Leader Spotlight: Using predictions to tell a story, with Gabe Huerta

Gabe Huerta is Chief Product and Technology Officer at Urgently, a provider of digital roadside and mobility assistance technology and services. He started his career as a developer at Citysearch.com before being promoted to engineer and, later, Director of Product Design. Gabe continued his product management career in various product leadership roles at companies like Wasserman Media Group, Axon, Evite, Connectivity, and Reputation. Most recently, he was VP of Product at Thinknear by Telenav (acquired by inMarket) before joining Urgently nearly five years ago.

In our conversation, Gabe talks about the critical role of predictive analytics in an industry as complex as roadside assistance, where everything depends on real-world events. He shares his experience completely redesigning Evite’s outdated, 15-year-old platform from the ground up in only eight months, as well as how having a vision, roadmap, and clear innovation phases is crucial to undergoing a successful business transformation.

Dispatch-as-a-service

To start, could you provide an overview of what Urgently does?

Our mission at Urgently is to provide the best mobility assistance possible. We use the term “mobility assistance” because we handle not just roadside events, but also battery delivery, EV charging, etc., to the customers of our customer partners, large American and European automotive companies. We also provide services to some of the largest American fleet businesses.

In terms of our technology and platform, we essentially offer dispatch as a service. This is what we use to access the networks provided by our partners, and we also have our own networks and other integrations. In addition to that, we offer case management capabilities. We’re able to track jobs, see if customers are having issues, chat with customers, and handle the entire case experience.

The third aspect of the platform is our dispatching software for roadside service providers. Providers use this to accept jobs. We have an experience very similar to most dispatching engines out there, so they can manage their own jobs as well.

Revamping mobile strategy in 2009

Several years ago, when you were VP of product at Evite, you had a career-defining moment where you completely redesigned the platform from the ground up. What was that process like?

I joined Elite in 2009 and, at the time, Time Magazine wrote a piece on it and said it was one of the worst websites out there. Evite technology had aged and was cluttered with ads. I joined toward the end of that era back when it was owned by InterActiveCorp. I was part of the group hired by John Foley, who later became the CEO of Peloton. InterActiveCorp made it very clear that we had eight months to turn it around, so we got to work.

One of the key things for Evite was honing in on a particular audience. It was a 15-year-old app when I got there and was being utilized by not just its core audience of moms, but by businesses, clubs, and sporting events. It was difficult to make everybody happy — we had to redesign an application and we knew we couldn’t get to parity. We honed in on the 24–35-year-old married women with kids audience and used that as our base to prioritize.

We had a great customer support team that gave us a lot of customer feedback, and one of the things that came out of that was that users didn’t like the ads. They were cluttering up the experience. We tested a lot of things, like redesigning the actual invitations and creating a premium product that eliminated ads. That option to give users more control over the invitation excited people about our product. We used focus groups, customer surveys, and raw feedback. We had a great team and met weekly with customer support to prioritize properly.

Once you launched the new website and features, was there any feedback that particularly surprised you?

We had a lot of analytics around the gallery, and we knew we had to update the application to make it mobile-friendly. One of the things that we learned later was that 46 percent of people were accessing Evite through their phone’s browser. Keep in mind that this was 2009 when mobile devices weren’t as heavily used as they are today. So, we decided to focus on building an application. We built native apps as well as mobile web, and we made it completely accessible on iPads, etc.

All of a sudden, the people that we originally lost because they found it difficult to use Evite on mobile devices started coming back. We grew from converting around 10 percent on mobile devices to 56 percent in one year. It was tremendous to see mobile adoption.

You mentioned doing lots of testing. How did your team at Evite approach A/B testing as a method to validate ideas?

At Evite, we used a lot of frontend tools for optimization for the gallery. We were able to change product in the UI without engineering, and that helped us optimize the galleries and conversion points. There are a number of these tools in the industry when you’re thinking about merchandising or ecommerce. Every day, we tried different things like moving certain products or categories and evaluating how those things would convert.

Secondly, and we do this a lot at Urgently too, we had a framework to create a test plan. We understand business objectives and the good stuff that comes with that test plan, as we use tracking tools to measure those things. We use those analytics to determine whether the test has passed or failed. And we do that a lot. About 30 percent of our product development is around testing. That has led to a lot of great ways to innovate and drive the business forward in terms of success, pricing, customer happiness, and more.

---

Over 200k product managers and developers use LogRocket to create better digital experiences

Learn more →

---

‘The instrumentation that we need to fly the plane’

For quantitative data, what’s your process for determining which metrics are most important when you’re assessing a user’s experience?

Going back to the Evite example, we were similar to an ecommerce business in that we were all about conversion. The quantitative component is being able to understand how your product is doing in the context of the business. For us, this meant conversion and the guest-to-host ratios. That led to the business model, which was about ads and impressions. That’s what the quantitative brings to the table — how does your product align with what the business looks at as key success metrics?

Every morning, I wake up with my cup of coffee and I look at my scorecards. At Urgently, this is the instrumentation that we need to fly the plane. Without that instrumentation, it’s very difficult for us to understand how the product is doing against our goals.

The qualitative is more about insights and validation you’re hearing directly from the customer. They’re saying, “I like your product. It would be great if your product did this.” The insights that you don’t see in the quantitative data become tremendously valuable for product managers to say, “Wow, that person identified this pain point. I’ve never thought about that.”

Do you feel like you get enough of that qualitative picture by looking at those scorecards? A lot of product teams do on-site visits — do you employ those as well?

Yes, we do ride-alongs. That’s helped us quite a bit with service providers, but we have product managers who are in tow trucks and go to tow shows to talk to those customers as well. It’s a very fascinating industry. A big part of our insights stems from CSAT and NPS scores that we get directly from our partnerships. OEMs and fleet businesses obsess over that data. We use that as a good way of understanding whether customers are having a good experience. We get a lot of insights from things that go wrong, so it comes in many forms. We do a lot of testing through these real-life scenarios and truly learn a lot from them.

Using predictions to reduce complexity

How do you utilize predictive analytics to assess what might be needed for providers going forward?

It’s been a transformation. Roadside is a very complex business — everything depends on real-world events. We spend a lot of time on dynamic pricing and making predictions on pricing models, as well as making predictions on the providers themselves. Those predictions help us tell a story about provider performance and conversions in specific markets. Because this is location-based, there’s a lot of complexity around that. These are things that we look at from a modeling and machine learning perspective — day of the week, whether the location is urban, suburban, or rural, events, holidays, severe weather, etc.

When I joined the company, we went through COVID, which had a massive impact on the industry. People weren’t driving. That meant fewer tow trucks, so they let a lot of people go. Out of COVID, there was suddenly a lack of supply. We need to understand how those anomalies impact data and how we can adjust for that.

Do you have an example of a time when you saw a pattern in the data, validated it, and ended up actually moving forward with a feature?

When I was at Telenav, a MarTech business, we learned that through location data, you could have better visibility into who your customers are through where they go and what they do. This is an insight that the automotive or restaurant and dining industries didn’t have. It was another way of saying, “Wow, this is a level of insight that I don’t think anybody knows about.” We were happy to share this insight and it impacted our roadmap of delivering ways that you could buy into audiences using location-based data. That was an R&D project and we quickly got validation that people wanted to buy these audiences.

As a more recent example, at Urgently, we wanted to test a hypothesis around whether we can be transparent about pricing as an industry. We learned that there’s some way to control the predictability of pricing. so we tested that. Our data science and dispatch engineering teams spent months focusing on how to create an algorithm to consistently predict pricing without that volatility, which often drives extreme cases of bad experiences, bad costs, etc.

What’s your approach for handling the trade-off between innovation and other constraints, like time and budget?

I look at this two ways. The first is the strategic path toward innovation. At Evite, we needed to make it great, not good. It had been bad for a very long time, so coming out with something OK wouldn’t have been acceptable. When I was at TASER and we built Axon, that was the first time tasers were available in the digital space. It had to be great, not good. Having a vision and roadmap, as well as understanding those innovation phases, is crucial for businesses that need to transform.

Second, and the reason I pivoted my career into data, is there’s a lot of innovation you can get just out of testing. I leverage data quite a bit by taking insights, creating reports to visualize them, and then defining use cases. That data gives me validation before I even have to build something. I’ll talk to salespeople or clients and say, “I see this information and I can build this.” That validation is crucial. It makes sure you’re not going to go build all this technology and then come back later if it doesn’t get accepted.

Being inspired by what you’re working on

Lastly, what words of wisdom or advice would you have for people working in product who are hoping to move up in their roles or into product leadership?

Always be curious. That’s helped me quite a bit. Every day, there are questions about the business and the process, so don’t be afraid to just dig into something. I do that to this day. I don’t mind rolling up my sleeves and digging into data or running my own scripts to understand exactly what information I’m seeing that puzzles me. I encourage that. Do not look at this role as a management job where your contribution is just to manage KPIs. It’s really about curiosity and being inspired by the thing that you’re working on. Every day brings a new puzzle to solve, and I love that.

LogRocket generates product insights that lead to meaningful action

LogRocket identifies friction points in the user experience so you can make informed decisions about product and design changes that must happen to hit your goals.

With LogRocket, you can understand the scope of the issues affecting your product and prioritize the changes that need to be made. LogRocket simplifies workflows by allowing Engineering, Product, UX, and Design teams to work from the same data as you, eliminating any confusion about what needs to be done.

Get your teams on the same page — try LogRocket today.