatch, where someone said, oh, I didn’t know this is part of my job, or that I was supposed to be doing X, Y, and Z, or that this was like an important parameter that I was being evaluated on. Did that kind of mismatch show up in the numbers, as well, in terms of the salary and where everyone thought they should be relative to what the data and kind of your own work determine they should be at? I guess that’s like a way of asking what you did about cases where there did seem to be a larger mismatch?

Jennifer: (00:16:43) So we decoupled the conversations from the salary adjustments very intentionally so that we could have those conversations first and untangle the differences in expectations on its own. And then we rolled out the actual salaries, the bands, and all that after. And that was intentionally done to make those conversations less fraught. At the time we did it, since we were only around 20 people, that wasn’t a huge sample size for seeing discrepancies. It turned out that we did not really end up adjusting that many salaries. But we certainly adjusted expectations. So I would say that the leveling alignment conversations were in some cases significantly more corrective but the salaries that we had assigned, were matching what levels we thought people were.

Wailin: (00:17:30) Is it actually a harder conversation, then, to have around expectations, when you realize that this person has been going along for X amount of time, with such different expectations than what you might have as a manager? Is that actually maybe a harder conversation?

Jennifer: (00:17:48) I think so. I think Everett, you actually, you had the ones that were the hardest.

Everett: (00:17:52) Yeah, I think… hey thanks for that in the end.

(00:17:58) I think those are really challenging. And I think, in particular, for positions that are not cut and dry, positions that have sort of a mix of tactics and strategy. At positions where outcomes are really important and those outcomes might not be anything that you can measure in a year, what you do to get those outcomes really becomes fluid. And the information we get about how to achieve those outcomes can also be fluid.

(00:18:35) I’ll give you a quick example. We hired a lead person to do government contracts. And we spent the first summer trying to build out a practice and try and get in a bunch of contracts. We were flailing finding partners. Now if I’d set up a measurement, which would have been the number of contracts per year, the number of partnerships per year, etc, etc. The person who I’d hired would have failed. It was the wrong measure. We had to go into a market, get in as a subcontractor, win one deal, and then build, that’s it.

(00:19:12) And we finally got to a person who broke it down like that, and then we started to change how we’re strategically going to operate. So I think the overall message is, and I certainly feel this as a CEO, I feel like my job changes at least once a quarter. Less so now but certainly in the early days. And it’s about kind of coming up with the rubrics and then being able, not to change them completely, but have them able to be defined in a way that as the measures and the outputs change with good information, the person can grow and develop.

Wailin: (00:19:45) Once Truss had its skill level system figured out, they matched salary bands with the levels, and then they were able to see how everyone lined up. They published a spreadsheet showing people’s names, levels, and salaries. It’s important to note here that no one got their compensation cut. If a salary needed to be adjusted, it was upwards only.

(00:20:04) Why was it important to publish all of the salaries with names for the whole company? And what did it feel like to do that, especially the first time?

Everett: (00:20:13) For me, it was really important because of questions that I encountered afterwards. When we sort of announced it and I talked in a public sphere. A lot of question was, did that include the executives? Did I include the owners? Did I include…? And yes. And I’d say, well, why wouldn’t you do that? Well, you know, duhduhduhduhduh. And I was like, how can you have a value of transparency, and only do it for a certain part of the organization? That makes no sense.

(00:20:40) It might create some different questions, which I think is really what people were afraid of. But I think that that is critical. If it’s a value that’s driving this thing, then it’s got to be across the entire organization. What it felt like was okay, here it is, right. And being real, actually, about, we’ve included people along the way. And we hope it’s sort of a ho hum, which essentially, it kind of was. There was not a lot of real kind of Sturm und Drang at that moment, because kind of people knew what was going on.

(00:21:14) And also being like, okay, everybody’s gonna look everybody’s curious to see, that’s just human. That’s what I remember of that feeling. But I was really pleased that it wasn’t a lot of storming afterwards.

Wailin: (00:21:27) And so does that list now get refreshed, like every year?

Jennifer: (00:21:29) We update it every time we adjust salaries, which is two to three or four times a year.

Everett: (00:21:34) This is happening anyway. And I know there’s a lot of folks who were surprised by this. But folks, particularly in and I don’t want to sort of generalize across generations. But some of the younger folks are telling each other salaries all the time. In New York, I heard the story that there’s a thing called a salary reveal party, where everybody has a get to know you, and they have their name on the front what they do, and on the back is their salary. And at a certain time in the party, they turn it over. That stuff happens all the time. So the cat’s out of the bag. If there’s leaders who think that people don’t share each other’s salaries, I’ve got some news for you.

(00:22:12) So how are you going to do that and address that becomes really the challenge I would give to anybody who’s listening, because it’s happening out there. And so, connected to the culture and the business model.

Jennifer: (00:22:24) In a way I’ve come to think of not publishing salaries is putting your head in the sand. What we found is that by just having it out there, that this conversation is off the table, and we can move on to other problems and issues. And if one of those problems is correcting for inequalities in salary, and that’s problem one to solve.

Everett: (00:22:45) For folks who are small, it’s much easier to do. It’s harder for a large company. So I’ve talked to folks, and like, hey. Start with a division. Start with an office. Start with something that’s relevant, but actually make sure everybody in that situation has their salaries revealed.

Wailin: (00:23:02) As your organization has grown, are managers now trained in having these conversations as well? Is the work of doing this, is it a little bit more diffused across like a leadership level?

Jennifer: (00:23:15) We actually have a process for reviewing levels. The individual employees look at our rubrics and evaluate themselves. Then they think about where they want to be headed and what they think they want to be working on. And then they bring that to their manager. Then they say, hey, this is where I think I am. And this is what I think I need to be working on. That sets it up so that the employee is the one driving the conversation, they’re being proactive, and they’re showing what their desires are for the mentor then to react to and say, hey, I’m aligned with you on these five points. But these two points, here’s my different perspective.

(00:23:56) So one of the things that’s really valuable about that is that when the manager comes back with this different perspective, it doesn’t come across as I’m judging you. It doesn’t come across as here, let me tell you your 10 flaws. It comes across as we’re working together to advance your career. And then let’s come up with a plan together. That’s the output of that conversation.

(00:24:21) We hired in a head of people ops this past year, a bit late to be honest, we probably should have done it a year or two earlier. We’d just been flying by the seat of our pants with Google Doc templates and whatnot prior to that. And what I’ve been seeing is generally a huge sense of relief. You know, a lack of clarity causes stress. And so the clearer and more explicit we can make all of this, the actually less cognitive load it is for the employees. And people have been really happy to see us work on this.

(00:24:53) Broken By Design by Clip Art plays.

Wailin: (00:24:56) REWORK is produced by Shaun Hildner and me, Wailin Wong. Music for the show is by Clip Art.

Shaun: (00:25:00) Truss is at Truss.works, that’s T-R-U-S-S dot works. Everett Harper is on Twitter at @everettharper. That’s E-V-E-R-E-T-T Harper. And Jen Leech is on Twitter at @jennifermleech. That’s spelled L-E-E-C-H.

Wailin: (00:25:16) Truss has written extensively on their company blog about why and how they did the salary transparency project. There’s more details we didn’t cover here so we’ll link to those blog posts in the show notes for this episode.

(00:25:28) You can find those show notes along with transcripts at rework.fm.