Velocity Motors: Edge AI Slashes CPL 38% in 2026

The year 2026 presents a new frontier in mobile advertising, where the precision of targeting directly correlates with campaign efficiency. One recent campaign for a regional automotive dealership group, “Velocity Motors,” demonstrated the far-reaching potential of edge AI for mobile video ads, achieving a 2.3x uplift in qualified leads compared to previous benchmarks. Velocity Motors sought to drive local showroom visits and test drives for their new electric vehicle (EV) lineup across three key dealerships in the greater Atlanta metropolitan area: one in Buckhead, another near the Perimeter Mall, and a third in Alpharetta. The question for us was: could on-device vision processing deliver granular, real-time targeting that traditional server-side AI could not?

Velocity Motors EV Campaign: A Deep Dive into Edge AI Implementation

Our objective for Velocity Motors was clear: generate high-intent leads for their new electric vehicle models. Previous campaigns, while successful, often struggled with audience saturation and a high volume of unqualified inquiries. We hypothesized that on-device vision processing, a core component of edge AI, could provide unparalleled contextual understanding, allowing us to serve hyper-relevant video ads at the precise moment a user was most receptive. The budget allocated for this three-month pilot campaign was $180,000, running from January 1 to March 31, 2026.

Strategy: Contextual Relevance Through On-Device Vision

The core strategy revolved around using the user’s mobile device to analyze real-time environmental cues. This meant moving beyond demographic and interest-based targeting. Our focus was on identifying users who were actively engaged with automotive content, specifically electric vehicles, or were in a physical context that suggested automotive interest. For example, a user browsing EV reviews on a publisher app while waiting at a charging station or driving past a competitor dealership could be identified. This wasn’t about tracking individuals, but about understanding the immediate context of the device user. We integrated with a specialized mobile ad platform that supported client-side AI processing, ensuring user privacy by keeping raw data on the device.

We specifically targeted iOS and Android devices within a 15-mile radius of each Velocity Motors dealership. This geographic boundary, encompassing areas like Sandy Springs and Dunwoody for the Perimeter Mall location, was important for driving showroom traffic. We configured the AI to prioritize signals such as interaction with automotive content (e.g., watching YouTube reviews of EVs), location data indicating proximity to EV charging infrastructure or competing dealerships, and even accelerometer data suggesting active driving. According to a 2025 IAB report, mobile video ad spend continues its upward trajectory, making precision targeting an imperative for effective budget allocation.

Creative Approach: Dynamic Video Personalization

The creative strategy was equally critical. We developed a suite of short-form video ads (15 and 30 seconds) showing different features of Velocity Motors’ EV lineup: range, charging speed, interior luxury, and performance. The key was the dynamic creative optimization (DCO) powered by the edge AI. If the AI detected a user was, for instance, in a high-traffic urban environment (suggesting range anxiety might be a concern), they would be served a video emphasizing the EV’s rapid charging capabilities and extensive city range. Conversely, a user frequently seen in suburban areas might receive an ad highlighting family-friendly features and quiet operation.

We produced 12 distinct video assets, each with multiple voiceover variations and on-screen text overlays, allowing for hundreds of potential combinations. This level of granular personalization would have been impossible with traditional server-side processing due to latency and data transfer limitations. The goal was to make each ad feel like it was speaking directly to the user’s immediate needs or interests, based on their real-time context.

Targeting Refinements and What Worked

The initial phase of the campaign, the first 30 days, focused on broad targeting within the defined geofences, allowing the edge AI models to gather baseline contextual data. During this period, our Cost Per Lead (CPL) was $75, with a Click-Through Rate (CTR) of 1.2%. While acceptable, we knew we could do better. The real gains began in February when the models had enough data to start identifying stronger correlations between on-device signals and conversion likelihood.

One particularly effective refinement came from analyzing users who frequently visited public EV charging stations, such as those at the Georgia Power EV Charging Network locations around Atlanta. The edge AI could detect when a user’s device had been stationary at these locations for more than 15 minutes. For these individuals, we prioritized video ads that highlighted the benefits of upgrading to a newer EV model, often featuring trade-in incentives. This specific segment showed a CTR of 2.8% and a CPL of $42, demonstrating the power of highly contextual targeting. Another strong signal was the presence of other automotive applications on the user’s device, particularly those related to car maintenance or used car listings. These users were clearly in an automotive mindset.

What Didn’t Work and Optimization Steps

Not every aspect yielded immediate success. Initially, we experimented with targeting users based on detected vehicle types from background noise analysis (e.g., engine sounds). This proved unreliable and led to low engagement, as the audio analysis was often inaccurate or misinterpreted. We quickly deprioritized this signal after two weeks, reallocating budget to more strong contextual cues like location and app usage. This rapid iteration was possible because the edge AI provided immediate feedback on signal efficacy, allowing for daily adjustments to the model’s weighting of different contextual factors. This is a critical advantage: you don’t wait for weekly reports. You see performance shifts in near real-time.

Another challenge involved creative fatigue. Even with dynamic elements, some users saw similar ad variations too frequently. To counter this, we implemented a frequency capping mechanism at the device level, ensuring no single user saw more than three ads from the campaign within a 24-hour period. We also introduced a rotating set of “hero” creative assets, refreshing the primary video narratives every two weeks. This simple adjustment helped maintain viewer interest and prevented a decline in CTR. We also found that 15-second ads outperformed 30-second ads by a margin of 15% in terms of completion rate, especially when served in mobile app environments. This led us to shift 70% of our budget to shorter formats in the final month.

Metrics and Results

By the campaign’s conclusion on March 31, 2026, the results were compelling. Velocity Motors achieved a total of 12.5 million impressions across the three-month period. The overall Click-Through Rate (CTR) for the campaign averaged 2.1%, significantly higher than the 0.8% benchmark from previous, traditionally targeted mobile video campaigns. The real success was in the conversions. We tracked 2,400 qualified leads, defined as users who completed an online test drive request form or called a dealership directly after viewing an ad. The average Cost Per Lead (CPL) dropped to $58, a 38% reduction from the initial phase and a substantial improvement over the $95 CPL from prior campaigns. Each qualified lead was valued at $220 based on historical conversion rates to sales. This yielded a total Return on Ad Spend (ROAS) of 3.8:1 ($220 * 2400 leads / $180,000 budget), far exceeding the client’s 2.5:1 target.

The cost per conversion, in this case, a qualified lead, was a direct testament to the precision of edge AI targeting. We saw conversion rates from ad view to lead form submission increase by 1.7x compared to previous efforts. The data speaks for itself: better targeting means less wasted spend and more valuable interactions. I often tell clients that the future of mobile advertising isn’t just about reaching more people. It’s about reaching the right people, at the right time, with the right message. Edge AI makes that aspiration a tangible reality.

The success of the Velocity Motors campaign highlights a significant shift in mobile advertising capabilities. By processing contextual data directly on the user’s device, edge AI delivers unparalleled targeting precision, leading to higher engagement and more efficient ad spend. For marketers, embracing these on-device capabilities means moving beyond broad strokes to craft truly personalized, impactful video ad experiences.