Listen to Crafted, Artium's podcast about great productsand the people who make them.

Listen and subscribe to Crafted: Apple Podcasts | Spotify | All Podcast Apps

Full transcript below — but we recommend you listen for the best experience. 

Sol Rashidi: When the business gets out of the way and lets us do some of the magic that I know we can do and they're willing to partner and experiment, it's unbelievable to see what the results are.

Dan Blumberg: That's Sol Rashidi, the chief analytics officer at Estée Lauder. Sol's been in the business of data insights since before it was cool, taking companies from merely collecting and cleaning data to leveraging it in powerful ways. She's worked with giants like Sony Music, Merck and Royal Caribbean, and she was a key player in launching IBM's Watson AI. On this episode, Sol takes us through her career and how she's gone from playing rugby on the women's national team to guiding businesses through the digital age.

Sol Rashidi: I had my hands in both strategy, use case development and understanding multiple industries and where their apprehensions were.

Dan Blumberg: But it's not only about hard numbers, we'll also talk about how important storytelling and EQ are and how what should have been one of Sol's biggest triumphs ended up being a painful lesson.

Sol Rashidi: The CEO patted me on the back and said, "This is exactly why you're here. This is why data's so important." And I felt really good. I'm like, "Yeah." Well, the problem was is I was in the same room with the three of the presidents from the line of businesses and they looked at me and they're like, "You are no longer invited to any of our meetings. You are dead to us."

Dan Blumberg: Welcome to Crafted, a show about great products and the people who make them. I'm your host, Dan Blumberg. I'm a product and engagement leader at Artium where my colleagues and I help companies build incredible products, recruit high performing teams, and help you achieve the culture of craft you need to build great software long after we're gone. Before you got into the corporate world, you were a professional athlete. And I'm curious before we get into the weeds of data and insights, how your experience playing water polo and rugby shapes how you work with teams today.

Sol Rashidi: You know what's interesting? I think in our field, the sport that I play now, which is being a professional CDO or a CAO, you definitely take a few hits and your only option is to stand back up and keep moving forward. And I think a lot of that grounding and foundation came from playing sports. Water polo is a tough game and I was never the best, but I was definitely one of the hardest working.

When I went to the collegiate level and played water polo for Cal Berkeley, I realized that division one sports is a completely different ball game, no pun intended. And the sad thing was is my coach and I had a conversation one day and I'm like, "It's my sophomore year". And she's like, "Listen, you're great. Love your attitude, love your tenacity, love everything about you, but you are just not built for this sport at this level.

You're five three on a good day, all the other girls are nearly six feet tall. You just don't have the arm reach and it's not going to happen. So you have to decide if you want to continue practicing and training knowing that." A week later I was walking down Bancroft Way and a woman named Kathy tapped me on the shoulder and said, "What sport do you play?" She's like, "Have you ever played rugby?"

I'm like, "What on earth is that?" She's like, "Well, the Americans call it football without pads, but it's a lot more graceful than that." She's like, "You've got the perfect physique for it. Why don't you try out? We're building a team." So I did. And the biggest lesson there was all of my inefficiencies and deficits in water polo were my strengths in rugby. I was five foot three, but that means I was shorter, had a lower center of gravity, harder to tackle, huge benefit.

I had all my ball handling skills from water polo because you manage a ball while treading water. And so I became captain within a matter of four or five months, and I played on the team, played on the national team for a few and it carried over. So I always say from there, two things. One, resiliency, work ethic definitely go hand in hand. And two, know your shelf life. It may not work in one thing, but you may prosper and thrive in another. So it's not always you, sometimes it's the environment you're in.

Dan Blumberg: So tell us then how you went from chemistry to engineering to data. What's the journey there as you figured out where you'd have maximum shelf life, as you just put it?

Sol Rashidi: Honestly, I couldn't have planned it this way. It's a little bit nuts, but I realized when I graduated, I wanted nothing to do with chemistry. I was more of a front office person, not a back office person. I was more of an extrovert, not an introvert. I thought maybe project management was a thing. And so I did it for a few years and I realized I wanted to be the decision maker, not manage the decision makers. And a lot of times I noticed that I could see something a few steps ahead from what other people could see. And so I just took some random jobs.

Dan Blumberg: Sol moved from project management to business development at an agency in Los Angeles. Her company was up against Accenture for a big ERP deployment deal, and suddenly they needed a data expert.

Sol Rashidi: So they shipped me off to Texas of all places, to Dallas, and my job was to get MDM certified. MDM had just become a thing in the early two thousands, and it was a major issue for all the ERP deployments because it wasn't the business processes that were struggling, it was the data. And master data management was a big, big topic. So they shipped me off to Dallas.

I was in charge of business development, but I happened to be the most technical person on the entire team, and I eventually became one of 11 that was certified. And so we came back, we said, "Hey, find anyone on the Accenture team that has a certification, Sol has it." And they're like, "No one does." And that's how we got the data deal. As soon as I put it on my resume and on LinkedIn, IBM, Deloitte, KPMG, Accenture, everyone started coming to recruit me. I ended up choosing IBM, which was the best choice I ever could have made. It fundamentally changed the trajectory.

Dan Blumberg: Sol went from team lead to managing the build of IBM's enterprise data management capabilities. And around that time, IBM was building Watson AI, which would soon beat the Jeopardy champion, Ken Jennings.

Sol Rashidi: When Watson beat Ken Jennings in 2011, IBM said, "We're ready to take Watson to market." Fortunately by then I had built a lot of visibility with a lot of key leaders within IBM because I was quote unquote "Their data girl." I was the one that could make data not boring. To quote Justin Timberlake, "I brought sexy back," if you will.

And so I just traveled the world and my job was to articulate the value, the need, the necessity, and how everything starts at this core and fundamental track. And so when they were ready to release Watson, I said, "Listen, AI is not possible or feasible without data. You need someone on the team that knows this area." They believed in me and they brought me over, and that became the game changer for me.

Dan Blumberg: One of the things you did at IBM was to establish a point of view on how enterprises could leverage AI for competitive advantage. Could you share more on what that meant in the early days of Watson and how large enterprises today should be approaching and utilizing AI?

Sol Rashidi: IBM was amazing when Watson first launched. 2011, 12, 13, I mean, talk about an opportunity. Literally Dan, my job was to fly around the world, Mexico City, Paris, Prague, meet with enterprises that wanted to be early adopters in this space, help them understand what AI was, help develop use cases that would be pertinent to certain functions in commercial banking and healthcare.

And so I had my hands in both strategy use, case development and understanding multiple industries and where their apprehensions were, but I also had the opportunity of working with the product team and helping them develop the product and the platform and features and functionalities in the roadmap. And so it was a beautiful amalgamation of both, and those themes have always carried forward.

I was one of the first that actually brought in a product mindset into the role itself that was no longer around running ETL scripts and doing migrations and integrations and quality and cleansing. That's what I call sort of the defensive playbook. I really brought the offensive playbook, which was around capability development, product development. And so that was a lot of fun and that came from IBM.

Dan Blumberg: Can you share more about what you mean when you say you're going on offense instead of on defensive measures with regard to data and analytics in AI?

Sol Rashidi: Yeah, so about the back office, the defensive playbook is always around infrastructure, cloud ops, InfoSec, migrations.

Dan Blumberg: Cost cutting.

Sol Rashidi: ETLs. Yeah, it's really the piping and the plumbing. And that was really the center of gravity when people talked about data. It wasn't really talked about an enabler or a value generator. And so when I had my first opportunity, I had a phenomenal leadership team that really believed in me.

The goal was not to do data for the sake of data. Aggregating, consolidating it and dumping into a data lake back in 2016 was not the value. What we were going to do with it was the value. And so I would always create an offensive playbook of here's what we can't do today, and here's what we're going to be able to do. Here are the personas that we've ignored historically, here are the personas we're going to focus on.

It's we are creating capabilities for the front end users of this data and we're going to not only build for them after understanding their pain points, but we're going to understand what's going to create productivity, efficiency and also go deeper when it comes to market penetration, new channel strategies, new consumer cohorts, understanding consumer journeys so that they can better market. Right time, right place, right?

That's not another pivot table, it's not columns and rows, but it's the fact that you could fundamentally increase consumer retention, consumer attention, consumer experience. And that's the fun stuff. So sometimes when the business gets out of the way and lets us do some of the magic that I know we can do and they're willing to partner and experiment, it's unbelievable to see what the results are. And so we started becoming business partners and not necessarily looked as the tech nerds. And I think that's where the pivot came.

Dan Blumberg: You were Chief Digital Officer at Sony from 2018 to 2020, and on your LinkedIn page you say you helped the company embrace a data analytics mindset in order to connect with fans, amplify current operations and not be threatened by the digital mindset. Can you share more about that mindset and why it might be threatening to some?

Sol Rashidi: The music industry, just much like media, very rooted in experience, intuition. And once you're in that industry, you tend to always stay in that industry. When Napster came about, that fundamentally shook up the music industry. And so there's always been an apprehension, but I think now after streaming services with Apple and Spotify became the way to listen music, they crossed that chasm.

However, when it came to decisions, they rely on intuition and experience a lot more so than data and digital capabilities. What I'd mentioned to them, I said, "Data doesn't replace common sense and data doesn't replace critical thinking, but you have to be informed. Because when you are in a bidding war with an artist outside of vibe and energy and the relationship, which by the way someone else may have as well, what else do we have to lean on?

Because with the younger artists, they want to know that we are maniacally following their success and that we are going to be very transparent with their success. What tracks worked, what tracks didn't, how did our marketing work, where it didn't. If we can project forward in saying, 'We are constantly informed of how you're performing globally and locally, and we have the following data points that we'll be able to show you in near real time,' that's going to catch their attention. Sometimes having data points helps that."

And I think that really resonated. And so they said, "Okay, we'll try a few things." And I think that kind of opened the door for opportunity. But there are a lot of legacy based industries, legal, insurance, media, that are still very much dependent on executives who've been in the industry a really long time and depend on their experience and intuition versus more of a balanced approach in being data oriented.

Dan Blumberg: What are some of the biggest challenges and opportunities to better utilize data, analytics, AI at companies like Estée Lauder, at companies that are investing in better, stronger data analytics and AI capabilities?

Sol Rashidi: There's a few themes. One, companies naturally gravitate towards a consumer oriented enhanced experience when leveraging these capabilities. But I think one of the things that I've always advised the companies that I've served and advise companies that I don't serve but pull me in for advisement is don't start with something that's consumer facing because that's your primary revenue stream.

If you don't have the muscle, if you don't have the talent, if you haven't built this capability yet and you're just dipping your toes into it, let's go with a high value, low risk area like an internal function, for example, before we go to something that's more consumer facing. I would say that's one. I think the second theme is around productivity, efficacy, and lowering the total cost of ownership.

Within a function, and whether it's supply chain, whether it's R and D, whether it's finance, you'd be amazed at how many people are still operating on manual entry and Excel spreadsheets and everything sits on their local drive. So even something as simple as automating a task that is menial repetitive is a no-brainer, and it's unbelievable the magnifying productivity that it provides back to the workforce.

So imagine infusing your teams with more time so that they could refocus their energy and effort on some of the newer things that are coming down the pipeline. When you spin the narrative of it's not about replacing your job, but you actually get to learn a new skillset, you could be relevant in this new workforce that's going to be needed in the next few years.

And we're prepping you for that. We're up-skilling you now, but we can only do it if we remove the 20 hours of manual labor that you do on Excel and PowerPoints. And folks always forget that. They assume it's replacing people versus enhancing folks. And I'm like, "No, this is an accelerator for you. This is an amplifier for you. Just be open to it."

Dan Blumberg: How do you approach questions of automation? When should a human be in the loop and when is it safe for marketing automation tool or AI to act on its own?

Sol Rashidi: That is an excellent question. I am a big believer that it should not act on its own. It may facilitate a task and it may provide an output, but it is your responsibility a thousand percent to take that output and apply your experience, your knowledge, your education, your critical thinking, and decide if that output is relevant for the situation at hand.

There was recently a publication where a lawyer had cited, and he used ChatGPT, had cited four cases as an argument to win his case, but then afterwards they found out that the four cases he had cited were completely made up by ChatGPT. They were not actual cases. And so this was a classic example where you've got a lawyer whose fundamental job is forensics and due diligence and did not do forensics and due diligence in the four cases that he had cited that were complete hallucination cases that ChatGPT had come up with.

So I don't ever think AI should act on its own. I think it's in its, it's not infancy, it's more mature than infancy, but let's say on a gradient of one through 10 in [inaudible 00:15:24] skillsets. We're at a four-ish, we're not eight, nine, robots aren't taking over the world and making decisions for humans just yet. But because of that maturity level, it still requires us to validate and understand if contextually it's relevant for the situation at hand.

Dan Blumberg: So that lawyer used ChatGPT, I definitely read about that case, right? It cited four what sounded like totally reasonable [inaudible 00:15:46]

Sol Rashidi: Totally.

Dan Blumberg: ... case studies. Robbers versus the state of Kentucky or whatever it was. And it's like, "Oh, that sounds like a real case" and he didn't look it up, which he should have. But I'm curious, what do you think product makers who are building technology like GPT should be doing to help?

And I generally don't like blaming the user when something goes wrong, usually I think, "Oh, the product maker designer could have done something to help this person." So I'm curious how you, as these technologies are rolled out, what are ways that the context of what the software is doing can be framed better so that people don't misunderstand that, "This is perfect and I can just go run with it."

Sol Rashidi: Yeah. And this is where I think there's macro and micro and local implications. At a macro level, anyone who's in the industry or in the know knows this is not perfect by any means, and it can't be relied on a hundred percent. But we've been working with it for a really, really long time. And what OpenAI did with ChatGPT, which was different from any of the other AI companies was they went direct to consumer and it's the first AI offering that's direct to consumer channel.

So their channel strategy was fundamentally very unique and that's what blew this space up. That aspect fundamentally created new challenges, but the company was very open, aware, honest and transparent about its hallucinations and about its limitations, but it did make it available for the average consumer to use. So at a macro level, I think the company did its responsibility.

At a micro level though, it's up to the individual. And I think that breaks out into two aspects. If your function or job has forensics and due diligence associated with it, like a lawyer, you better damn well be responsible when you're using ChatGPT and do your research. So in that particular case, I actually do blame the lawyer. I'm like, it is your job to research and you did not research four court cases that you were spitting out, shame on you.

So I think at an individual level, if you're going to use something like ChatGPT or any of the software out there, you still have to apply the common sense and the critical thinking and your experience and ask the questions to make sure that what you're seeing in the output that you're reading aligns. But folks just aren't thinking like that these days. And I think the part that I'm most worried about is intellectual atrophy.

We have a lot of devices and apps and tools available to us that has decreased the mental workload. And I talked to some individuals, and I can just tell that intellectual atrophy is kicking in. We're depending too much on the apps to do things. So at a macro level, it is up to the company to describe what it can and can't do. Then there's at the local level, which is as an individual, if my role and function has an element of forensics into it, don't skip it, apply it. Don't take it for granted.

Dan Blumberg: I want to get away from the hard numbers here and talk about EQ for a second because you've shared with us how important it is to not only unearth great insights using data, but you said you're a translator and it's important to translate those insights in a way that actually affects change, doesn't just scare people. I wonder if you could share more on that.

Sol Rashidi: The evolution of my role has changed year over year. There's a harsh reality. There's a lot of expectation without a ton of empowerment and control, and we have to split the decision making amongst several leaders, and so it requires a lot of alignment and you're oftentimes not keeping pace with where the market is going and et cetera, et cetera. I can give you a hundred different reasons.

But I think if there is a chance of success and if you do want to increase your probability of success, I do think that there is a formula that you can't skip, and it's a combination of IQ, EQ, and DQ. Your intellectual quotient is all about your competence and knowledge in the space. And that can be learned, that can absolutely be learned. The digital quotient can also be learned. Now, it does require a different skillset.

It is connecting dots to understand new opportunities, new capabilities. That can also be learned. EQ is a tough one. The emotional quotient is you've got to understand the industry, like where I was with at Sony Music and knowing fundamentally going in that this was going to be an uphill battle because it's not an industry that's necessarily leaned into technology to run their business and operations.

And that's something that I think not everyone carries forward with them, and I think that's a very important piece of the puzzle to make your role as successful as possible in the data and applied AI space. It's also where it's gotten me in trouble. It was probably one of my most epic failures, but probably my greatest success. It's a story that gave me a good kick in the pants. Earlier in my career the goal was to consolidate all data sources that had consumer data in it, and it was about 76 different databases.

We defined that 21 of the 76 were really pertinent for us and important, so we aggregated into a data lake and we're like, "Okay, it's ready for consumption." Our goal was to generate new insights now that we have all of the consumer data aggregated. It took us six weeks, but we discovered something. With our loyalty program we had three tiers. And our highest tier program, we bent over backwards for them because we thought that we were making so much money because they were purchasing just a lot of goods from us.

However, when you combine data from our revenue management system, a few things we noticed. One, none of those call center systems and applications were integrated. So if you had someone from that highest loyalty tier call Wichita and complained about something and they got comped, they could call Tacoma, complain about something and they would get comped. Then they can call Jacksonville, complain about something and get comped.

Well, when you combine all the products and services that we had comped combined with the revenue that we were generating from these most revered loyalty members, we were actually losing money, not making money, and no one in the company had known that. This is huge, and I shared it with my boss and he says, "Wow, Sol, this is massive." Shared in the next EC meeting. So I went to the executive committee meeting, I shared the findings.

My boss said I had done a great job. The CEO patted me on the back and said, "This is exactly why you're here. This is why data's so important." And I felt really good. Borderline cocky almost. I'm like, "Yeah." Well, the problem was is I was in the same room with the three of the presidents from the line of businesses, and they looked at me and they're like, "You are no longer invited to any of our meetings. You are dead to us."

Well, you can imagine my surprise, and I just looked at them. I was like, "What? I just saved us around 120 million dollars operationally year over year. This is fantastic insight. Why would you guys not use this?" And what had happened was, and this is where EQ kicks in.

Yes, it was great that we came up with a new insight. It was the fact that I came up with a fact that they had not. I discovered a business insight that the business had not, and I'm not one of them. What's worse is I surprised them, I sidelined them and I didn't bring them in the forensics of researching to see if this was really a fact or not, but instead there was organ rejection.

I had the opposite effect. Because it had nothing to do with me being right or saving money or coming up with this miraculous insight. It was the fact that I presented it without running it by them, and I did it without including them, and I uncovered something that they had not uncovered in 40 years of operating. And it was egg on their face. And it took me 14 months to recover essentially. Sometimes it's not about being right.

Dan Blumberg: This feels like this goes back to the team sports stuff we started talking about at the beginning.

Sol Rashidi: Oh, yeah. Oh, yeah.

Dan Blumberg: Sol, thank you so much. I really appreciate you taking the time. We've learned a ton today.

Sol Rashidi: Oh, it was my pleasure. Thank you guys so much. This was incredible.

Dan Blumberg: That's Sol, and this is Crafted from Artium. If you're looking to leverage data and AI in new ways, let's talk. We're helping a number of companies looking to advance their AI capabilities right now. At Artium, we can help you build great software, recruit high performing teams, and achieve the culture of craft you need to build great software long after we're gone. You can learn more about us at thisisartium.com and start a conversation by emailing hello@thisisartium.com. If you like today's episode, please subscribe and spread the word because you improve as a person with each and every listen.

Sol Rashidi: This is an accelerator for you. This is an amplifier for you. Just be open to it.