The mobility retail market is in for a period of change and disruption as the world shifts away from traditional fuels. Oil demand growth shows signs of slowing and across multiple scenarios demand is expected to begin to decline by 2030.1 This shift in fuel focus is presenting an opportunity for both established mobility retailers and market entrants to define new offerings using AI and advanced analytics (AA).
As electric vehicles (EVs) become more prominent, the customization of service stations may prove crucial to future success.2 Intelligently harnessing AA or AI can help mobility retailers customize their value proposition for each station and unlock a wide variety of use cases that could help improve returns, achieve cost savings, and drive customer engagement.3 Some of these use cases include personalization in customer loyalty programs, fuel price optimization, labor activity improvements, station network optimization, and convenience retail optimization.4
This article dives into the latter two use cases—station network optimization and convenience retail optimization—exploring how mobility retailers can use AA and AI to tailor each service station to their customers’ preferences and seize opportunities in the changing landscape.
How the convenience revolution could advance business
Consumers demand more convenience today. They are looking for fresh and frequent consumption, delivery of goods, their needs being met while on the go, and a frictionless customer experience. This trend could play into the favor of mobility retailers—if they offer the right product, in the right place, at the right time, and for a fair price.5
The intersection of the energy transition and the convenience revolution presents several opportunities for growth—the fundamental question now is how to effectively harness both trends to position a mobility retail business that best serves customer needs. Businesses that differentiate themselves, either via lower costs or a more elaborate value proposition, could be the ones to succeed.
The new mobility retail opportunities emerging from the energy transition are showing attractive value growth opportunities, specifically in EV charging infrastructure (EVCI) and nonfuel retail (NFR).6 Fast growth in these areas is creating space for new entrants and may spur traditional forecourt retailers to rethink their approach and business model (Exhibit 1).
Together, new mobility opportunities represent more than $100 billion in value spread across the globe.7 This value growth differs geographically, driven by GDP growth forecasts and the impact of the energy transition on different regions. By 2030, Europe could hold $32 billion of this value and see 2.1 percent growth per year in the retail EBITDA value pool.8
McKinsey analysis shows that many large global retailers in select markets have already seen NFR become more profitable than fuel. This has resulted in increasing EBITDA multiples in M&A transactions and new fuel retailers entering the industry.
However, even within a region, new value pools will not be the same across all stations. The performance of a quick-service restaurant, for example, will strongly depend on the location of the station, the characteristics of its customers, or the existing restaurants nearby. Similar issues will affect an EV charger or a car-care service, for example. Determining where to deploy each service and value proposition to maximize the return of every square meter will likely be a critical success factor for fuel retailers.
Understanding customer behavior and building the right offers at each station will therefore be key—and AA can be applied in three domains to capture value in the retail business: building the network of the future, optimizing convenience retail by leveraging transaction data, and driving value from fuel pricing, loyalty, and operations.
Building the network of the future
Most mobility retail networks are designed and optimized for fuel and its associated services. However, with the increase in transport electrification, mobility retailers are transitioning from selling fossil fuels only to a variety of different types of energy sources that power passenger cars, light commercial vehicles, and trucks.9 Compared to today, the heavy-duty market in particular will see an amplification of required energy sources.
In the age of decarbonized mobility, designing the network of the future could prove challenging. Tapping into the electric mobility opportunity is not as simple as installing a charger and a plug, given the varying potential for individual service stations to provide such services.10 Some stations are well suited to serve as flagship mobility hubs, while others may remain better suited to traditional fuel and convenience stores (Exhibit 2).
Transaction data can help mobility retailers gain deep insight into consumer behavior and, if it is translatable, to EV driver preferences. These two dimensions combined provide a comprehensive picture of station-level economic outlook regarding fuel, EVs, and NFR.
Determining when and where to build EV charging infrastructure
How to build the network of the future
Fuel retailers can use transaction analytics combined with customers’ likely EV charging behaviors to design the network of the future. EV uptake will impact the profit and loss of individual stations differently, mainly driven by three factors:
- Location, which determines the EV penetration rate, overall vehicle traffic, and the share of the market for on-the-go charging.
- Consumer behavior, which determines the role stations play in people’s lives.
- Station properties, which determine the potential for EV charging infrastructure (EVCI) installation from the perspective of electrical power and parking space availability.
A three-step, station-level opportunity assessment can be done to see which stations are ready to refine their offerings. The assessment first defines the station-level profitability; second, its fuel and NFR potential after repurposing; and third, the optimized future network that maximizes new value mechanisms such as EV charging.
This type of analysis can help mobility retailers quantitatively understand the different station archetypes and how to focus their investments on the right places within their networks (exhibit).
By using carefully chosen machine learning (ML) models for traffic prediction and consumer behavior modeling, mobility retailers can make data-informed investment decisions and prepare their network for an electrified future (see sidebar “How to build the network of the future”). We have seen ROI improvements of between 10 and 20 percent from data-driven decisions in capital expenditure (capex) spend.
How retailers are using machine learning: Case studies
Machine learning (ML) has already been used to improve the mobility retail offer at different service stations. Here, two case studies highlight what fuel retailers are doing to succeed in a changing environment.
One retailer took clear action to lead in the energy transition and capture value. After analyzing more than 200 million transactions from the point-of-sale system, an unsupervised ML model suggested clustering stations into archetypes based on the role these stations play in people’s everyday lives (exhibit). This allowed the retailer to map the future network footprint and release significant capital expenditure (capex).
While these insights from ML models are very helpful, execution is key. Building the right set of capabilities to act on insights may define the most successful retailers and ensure they lock in the best locations.
In another example, a retailer used a model to list the important predictive factors of each location, with the team validating locations against each factor. Building such capabilities can bring transparency to the site selection process and, once the site has been set up, support continued monitoring against established KPIs.
Some retailers are ahead of the AA and AI curve and have built in-house capabilities for EVCI based on ML algorithms that can make hypergranular predictions about EV charging locations using a three-step approach (see sidebar “How retailers are using machine learning: Case studies”).
The first step is to identify the local data that impact EV charging, such as local EV registration volume and traffic, socioeconomic factors, competition data, points of interest (POIs), and number of available chargers in the market.
The second is to predict how patterns will shift in the future by forecasting the relevant drivers—including socioeconomic factors, car sales, and EV penetration—to generate scenarios on how competition may evolve and understand the required charging power.
The third step is to train and run ML algorithms to determine the impact of EV uptake scenarios on different operational metrics—for example, how much the chargers will be used and the corresponding energy sales. These models enable operators to achieve hyperregular advice on suitable locations for EV charger installations, the number of chargers needed, and their ideal power output.
Using transaction data to improve nonfuel retail
Customers who are charging EVs will have an extended dwell time of over 20 minutes, making the nonfuel offer more important than ever.11 While global fuel and NFR value pools are expected to grow in the years to 2030, with NFR projected to increase by up to 40 percent, retailers are uncertain about how to capture this value from NFR.12
For pure-play retailers in other industries such as grocery retail, having a differentiated customer value proposition has been key in driving value.13 Mobility retailers have started taking notes, focusing on reinventing their customer value propositions (such as understanding customer behavior at the individual transaction level to define microassortment); executing the customer value proposition, including tracking stockouts, promotions, customer engagement, and experience; and developing models to assess the “share of wallet” for each location and identify gaps between the current versus “fair” share of income.
Best-in-class retailers are leveraging large cloud-based platforms and ML models to collect store information and analyze over one million transactions across their network on a daily basis. Transaction analytics have helped retailers identify operational interventions across different areas, for example, category distribution, top-selling SKUs, and category sales rhythm.
In a recent pilot store in Asia, a wide variety of AA- and AI-driven modules were tested together to understand the ultimate potential of a retail outlet. The techniques used ranged from a system automatically generating tasks for store teams to automated store-level assortment, automated planograms, seamless checkouts, personalized promotions, and many others (Exhibit 3). The result was an absolute gross profit increase of over 40 percent, largely driven by cost optimizations from the automated store.
One mobility retailer used its data to prioritize the sale of popular items across its network of stations, leading to an increase of 5 to 10 percent in gross margin on convenience sales across its stores.14 Other retailers have typically seen a doubling of convenience EBITDA over four years by implementing a combination of levers, such as dynamic pricing, stockout prediction, and cross-sell bundle recommendations.
Three steps to building the network of the future
Mobility retailers are sitting on valuable data. Executing on AA and AI insights is not simple but doing so is critical—failing to harness these insights could leave mobility retailers with the wrong product offering when consumers seek out new types of services and convenience. Mobility retailers looking to embed AA and AI could consider the following three steps:
- Getting the basics right. Having the digital infrastructure in place throughout the organization to run and interpret data insights is a foundational requirement for mobility retailers. However, it can be difficult to own “clean” data that is complete and correctly formatted across offerings and operating models. Some retailers do not have access to data from their dealers or real estate partners, leading to a partial view of their customers and missed insights that would allow them to embrace their full potential. Ensuring a complete view of data is a big, yet complex, value unlock.
- Setting high ambitions. Retailers can help drive the decarbonization of the transportation sector by setting ambitious goals and repurposing existing assets. Tough choices may have to be made. Some stations are not future-proofed or may require repurposing as fuel sales decline, which may reduce their convenience retail footfall. Key strategies to navigate this change include selecting the right use cases, developing a road map, being agile enough to test something fast in the market, and adopting a customer-back proposition that starts with designing products and services from the customer’s perspective—leading to a product pull versus the historic product push.
- Honing digital capabilities. Customers are used to integrated customer journeys in the online world and they increasingly expect the same from offline retailers.15 Customers want to be recognized, rewarded, and receive personalized offerings across a retailer’s portfolio. Retailers who do this well earn a bigger share of the customer’s wallet.16 Mobility retailers can analyze past transactions to offer personalized promotions and offers on seemingly not-so-adjacent products. Investing in experts (such as data engineers and data scientists) can allow retailers to use data to create meaningful partnerships (for example, with hyperscalers or established retailers that allow access to new capabilities or value pools). Data could become the foundation for predictive use cases, based on generative AI, such as higher-level personalization, optimization of repetitive work (such as customer service), and supply chain optimization. It is therefore not surprising that 56 percent of CEOs globally consider these AI cases a core focus for their organization, with 64 percent expecting to increase their investments in the technology.17
However, while the opportunity is clear, many companies fail at leveraging data and technology. This is because “softer” factors come into play, too. Analyzing a wide variety of digital transformations reveals important elements for a successful approach, such as strategy, talent, operating model, and change management. These journeys start with small steps to ensure that the entire organization is on board, not just top management. The organizations most likely to see the highest returns from AA and ML may be those that take a holistic approach, rather than treating technology as a “bolt-on” capability.
These are unprecedented times for traditional mobility retailers. However, AA and AI can give mobility retailers the insights they need to survive and thrive. Those retailers that take a customer-back approach and build their internal muscle to turn data-driven insights into action may be at the forefront of enabling and harnessing this new fuel wave.
