Jeff Dirks and Carter McHugh

[00:00:14] Jeff Dirks: Hi, I'm Jeff Dirks. I am the Chief Information and Technology Officer for TrueBlue.

[00:00:21] Carter McHugh: Hi, and I'm Carter McHugh. I'm the VP of Technology for PeopleScout, a TrueBlue company. Very excited to be here today with you, Jeff, and addressing the DevOps Enterprise Summit for 2021. This is a very exciting time for me. I've been an Enterprise DevOps Summit attendee since 2016, so this is my fifth year and my first presenting. So, very excited to be here and can't wait to talk a little bit about the transitions that TrueBlue has made over the past few years as we've embraced a lot of DevOps and Lean concepts.

[00:01:03] Jeff Dirks: All right. That's great. Let's first introduce TrueBlue. And we are a $2 billion revenue company that spans the staffing continuum in North America. We put nearly 400,000 people to work per year. We do about 30,000 payrolls a week, and we're a force for good, making it easier and simplified for people to connect with work, whether it's day labor, recruiting process outsourcing, or staffing pick, pack, and ship operations for key clients across the United States.

[00:01:43] Jeff Dirks: In terms of our technology stack, we have three primary products that span our three business lines. JobStack is our iOS and Android native mobile applications for associates who are looking for day labor and clients who are looking for laborers to do work. And the focus of JobStack is to simplify how we match the need for work with folks who want to get work done. Our StaffTrack application for people management helps to operationalize and manage workers across various different job sites in terms of staffing, plant, pick, pack, and ship operations. And our Affinix application automates the recruiting process outsourcing services that we provide for our clients here in the United States and abroad.

[00:02:34] Carter McHugh: Thanks, Jeff. That's a great introduction to TrueBlue, and TrueBlue's really a great company, a force for good because we connect people with work. And if we go back in time a little bit, to 2018, which is the first year, Jeff, that you started in your role as CIO and CTO for TrueBlue. And at the time, there were a lot of challenges we faced. Talk to us a little bit about some of the challenges that you came into when you entered the organization.

[00:03:02] Jeff Dirks: Yeah. From a challenges standpoint, in terms of technology supporting the business, it was very much an old-school IT service management, ITIL focus, very siloed, very disconnected with thinking about what our respective businesses needed or were going to need in terms of anticipation. And I really call this the flywheel of decay, that our revenue growth was beginning to slow. We had margin erosion, which of course drives cost-cutting pressures, which leads to more people having to get more work done across more systems. And that really put us on our heels versus leaning in and thinking about where the business was going. We were caught in this flywheel of decay, which is really stagnation.

[00:03:50] Jeff Dirks: As the business was trying to accelerate and anticipate how we could simplify how people connect with work, creating more client value, generating more growth and value as a result for our clients. But we were bunkered as a technology organization. We were focused on cost cutting, focused on controls, oversight. We weren't nimble. We had a significant amount of investment out of our total IT resources. About 70% was spent on back-office KTLO, and we were understaffed trying to do too many things with aging systems. And we were very much disconnected from what the business needed out of technology to be competitive in the North American staffing market.

[00:04:35] Jeff Dirks: And so what did that mean? We were in the danger zone with respect to being disintermediated by technology innovation in what has otherwise been a fairly old-school industry, with laggard technology adoption. It was beginning to transform and be completely optimized through technology innovation.

[00:04:57] Carter McHugh: Thanks, Jeff. That's a great introduction to a lot of the challenges that we had when you entered into the business. And I entered into the business not long after you did, and that was kind of at the end of 2019. And you had a big vision for where we wanted to start and how we would go about transforming our technology organizations to really better serve the business.

[00:05:21] Carter McHugh: And we started with these things with a lot of the great concepts that existed in modern technology organizations and modern software delivery cycles. We had the vision for two-pizza teams. We had a vision for adopting public cloud and getting out of our data center infrastructure. We wanted to shift security left, right? We talked about we were very control-focused, and we really wanted to get into modern DevSecOps. We wanted our two-pizza teams to run what they build and stop utilizing so many shared services capabilities that were not serving the business very well. And we wanted to adopt continuous delivery. We wanted to deliver more value more quickly to the business. And when we started this, there was great excitement around this from a lot of the teams to do this and reorganize around these concepts.

[00:06:13] Carter McHugh: And so we got to work right away. We reorganized into two-pizza teams. We migrated a lot of our infrastructure away from our proprietary data centers and into public cloud. We bought tools to help shift security left and integrate that into the software delivery lifecycle. We bought more tools to help us run what we build. We bought even more tools to help us continuously integrate and deliver our software.

[00:06:42] Carter McHugh: But the questions really came pretty quickly after that from the teams. Despite all of our investments, we really wondered: are we better? And what are we really better at? Do we better understand the business' needs? Are we more efficiently delivering greater value to our business partners? Do our business partners really understand and value technology and what it can do for their problems and what solutions it can offer? We summed it up and really had to ask ourselves, what do we know about ourselves? Are we capable of delivering more of what matters for our business?

[00:07:27] Carter McHugh: And so as we entered 2020, we had a lot of opinions, but not many facts. And most of these opinions were from a lot of leading technology companies, the Facebooks, the Amazons, the Googles. They all had great engineering practices and great technology and great opinions about how to implement those practices. But we needed more purpose and guidance beyond, \"that's what they do, so we should do it as well.\"

[00:08:00] Carter McHugh: And very fortunately, we were gifted the book of \"Accelerate\" by Dr. Nicole Forsgren and Jez Humble. It really gifted us the data-driven insights into how high-performance organizations of all kinds perform. And so we set out that winter as a leadership team. We all dove in and made sure that we read \"Accelerate\" cover to cover and really understood what it was saying.

[00:08:27] Carter McHugh: And when we returned in February of that year, we had a leadership meeting all armed with facts, and we understood what performance levers we could measure against: lead time, our deployment frequencies, our mean time to restore our services, and change fail percentages when we make changes. So this was great. This gave us a great start on what we needed to measure and how we could set some organizational goals around these areas.

[00:08:57] Carter McHugh: And we made some decisions really quickly, and one of the first ones was that every team would baseline their current metrics, and they would measure these things quarterly. And it was important that we set this for every team. We didn't let any teams opt out. We didn't let teams say, \"I have too much legacy systems to take this on.\" We didn't let teams say, \"Well, I'm just a corporate shared-services team, so this doesn't apply to me. This only applies to business applications.\" We knew it was important for all the teams to adopt these things so that we could get improved performance as a high-performance organization across every part of our business.

[00:09:39] Carter McHugh: And a second important decision that we made was that we were not going to set any universal improvement targets, such as everyone must increase by 20% or 100% or 1,000, move from four releases to a release every day. Instead, we let teams make their own commitments around these measures. And we learned some things. We learned that some teams were very greenfield. They didn't have a lot of legacy burden. They might've been a new application or a new platform that was being developed. And so they set some very aggressive goals, sometimes even around deploying every day or every user story that was coded and reviewed and tested.

[00:10:19] Carter McHugh: And some teams were really more burdened with running our legacy systems. And there was a lot of complexity involved in those and a lot of business demands to keep those tools operational and stabilized. And so they had a little bit more modest improvement goals, a little bit less room to work right away. But nonetheless, they had goals, and they set out to improve and try and reach those goals.

[00:10:46] Carter McHugh: And the last thing that we determined is that visibility was required. Teams were required to come and publish their metrics each quarter, and then they had to share their learnings about those metrics quarter over quarter. And this really helped us encourage a culture of transparency and learning at TrueBlue. So before, as Jeff opened, it was very siloed and no teams talked to each other. Business units didn't talk to each other very well. This helped encourage a culture of transparency and learning across the organization. And by doing these three steps in this way, it invited healthy competition, but it avoided having unfortunate shaming of other teams if they felt that they were falling behind or not meeting the same level of targets for other people. And so this was really important because, again, we wanted to have that culture of transparency, but also make and allow people to have room to fail and improve and learn.

[00:11:55] Carter McHugh: And this was great, but of course, it invited even more questions. Teams came to us and said, \"Well, what constitutes a deployment? Does every team need to measure the same? Right? How can one team count every user story as a deployment and another team count just change records in the change system as a deployment?\" People had questions about, does this reward poor quality? Could people game the system and release untested and unready software into production, a requirement, and then take credit for fixing it as a deployment really rapidly? People had questions about, do we have to follow the change management process anymore? Can we just deploy whenever we want? And some people even had statements and said things like, \"Well, our customers and our partners are very happy with our current cadence today, and shouldn't we not measure for deployment frequency? Shouldn't we be measuring for business value impact? What if I just release once, but it has such business impact that that's a huge success for our partners and customers?\"

[00:13:01] Carter McHugh: And we had to admit that these were all great questions, and that our difficult truth was we didn't know the answer to these questions. And though we were armed with facts, and though we had the data from \"Accelerate\" on what high-performance organizations did, we really treated this more as a compass. Right? It wasn't a map to how teams had to do something. It was a compass. It allowed us to measure, to try different things, and then to course-correct. And so we really encouraged teams to think big. If you want to deploy every day or every hour, that's a big goal, and you should think big towards that goal. But work small, just make small daily increments towards those goals. And no matter what your platform is, legacy or greenfield, we ask teams to optimize for forward momentum. Even the smallest steps that a legacy team could take towards moving forward and increasing their deployment frequency or reducing their mean time to restore was important.

[00:14:01] Carter McHugh: And so what did the teams discover as we started to go through this over the course of 2020? One, we really discovered that the increase of the deployment frequency started to lower our change fail rate. Again, small batches of work. By decreasing the amount of work that went into each deployment reduced the number of things that could go wrong, and if something did go wrong, it allowed teams to respond more quickly because they were much more aware of the small thing that changed and be able to fix that or roll it back.

[00:14:36] Carter McHugh: And so the combination of those two things started to lower our mean time to repair and restore. We had less severity-one incidents. We had less major customer-impacting incidents because of these things. And when we did have an incident, again, it was faster and easier than ever to recover or fix because we had focused on increasing our deployment frequencies and because we had had smaller batches and lower change fail rates.

[00:15:05] Carter McHugh: And so finally, all of this together reduced the amount of time that teams had to spend focusing on those things. And so that started to free them up to work more closely with the business to deliver more business value. And so the sum of all of these things really started to shorten our lead times. And so when we look back at the four levers that we wanted to pull, defined in \"Accelerate,\" we realized that by really focusing heavy and setting and encouraging teams to increase their deployment frequency really compounded on itself and allowed us to move those other levers with much greater ability and much quicker.

[00:15:44] Carter McHugh: And when we summed it all up, when we look back at 2020 and as we look forward and how we've operated through 2021 across all of our business, we found that our total changes deployed were increased by four times in 2020, and then that allowed us to set a goal for increasing 10 times on that four-time improvement for 2021. So you can really see the flywheel starting to reverse as we start to accelerate the gains in 2021 that we had over 2020. We also measured the time we spent in our keep-the-lights-on operations, KTLO, and we had a 13% decrease in that in 2020, and we set goals and are well on the way to a 30% decrease on top of that in 2021.

[00:16:32] Carter McHugh: And we also measured our IT shared spend ratio to our line-of-business spend. So how much time are we spending in legacy platforms and shared services and data center operations compared to business applications and frontline business innovations and delivering more value for our business partners? And so we saw that we were able to shift the IT spend by 10% towards the line of business and out of costly shared service centers in 2020. And we've set another goal of an increase of an additional 6% on top of that in 2021.

[00:17:08] Carter McHugh: And then finally, we said, let's measure our employee satisfaction. Let's make sure that these improvements we make are manifesting themselves in greater satisfaction from our technology employees, and we're not increasing their overwork or burnout as we increase these changes through our system. And so this was also a net positive. We saw an increase of seven NPS points on our net promoter scores for our technology employees, and we've set another goal for another increase of 4% on top of that. And so you really see that we've been able to switch from a flywheel of decay that we were in just a few years ago into a flywheel of innovation and growth.

[00:17:52] Carter McHugh: And so now we've got the ability to have sustained delivery and empower teams, and that translates into greater revenue growth, expanded margins, which allows us to invest more of the money we make back into true strategic investments, which again, excite teams, make them feel more empowered, and they invest more in continued sustained delivery for the business. And so this has really enabled TrueBlue to move from project to product-centric organization. So we were able to sunset a lot of our brittle infrastructure and technology products. And with these new capabilities we had, we were able to reinvest a lot of these savings into new product management functions that allow us to focus more on the technology stacks, JobStack, and Affinix, and StaffTrack that we talked about at the beginning, that are delivering innovation and value for our business partners. And so we've really been able to rapidly innovate for our business through these technology, and all because of the ability of what \"Accelerate\" has shown us and how we really adopted a Lean, compass-driven, measure-try-repeat process.

[00:19:08] Carter McHugh: And so I'm going to sum it up here a little bit and talk a little bit about the journey that we were on and what you might be feeling here as you're attending this conference. I know in the first conferences that I attended, everyone else felt like they were already at the top of the mountaintop. And here I was back just starting down the road on this journey. And sometimes it felt like, well, why even bother? All of these people are so far ahead. All of these transformations, they just seemed very intimidating. The leaders in these spaces were very vocal and opinionated about things. A lot of the laggards were kind of quiet and invisible, keeping their head down, trying to learn as much as they could. And so what I really learned and took away from that is that it's very important to start your journey wherever you are. And don't be intimidated by some companies being much further ahead, because there are value in starting your journey wherever you are and taking that next step. Again, optimizing for forward momentum.

[00:20:05] Carter McHugh: And to recap a little bit some of the things that we really learned around setting goals, right? We learned that the methods, tools, and commitments, they have to be set by the teams, right? A top-down organizational mandate from Jeff or me as a vice president and a CTO can often be really difficult for teams to adopt and met with disillusionment and difficulty. But by bottoming this up, by setting the vision and the goals of what we wanted to improve, we literally let the teams select their own methods, tools, and set their own commitments, and that really increased their empowerment and their buy-in towards this.

[00:20:45] Carter McHugh: And we were also really careful in how we measured this. Again, Goodhart's law kind of says, if you set some targets, don't be surprised if those targets that you set out to measure quickly become gamed, right? People start to improve against those targets to the detriment of other goals that you may have. And so that's why we really were very thoughtful in not setting specific targets of you have to improve by 10 or 20%, but really letting those teams set their own goals, no matter how modest or how aggressive, and share those learnings back with us.

[00:21:18] Carter McHugh: And then sharing those learnings were really important because we demanded transparency of the teams. They had to report on their metrics. They had to tell us what they learned, but we were very allowing for the room to fail. Right? So when they came in and they had these learnings and we said, \"Well, that's great. We're glad that we learned these things. What can we do to get better?\" Right? \"And what can we share with other teams about our best practices?\" And so that really generated a lot to the culture that we transformed from the siloing and the bunkering and being really scared to innovate for the business or do something different into a culture that allowed people to experiment and feel safe in experimenting and deliver the results.

[00:22:08] Carter McHugh: So now I'm going to turn it back to Jeff a little bit, who's going to talk a little bit about executive buy-in with our leadership at the executive level, and also what we're looking for next for TrueBlue as we go forward.

[00:22:22] Jeff Dirks: Great, Carter. Thanks. So in terms of gaining executive buy-in, we've learned a lot over my last three years here at TrueBlue, and I think to sum it up, to build buy-in with our CEO, with our CFO, and with our brand presidents really requires us starting with the end in mind and talking with them about how they're thinking about not just where the enterprise is or where their businesses are at, but their vision for where they expect the trajectory of their businesses to take them over the next three to five years. And then working backwards from the quantification of what that bullseye is. Growing higher-margin revenues, increase in market share, technology-driven innovation, simplifying how we recruit or how people connect with work, and working backwards towards how do we engineer the technology to build those solutions so that business benefit can be realized. And then the quantification of benefit can be made tangible in terms of revenue margin, market share, et cetera. And then being relentless on going back to the metrics that describe that trajectory over time in terms of the business value realization, which is something we've worked very hard on here within TrueBlue Technology with our senior business leadership.

[00:23:35] Jeff Dirks: In terms of what's next for TrueBlue, we continue to have a big vision for how we expect people to connect with work into the 2020s. And we think technology-driven innovation will be a big part of simplifying how folks that want to work can be connected with people who need to have work done. We think there's a big opportunity in terms of the emergence of the gig economy. And overall, we'll be focused on working backwards from business opportunity and business problems to figure out how to best lever technology to deliver that benefit. Making the transition from a project focus, which is classic IT, to product focus, and that product work never ends. It's a continuous feedback loop in terms of sense and respond and how we evolve over time based on the needs of the market. And then making it easy to invest in innovation and discovery. And whether it's things like robotic process automation or data science and machine learning or conversational AIs or blockchain, making it easy from a capital-efficiency perspective to think about innovations and how we can continue to add value in the market. And lastly, Carter and I are very honored to have had the opportunity to present here today, and we'd love to connect and looking forward to meeting everybody virtually. Thanks.

[00:24:54] Carter McHugh: Thanks so much, Jeff.

[00:24:56] Jeff Dirks: You bet.