Unlocking data for fleet customers.
Completed in 2021 & 2022 while Director, Experience Design at VML.
All content is owned by VML and Ford.
How can we unlock data into a newly formed entity just for fleets and dealers?
After becoming it's own branch of Ford in 2021, FordPro sought to explore the depth of capabilities and possibilities that could be created using the mountain of data it launched with.
While at VML, s part of FordPro’s digital innovation team, I led design efforts across a suite of exploratory and applied projects aimed at redefining how commercial customers, fleet managers, and dealers interact with data-driven products. My role focused on strategic design leadership, driving experience vision and early product definition for initiatives that explored how FordPro could evolve its digital ecosystem—from intelligent recommendations to connected lot management.
FordPro was at a pivotal moment, transforming from a vehicle manufacturer into a data-driven mobility partner. Our goal was to unlock value from connected vehicle data and fleet behavioral and shopping data—helping businesses operate more efficiently while creating new digital products and services.
Across each initiative, we asked a central question:
“How might we use vehicle and behavioral data to make every decision smarter—for the driver, the dealer, and the fleet manager?”
The team I led, made up of a designer, a few data scientists and a project manger, looked at four critical ideas.
Can we look at vehicle data to proactively incentivize fleets to replace aging vehicles with electric versions?
How can we leverage information from vehicles to be sure that 100% of a lot can be ready for test drives or selling at a moments notice?
How can we use predictive data to help dealers get people in a suite of fleet products?
How can we use range data for fuel and battery powered vehicles to best map routes, time refuels, and save money?
Similar to the challenges Ford’s GDI&A faced, (link to that case study when live), Ford has not owned suites of digital products, instead has opted to outsource key needs to third parties or simply work in single use dashboards or excel-based products. For example, an independent third party, though branded as Ford, created their dealer portal Vehicle Locator+.
FordPro wanted to change that, creating Ford-owned, data-driven products that support dealers and fleets. The overall strategy on how to leverage the mountain of data Ford has was just being formed, so there was very little to go on. And, given the siloing of teams, there were mountains of overlapping ideas and competing analyses.
My team was tasked to tackled these main questions and along the way uncovered over $3 million in cost savings due to redundant initiatives and tabling ideas that users rejected.
Question 1
can we look at vehicle data to proactively incentivize fleets to replace aging vehicles with electric versions?
The creation of FordPro, by design, launched with vehicles like the Ford F150 Lightning and the Ford eTransit. These new vehicles tied with company and federal government incentives made them a prime target to replace aging fleets with new, electric vehicles.
One of the early projects my team was tasked with was to undertake how to use connected vehicle data to create a targeted microsite that looked at a fleet’s daily driving habits to identify where certain vehicles could be replaced with an electric counterpart. The product itself was straightforward, use a fleet identification number and pull in their data. While our data scientist dug into the available data within the API call, my designer and I dug into the main concerns of fleet managers and their reservations to switch to EVs.
Geographic Location
Geography would give us a few key pieces of data to consider. How temperature would affect battery life, the prevalence of chargers, and the change in incline of common routes.
Overnight Behavior
If vehicles were driven to the employee's homes each night, both distance and charging would be an added hurtle compared to vehicles being able to charge overnight at a company depot or warehouse.
Vocation or Fleet
The vocation or type of company is another key data piece, attaching a score to the type of work and therefore predicability of routes, the time the vehicle is parked, and whether the on-vehicle power would be a benefit and used.
Daily Routes
Tied with the fleet ovation, analyzing a specific vehicle’s daily route would quickly identify if the daily miles would be within the safe margin for an EV. These routes would also be used to look at proximity to chargers and elevation changes.
We took these data points and our initial algorithm that would provide insight to various teams across not just FordPro but the full Ford ecosystem. Getting approvals and the names to hold these workshops across the Ford organization was an up hill battle. However, our ultimate goal was to test the algorithm to be sure it was using the best method to predict battery life. I scheduled and ran individual meetings and one workshop with all teams to learn from their expertise. It was here we discovered two key issues.
First, we found five of the teams we talked to were undergoing similar products. All were looking at algorithms that would determine specific VINs that are good candidates for upgrades to battery. Secondly, each of these teams approached it differently, and each team has reasons why another team’s data would result in stranded drivers.
Ultimately, this led us to take a single recommendation back to the FordPro C-Suite - work with counterparts in Ford to pause all efforts in this until Ford can create an align approach to predict battery effectiveness in various uses, climates, and vocations. By shutting down the various, inconsistent approaches to this issue and collaboration together on one saved Ford an estimated $1 million in wasted efforts.
Question 2
How can we leverage information from vehicles to be sure that 100% of a lot can be ready for test drives or selling at a moments notice?
Vehicle Locator+ is a common tool across Ford dealerships. It is managed by FordDirect, a third-party partner that licenses data from Ford to serve dealerships directly. FordPro, however, sought to expand this ecosystem by introducing predictive data capabilities built from connected vehicle insights. The base version of the system is inventory management and the ability to trade vehicles between dealerships. In the advent of connected vehicles, Vehicle Locator+ was created to provide real-time data on a vehicle’s battery, tire pressure, fuel level and location.
This popular add-on came to fame after a string of thefts from dealers resulted in recovered cars since each vehicle was quickly located due to the vehicle’s location services and then reduced future theft. For Ford, the incentive to improve Vehicle Locator+ was to get all dealers connecting all vehicles to the cloud, providing direct revenue from dealers to Ford by enabling that feature. To increase that revenue, FordPro was looking to leverage the data and their data science team to increase the value of connected vehicles by diving into predictive AI.
First, my team looked at the available data from a connect vehicle and then began to craft a research plan that sought to understand what additional data dealers would be enticed by. Our research was split into three rounds.
Discovery Interviews
Our first round of research got us talking to managers or owners of dealerships, asking about their experience with the tool and where they felt it fell shot. The challenge in these interviews was finding that managers and owners were not as in touch with these tools as we expected, and our team, being representatives of Ford received unrelated pain-points about their overall experience of being a Ford dealer. Even these unrelated pain points were helpful, as it highlighted a disconnect between Ford and their dealers, it also led us to collect the specific names and generic titles of those who would help us understand how to shape the tool to their benefit.
We quickly dove into the second round of discovery interviews, using the information collected, we scheduled interviews with dealerships that focused on two roles. One was to target the maintenance and vehicle management team who takes care of daily car movements and preparations. The second is the lead sales manager and their team. In these interviews, we dug into the key concerns with each vehicle and their routine day-to-day tasks.
Sales managers emphasized that any customer should be able to test drive any car at any time — a challenge when hundreds of vehicles must remain ready on demand. They also sited our second user set as in their way. Porters and maintenance would tell sales managers that time was needed to identify what it would take for a car to be ready for a customer. We asked sales managers what would cause a test drive or sale to be delayed or hampered. The main concerns were ones that existing car data would solve, battery health and fuel level. These were items that would immediately delay a test drive. Other concerns were issues of the perception of the dest drive largely in vehicle cleanliness and the presence of flat spots on tires form vehicles that have not moved, causing a bumpy ride.
When we then asked to have the porter or maintenance staff join our interview, we heard similar issues - sales teams had high demands to prepare cars for test drives that had to slot between customers in for service and the list of vehicles that needed to be serviced on the lot. These teams highlighted that each day the priorities would shift as sales teams put in a list of needed cars and as the Vehicle Locator+ tool would highlight a vehicle low on fuel or battery.
Product Ideation
We asked ourselves: what would it take for every car to be test-drive ready? That led us to focus on four signals — battery, cleanliness, tire health, and fuel. From there, we explored how connected data might help dealerships anticipate issues before they occurred. From these interview we created initial prototypes and dove into the possibilities of the data against these goals. Some of those were clearer in the data than others. Our data science team worked to leverage AI to predict battery healthy by taking hourly battery data. We could fairly reliably inform dealership when a battery was within two days of dying based on the decay of the battery. We used location and time data to detmeurine what vehicle are likely to have flat spots in tires due to being stagnant.
We also attempted to use satellite data and location data to determine what vehicles may need cleaning based on proximity to trees or the dealership perimeter suggesting likely dust, this idea proved a large lift and unable to be predicted in our initial exploration. We also identified potential of using the dealers site data to predict what vehicles are showing interest and tie them to specific VINs on the lot, basically matching online customer interest to VINs that might be test driven.
We took screenshots of the existing Vehicle Locator+ tool and overlayed how these new insights might live within the tool with the goal of taking high level ideas back to dealers to determine value and provide a recommendation on an API between FordPro and FordDirect to implement these data insights.
Validation Interviews
Armed with our light designs and explanations of what was feasible with the insights, we met with some of the same dealers and new ones to talk about how these products might influence their use of Vehicle Locator+ and effect their overall business. In these interview we brought up our new product features and how they may live in the system.
Ultimately, dealers were highly interested in the ways that the tool could solve the disconnect between sales and maintenance. They spoke highly of the various features we add, it was, in these third interviews we uncovered an issue that plagues a lot of Ford products, disconnection. Sales teams use and love Vehicle Locator+ to trade vehicles and quickly check in before they take a customer to them. However, Vehicle Locator+ has no real tools fro maintenance teams, they leverage various third-party tools to manage daily tasks and plan their teams. So, without also supporting a suite of tools for maintenance, the effort to bring these the data to life with an API would not provide the value unless FordDirect underwent a significant shift in the priority and target users of their product.
Results
As a result of our in-depth research and exploration, we were able to identify key data insights that could be added to a backlog and showed value. However, by advocating for meeting with dealers and conducting in-depth research we were able to recommend a pause in creating an API that would provide minimal value given the challenged in Vehicle Locator+ that are outside of Ford’s control.
This work was in partnership to the word to establish a UX practice in both GDI&A and FordPRo. Projects like these didn’t just change backlogs or products, it showed the fundamental benefit of strategic design thinking and changed both the perception and feeling of bringing product designers into data products early on the product lifecycle. Rather than push forward with a costly integration that would deliver limited value, our research allowed FordPro to make an informed decision — saving $1.5M in development costs while identifying the deeper opportunity to build a truly connected suite that serves both sales and maintenance.
Question 3
How can we use predictive data to help dealers get people in a suite of fleet products?
The FordPro leadership initiated a directive to build a recommendation system similar to Netflix and Amazon, but with significant ambiguity around the target users (customers or dealers) and product scope (vehicles, accessories, or services). This high-priority initiative required clarification on fundamental questions before development could proceed effectively. My team, Analytics Expansion, was tapped to explore how a peer-based system could identify opportunities for fleet customers.
Task
I was tasked with leading a small team consisting of technical partners and a designer to tackle these foundational questions while partnering with a third-party tasked with the LLMs creation. We needed to map out FordPro's product and service ecosystem, understand the customer journey, and identify decision-makers in the purchasing process. Our objectives included conducting internal expert interviews to extract institutional knowledge, collaborating with Jackson Dawson (who was researching tools for Commercial Account Managers or CAMs), and aligning with the technical team already exploring Large Language Models (LLMs) that could connect similar commercial customers and match them to products based on value and usage patterns. The Product Design team faced challenges in securing backing and access to dealers for validation interviews, making it critical to develop a clear problem statement that would gain stakeholder support.
Action
I initiated internal interviews with experts across Ford and FordPro. I leveraged existing research from Jackson Dawson on tools supporting CAMs to gain insights without duplicating efforts. I facilitated numerous cross-functional workshops to synthesize findings and identify key opportunity areas. Working alongside the technical team, I helped evaluate how an LLM-based system could identify patterns among similar businesses (e.g., plumbing companies of comparable size in similar regions) to generate relevant product recommendations. Through discovery interviews with CAMs, I uncovered critical insights about their workflow and identified a specific gap in their product knowledge: while CAMs were confident with physical products (vehicles, add-ons, accessories), they struggled with non-physical offerings like insurance, service, and financing products.
Result
Our research led to a pivotal insight: a recommendation tool focused on non-physical products (insurance, financing, service plans) would provide the greatest value to CAMs, their customers, and Ford. This finding significantly narrowed the scope from the initial broad concept to a targeted solution addressing a genuine user need. However, we identified a critical implementation barrier: the main tool used by CAMs is not owned by Ford, creating concerns about integrating Ford's proprietary LLM technology into a third-party platform that might expose valuable insights to competitors. Despite developing a viable concept with clear user value, the project was placed on hold until Ford-owned platforms could be developed to host this capability. The work remains valuable for future implementation when the right platform becomes available, and the research insights continue to inform other FordPro initiatives targeting CAM support and commercial customer needs.
More Coming Soon!
I am still writing the last story - our route management tool. I am always happy to chat about this work in the meantime, pleas reach out!