Patrick: Today’s conversation is with David Swanagon. David is the Head of HR and Analytics for Technical Services at Saudi Aramco. Saudi Aramco is the world’s largest energy company, and last year they completed the largest IPO in history. Dave is responsible for multiple people initiatives, including leadership and professional development, diversity and inclusiveness, talent retention and data analytics, just to name a few. David, you sound like you’re quite busy, so thank you for spending some time with us today and welcome to the podcast.
David: Thank you, Patrick. Appreciate the time and I hope everybody is staying safe over in the UK right now during COVID.
Patrick: Perfect. Let’s jump right into the conversation. David, could you just give us a quick overview of what is people analytics?
David: Sure things. There’s really two components of people analytics. There’s the technical aspect in terms of how to design models that lead to executive decisions, and then there’s how measurements and analytics fit within the overall improvement method that a company utilizes to optimize these operations. To me, people analytics is understanding the technical components that allow you to create strong statistical models that has the rigor from a data perspective, but more importantly is to understand how those models fit within a portfolio of analytics, and how that helps influence an overall improvement method that a company utilizes to implement change ideas and to continue optimizing their results.
Patrick: I have to say, David, that is probably the best explanation of people analytics in a way that is extremely tangible, easy to understand. How have you used or utilized analytics and data to help Saudi Aramco with successful implementation of decisions or people programs?
David: Sure, absolutely. I think this is definitely a compliment to the company. One of the nice things about working in Saudi Aramco is all of the leadership and your colleagues are engineers, so there everyone loves mathematics. Typically the challenges you would have at other organizations of getting people comfortable with statistics is not really a challenge within Saudi Aramco because of the technical expertise that exists within the workforce. The issue that we have, though, is really managing the portfolio of analytical models and how the government should be managed in terms of making decisions.
For example, we’ve really developed and it’s over an evolutionary period, we originally started with more of a descriptive statistics methods where it’s just creating KPIs that tell us, “Here’s how we did in terms of talent acquisition, talent development, organization effectiveness, and talent retention.” All of the activities that fit within those pillars, we would develop metrics that said, “How did we do last quarter, last year?” It really didn’t tell us that much, especially if you were to enter way that with employee engagement surveys. If you’re looking at how employees felt at a particular time last year, it’s really hard to make any change ideas or identify the key drivers because the business environment is always evolving.
What we’ve done over time is shift to more predictive model, which means collecting a ton of data around employee sentiment, employee time and motion, also around the traditional [unintelligible 00:04:28] components, which is their academic background, their certifications, and so forth. Then also pulling in a lot of external data sets that tells us what are the demands for the jobs, what are the challenges for the company’s corporate objectives, and so forth.
Then from that, it’s really moving into this machine learning perspective, where we try to pinpoint how does HR enable the business from a strategy perspective? I think this is really the first step in people analytics, is that you have to look at the overall corporate objectives of the company. For Saudi Aramco, their major objective is to become the world’s leading integrated energy company. Part of that includes initiatives for environmental protection, moving into non metallics, and so forth.
The first step in people analytics is trying to understand, what are the enablers within the HR space that will help Saudi Aramco be successful in achieving their corporate vision? Then once you understand what those enablers are, for example, hiring the right people, developing engineers in a cost-effective and timely manner, making sure that employees are trained on the right domain disciplines so that they can respond to changes in the marketplace quickly, and then also retaining your top performers.
All of these things for us is really the basis of the models we develop, and then what we use is machine learning techniques to identify what are the key drivers influencing a good hire or somebody who may resign but hasn’t yet, or an engineer, what is the optimal training hours for them, and so forth. For us, it really starts with what are the enablers to the corporate strategy, and then from that, we focus on what are the key drivers of those enablers so that we can design our analytical models to support decision making.
Patrick: There are two points in that, that really stood out to me. The first is this concept about HR being a strategic partner for the business. That is absolutely something that we are seeing as we all move into this new world of work, and we’re looking at the future of work, and how HR specifically has been elevated in so many different ways over the past couple of months, and how HR is really being looked at as a true business partner as opposed to maybe in the past a supporting business function, but really now being looked at as a leading business function?
You also mentioned that the pace of change at organizations is significantly fast. We’ve been seeing that pace increase over the past couple of years, but over the past six months that pace has changed significantly. What have you learned at Saudi Aramco about your people over the past six months? Obviously, the world has been an interesting place. Have you learned anything about your people, specifically in the past six months that’s helping you with your predictive side of your equations?
David: Absolutely, It’s a great question. Really, if I can extend this to maybe the last year, there are many critical events happened within Saudi Aramco that I think really displayed the excellence of the workforce, but also gave us new insights into it, the best out of our performance. Last year, in addition to doing the world’s largest IPO, we also responded to missile strikes in our Abqaiq plant, and these missile strikes actually shut down 5% of the world’s global oil supply.
It required because of the logistics of the situation that our young engineers had to respond to the field site. They had to diagnose the issue, they had to come up with very novel solutions because a lot of this equipment requires long lead time. Based on the way these individuals were managed by our senior leaders, and empowering them and giving them a key objective with a key challenge, but not micromanaging the solution, our engineers were actually able to really solve the missile strike in a timely manner, which actually fed into the IPO being successful in subsequent quarters.
I think what makes Aramco so special is that over the last year period where they’re challenged by the missile strikes in the IPO, that adversity actually built an operational muscle that helped Aramco be successful when the pandemic started. I think that it’s funny if you had told us last year that those particular challenges were actually going to lay the groundwork for the company being successful once the pandemic started, I think everyone would laugh because they were such difficult complex challenges in their own right.
What we found is once the pandemic happened, is that the same leadership style that got the best out of the employees for the IPO, as well as the Abqaiq attacks, is the same leadership style that allowed the company to really optimize its operational model when it shifted to more of a remote site. It had to deal with significant changes to safety policies related to how employees interact, both in the workplace as well as on the camp.
In addition to that, what we found is that if you allow the freedom with employees, that innovation, it’s interesting as you can manage almost everything of employee except innovation, the innovation is not a hierarchical function. Executives really have to be comfortable with allowing all types of employees to network with each other, to come up with ideas and with novel solutions to challenges, and then have a filtering mechanism that allows those ideas to be brought to the key decision makers.
I think from a people analytics side, what we were able to achieve is we had a lot of data capture in terms of what solutions were going to possibly influence the way Aramco responds to the lockdown. Again, I think it goes back to the original thing we were talking about. Companies need to have a really strong understanding of what their corporate strategy is. When you face any type of adversity, especially with how dynamic geopolitics or economic markets change, if you don’t have a clear understanding of what your corporate strategy is, it’s easy for you to get distracted as adversity happens.
If you know exactly what your company stands for, exactly how you make money, and you also know what your value proposition is to employees, you really have those three things locked down in a clear way, and it’s embedded into your culture. Then as situations arise that require significant changes to the business model you anchor your policies against those values. Where analytics for us were really successful is we knew that Aramco still wanted to achieve certain objectives around its productive capacity for oil and gas.
We knew we wanted to keep employees safe, and we knew that we wanted to allow innovation to flourish in the company so that any employee could come up with a good idea that we would implement. From that perspective, our job within analytics was to capture all the data points happening within the company that would help executives deploy solutions that fit those three things. What’s our strategy, how do we keep employees safe? Then how do we continue to drive innovation in a more crowdsourced, egalitarian fashion?
Patrick: You’ve mentioned something that is an extreme passion of mine as well, which is this concept of innovation and empowering every employee at an organization to feel that they can innovate. That they’re part of the innovation process at their organization. One of the important things as new, the newer generations are coming into the workforce is that they’re expecting organizations to share some value either with their personal values, that corporate values match personal values, that they’re able to be in an environment where they can think creatively and innovate.
You’ve mentioned innovation a couple of times, and that really struck a chord with me. Could you give me one example of, you have these three strategic initiatives moving forward that you just mentioned, one of them being to continue this concept of innovation? Is there one specific way that you’re thinking about implementing a program or a process to encourage that and to allow that?
David: Oh, sure, absolutely. Again, I think what makes Aramco successful and really one of the best practices we could apply at other organizations is you need to ensure that the analytics is part of a overall improvement method and a governance framework that allows change ideas to be properly tested. Really what Saudi Aramco tries to do, one initiative they created is a digital hackathon, for example.
The idea of this digital hackathon is that we want to develop preliminary use cases for how to employ analytics and digitalization within various parts of the business. Obviously, this hackathon allows through its competition and its PR and its marketing, any idea to be considered on its merits, without it being tied to leadership hierarchy [inaudible 00:14:02]. What makes Aramco special is how they filter those ideas.
Since we incorporate analytics into an overall improvement method, what we try to do is that if a use case is approved from a pilot perspective, we tie it into what anchor analytics can we do to understand what the current baseline is, as well as how would these proposed change ideas influence the business and not simply in an individual way, we have to also tie this change idea to the overall portfolio of the company’s improvement methods. What we’ve found is really interesting is that a good idea actually can have a detrimental impact if it’s not properly combined into the portfolio of change happening in the company.
It’s a lot like a stock portfolio. It’s you want to diversify, you want to make sure that something you do in one part of the business does not have adverse effects in other parts of the business. We use these competitions to really inspire all employees to develop the use case ideas. Then once we have this collection of ideas, they’re put into a structured improvement method that ensures we’re collecting baseline analytics, we’re developing machine models to test them from a piloting perspective. Then we’re also developing driver diagrams that help us understand if this change works, does it have positive or negative effects on other parts of the business?
What we found is that once employees understand that their idea is part of a bigger system of ideas and that we want to make sure the whole company improves, not simply the sector they’re focusing on, their innovation becomes far more complex and becomes far more rigorous in terms of the ideas they submit, because they understand we’re not just trying to improve employee engagement in XYZ department.
We’re actually trying to ensure that employee engagement in this department leads to a standardized practice that helps employees across the company. When you apply that rigorous standard, employees actually come up with ideas that are far more substantial than if you said, “We’re just looking for great ideas and we’re going to allow any idea to be implemented, and we’re not going to track it on a portfolio basis or part of a broader improvement method.”
Patrick: That’s incredibly impressive. I have to say, I mean, I could talk to you for hours about this because we share a similar passion around analytics. [laughs] I just have to say, it’s very impressive to hear the work that you’re doing. It’s impressive to hear the work that Saudi Aramco as an organization is doing. As we wrap up, you obviously have quite a bit of experience in this field. You’ve been at Saudi Aramco for multiple years, have a long tenure in this industry and the business. What would you suggest to other organizations, a reason why they should consider building out their people analytics function?
David: That’s a fantastic question. Now, the caveat I would stay regarding this is that analytics is a double-edged sword, is that if you’re going to go down the path of building out an analytics function, then the company needs to be prepared to really invest in an employee’s statistical understanding. In order for a model to be rigorous from a technical perspective, we need to ensure that the data collection methods are done correctly. That the design parameters apply best practices. That actually the choice of the machine learning methodology matches the use case. Then also the overall findings don’t have structural issues such as multicollinearity and so forth.
A lot of organizations I think are jumping on the analytics bandwagon without really putting in place the governance or the technical abilities to ensure that the findings presented to management make sense and are structurally sound. For me, I think progress around data sciences is important, but it has to be followed with the right investment in the technical capabilities of the business. That also includes ensuring that executives are not only trained but are prepared to spend the proper amount of time vetting the models that are presented to them.
One example I’ll give you about how important this is that, for example, if you develop a machine learning model that relies on multilinear regression, multi variant regression, for example, well, that model is going to pull historical data, and then it’s going to utilize its mathematics to regress against what are the variables that determine XYZ event. For example, if you were to do a machine learning model that looked at leadership selection, and it tried to predict who would be great leaders in the future, if you choose a model that relies on historical data, you have to make sure that that historical data doesn’t have unconscious bias, for example.
When you look at stuff such as diversity and inclusiveness, if your goal as a company is to drive female leadership selection, then you have to be very cautious if you’re going to use some machine learning model that utilizes historical data where females have been disadvantaged, for example. There’s all these structural changes that have to happen to ensure that the model is correct, it’s aligned with strategy and that can’t happen unless the executives have really bought into the technical aspects of analytics.
That aside, if you have an organization where the executives believe in analytics and they’ve invested in the technical capabilities, what this does is that the more you understand a problem, it actually allows you to investigate parts of the system you didn’t even know exists. I think this is really the great finding is that whenever I’ve worked on an analytical model as we go through the process of solving that particular issue, it’s helped us expose other problems we weren’t even aware, other lines of inquiry that we could drive further and focus more on.
Analytics can do that because the systems of companies are so complex. Employees are so unique and interesting that if you just rely on your subjective opinion, you often summarize the experiences you’re having with them, but if you use analytics and you use it in a robust way you are able to see that, “Wow, employee engagement changes based on the environment that you’re in. That a leader can have an impact in certain areas, but not an impact in other areas and that engagement is important when it’s combined with safety, for example. It may not be as important if you’re dealing with cost perspectives.”
There’s all these interesting combinations that happen with analytics that if you’re just relying on your subjective experience, you either don’t spend the time focusing on it or your opinion isn’t powerful enough to sway stakeholders. In my opinion, I think as long as you develop a robust analytical framework that includes technical capabilities you’ll be able to explore all kinds of problems. Then the findings that you generate typically will influence executives enough that they champion change ideas in the business.
Patrick: David that’s excellent. Your insights are incredibly helpful and beyond helpful, quite inspirational. I also think that our audience is going to be very inspired around the idea and the impact that a true analytics program can have at an organization or incredibly passionate about data and analytics as well. We share that passion with you, but I just wanted to first off say, thank you for the insights. As I said I think you have articulated people analytics in one of the most concise and usable ways that I’ve ever heard. Thank you for that.
David: [laughter] Thanks my friend I appreciate it.
Patrick: Honestly, also just thank you for spending some time with us today and for joining the podcast. We appreciate it.
David: Much appreciate it and best of luck to everyone. Please stay safe like I said and thank you very much.