API first strategy and Mulesoft help Pilot Flying J break down data silos

Pilot Flying J uses MuleSoft, which enables the integration of data from locations across North America and increases the quality of service provided to its customers.

API first strategy and Mulesoft help Pilot Flying J break down data silos Pilot Flying J uses MuleSoft, which enables the integration of data from locations across North America and increases the quality of service provided to its customers.

At Dreamforce 2019 in San Francisco, TechRepublic's Bill Detwiler spoke with Pilot Flying J Senior Director of Digital, Loyalty, and Brand Marketing, Tyler Tanaka about interoperability and how linking multiple datasets has increased the performance of the company. The following is an edited transcript of the interview.

Bill Detwiler: How do you bring together multiple companies that are built up over multiple M&As? How do you bring together those systems in a way that's efficient, that works, that benefits the customers? I'm here with someone to talk about that at Dreamforce. Tyler Tanaka is Senior Director of Digital at Pilot Flying J. You may know them from their truck stops and gas stations, but there's a lot more that Pilot Flying J does. They manage one of the largest fleets in the industry, so I'm really interested to hear how they manage to keep everything running so smoothly.

Tell me a little bit about Pilot Flying J. How you've grown through M&A, how you've had to connect all this legacy hardware, software, data repositories, and bring it together in a way that works.

SEE: Hiring kit: Salesforce Developer (TechRepublic Premium)

Tyler Tanaka: It's been an interesting journey. We're over 60 years old, a family run, family owned company, but a lot of organic growth combined with M&A, and some joint ventures, some acquisitions, and some big mergers. And so there's been a lot of technical debt and legacy architecture and legacy systems and new point of sales and old point of sales that we've all had to bring together. One of the keys for us has been this move to an API first strategy, and unlocking all of those monoliths or siloed datasets into one nice, organized data lake, that is then able to be distributed and extended wherever we want. Our goal obviously is to make a better day for our guests on the road, professional drivers or a gas customer, and having that data be unlocked and accessible is extremely important.

Having all that data organized to begin with, the hard work that goes into putting it into an EDW and having it ready to be used. This API first strategy has given us the nimbleness and agility to be able to do what we want, even if we were to go in and merge or do an acquisition on a large property of locations. If we didn't have that flexibility, it would take a lot more work, a lot more time, and we wouldn't have the ability to be as flexible as we are as a business.

Bill Detwiler: Give me a concrete example of how you're putting that API first philosophy into practice. You can speak to that, whether it's a customer touch point or whether it's something behind the scenes.

Tyler Tanaka: I'll give you a good example. For some reason, we had ended up with multiple shower reservations systems being built. A lot of people don't realize that we're also a hotel; we're a concierge service. We have professional drivers staying with us in the parking lot, sleeping in their cabs overnight. We have over 80,000 parking spaces, and we do nearly 70,000 showers a day to help serve our guests.

Bill Detwiler: So it's a very distributed hotel, in a sense.

SEE: Hybrid cloud: A guide for IT pros (TechRepublic download)

Tyler Tanaka: Exactly. We're a conglomerate of hotel stops and a variety of different brands. For whatever reason, through building up over the years, we had ended up with multiple shower systems, and one system did maybe one particular thing. We wanted to give utility through our mobile app to these professional drivers who needed to reserve showers, so we built a shower reservation system that allowed drivers--once they break our geo-fence--to be able to use our mobile app to get in line. It then delivers them a code that they enter into the cipher lock on the door, and they can go take their shower.

Having an API strategy allowed us to be able to take all of those disparate datasets and disparate systems from over 750 locations in North America, and to be able to know how many showers are at each location, how many showers are available right now, and how to queue a driver in line, and then deliver him the code to be able to make a reservation and notify him when the shower is available. The goal was to provide this great utility to the professional driver, to allow him to save time and not be anxious and stressed out about his routine when he arrives on our property. Of course that API functionality let that take place, to take all that old architecture and turn it into something that is modern and new and available inside our mobile app.

Bill Detwiler: I imagine there were quite a few challenges to building the API that would bring the two systems together into a single lab. Talk a little bit about some of those challenges and how you overcame those challenges. What was it about the platforms you chose that made that process easy or even possible in the first place?

Tyler Tanaka: We chose MuleSoft prior to the Salesforce acquisition, so feel really good obviously. The companies that Salesforce chooses to buy are quite phenomenal. We've got quite a string going of other companies that we've chosen their products that have ended up getting acquired by Salesforce as well. So we chose MuleSoft because of that single runtime that platform delivers. It's removing all of those point to point--.net in our case--point to point in APIs previously, and then building the system layer and process layer APIs within the Mule code that then allows any developer to go in and write a new experience layer for whatever function we want to build out. We don't have to rewrite work over and over and over again that MuleSoft platform lets us build this muscle memory that you do the hard work once up front, and you've got all this reusable and repurposable code.

I think the big challenge is, of course, that there's a learning curve, like learning a new language, with anything. The investment that you make in your existing team members to train and go through the change management learning and adopting a new platform, I think was probably the hardest thing for us, but our team has been remarkable. They've adapted well; they have taken the time to learn and get trained up and have really embraced a new way of doing things. It just takes a little bit of time to get that momentum going.

Bill Detwiler: One of the things I talked to Uri Sarid, who's the CTO of MuleSoft, about was the leadership and a vision and the will within the corporation to invest the time, which you said, make those changes and start with that strategy. Was that a challenge within Pilot Flying J, or were stakeholders and business unit leaders able to see the benefit of moving to an API first structure from the beginning, or did you have to overcome any resistance--this is the way we've always done it, so we should keep doing it this way?

Tyler Tanaka: Yeah, that always exists. Change is hard for everyone. The real credit goes to our CIO, Mike Rodgers, who really has embraced an agile methodology within the IT organization and moving off of the traditional scope of work and waterfall. I think Mike is one of those leaders that quickly sees where we need to be further out and wants to empower and enable us as an organization to let IT be the business partners to enable all of the other business units towards success. The only way that we can do that is if we can be iterative and agile and athletic. He has done an amazing job of helping the company understand that if we want to get there, here's the important things that need to be required in order to help us get to that successful end point versus the traditional IT business relationship of 'go do my work,' 'go do my building IT.' We're really in a partnership alignment now through his leadership, which has been extraordinary to see.

Bill Detwiler: Something that we talk a lot about on TechRepublic is the shifting nature of IT. It really has shifted from people. IT was traditionally seen as a cost center, as the people that always say no, as a point of friction for the BUs. I think that barrier has had to come down really rapidly in the last decade. IT can no longer be an entity that is seen as just a roadblock or a barrier. It really has to be a symbiotic relationship and a partnership with the stakeholders on business.

Tyler Tanaka: That's right. Sometimes--and this is in my conversations with other digital leaders in this space, and I've heard a lot about this at Dreamforce this year--you can't say yes. There's no possible solution unless you started doing work a year ago or two years ago in order to make that possible and feasible today. This shift to adopting Salesforce as a platform for us, we're well vested in Salesforce and have committed to them as a platform. Sales Cloud, Service Cloud, Marketing Cloud, and a variety of other tools that they get bolted on.

We're a Tableau shop; we're a DMP; we're a MuleSoft customer now for APIs. That commitment three plus years ago has laid the foundation, and all the hard work that went into standing it all up now means that more often than not, the answer is always yes to whatever the business needs. The digital and IT teams are able to be extremely flexible. And now it's about problem solving, not about re-architecting or ripping and replacing. It's been really remarkable how fast we're able to move to help support the business needs and pivot in whatever direction we need to move as an organization.

Bill Detwiler: One of the other things that large organizations are dealing with is interoperability. Dealing with all those we touched on earlier, with legacy, this hardware you talked about, using the API strategy, API first strategy to overcome that, to bring those things together. Thinking about it now, were you already a Salesforce customer before the MuleSoft acquisition?

Tyler Tanaka: We only had Sales Cloud, and we had Marketing Cloud at that point. What we realized is that we still didn't have a way to move data even within the Salesforce environment because you had core Sales and Service and Marketing Cloud on two completely separate environments and two code bases. It was like you could do it through Salesforce connectors, and you could do it on point to point, but you didn't have the flexibility to do what you really wanted to do to move the business forward. That interoperability was a huge advantage to say, 'You know what? We're not going to rely on Salesforce to determine our own success here. We can't wait and pace their general releases or their connected strategy. We've really got to attack this on our own.'

A really good example is Josh Birdwell, who runs our retail technology; he's my peer from the business and technology side. He's stuck in the same thing with legacy architecture on the point of sale, and without having a way to tap into all these disparate datasets via moving this around in APIs, we'd be handcuffed. He and his team have done an amazing job of allowing us via mobile app or the kiosk--or whatever we want to invent on the digital product team--to be able to tap into that retail technology that's inside the store. It works on both cases in that example.

Bill Detwiler: You were using two different Salesforce products. It was MuleSoft and that API that you were already using to bring those together. Then the acquisition comes along, and the new announcements here at Dreamforce about some of the changes they're trying to do with APIs, with MuleSoft strategy with the APIs--in your eyes, are they going to make it even easier to do that.

Tyler Tanaka: That's right. I obviously thought that it was a brilliant acquisition because I already had seen the value of what we had as an organization using the product. I don't think that Salesforce may have even known how valuable it would be. The reason is because Salesforce obviously started as CRM and has grown into marketing. I actually think that all of the Marketing Cloud customers can now use MuleSoft to unlock extra value going the other direction. If you were just using Marketing Cloud, I think now MuleSoft opens up this entire additional cloud opportunity for those customers to move back into the core product versus core products moving into Marketing Cloud.

Bill Detwiler: On the other line.

Tyler Tanaka: Yeah, that's my take.

Bill Detwiler: I love the point that you made earlier about taking the time to skill up your staff, to take the time to invest in that. One of the things I know that MuleSoft announced was greater integration with Trailhead, Salesforce's training platform, around helping people create APIs. Talk a little bit about that. Is that something that you see as, 'Hey look, since we have this API strategy going forward, this is definitely something that our people can take advantage of.'

Tyler Tanaka: The last thing that we want to do is invest in a platform and rely on a systems integrator or an outside third-party to do all the heavy lifting and then leave with all the knowledge about how they built it all. Being able to have a progression of training from a junior developer all the way up to a senior architect is incredibly valuable. Being able to have that continual learning and continual knowledge growth to take the dependence off of these third parties that we would otherwise have to StaffOrg or run a hybrid is really important, and of course, it's free. It's included in what the platform already delivers to us. You're not having to go out and pay for additional training, and it's on their own pace; it's on their own time. It should be a huge advantage to be able to have the team go through that training.

Also see

Dreamforce2019-Detwiler-Tanaka