06 Jun
In November 2022, a leading automotive brand launched its new luxury SUV, aiming to reach the right audience and boost sales through a targeted marketing campaign. Leveraging Facebook Ads and lookalike audience targeting, they sought to connect with potential customers who mirrored their ideal buyer profile. This case study explores the success of using smrtr’s lookalike audiences for lead generation within the automotive industry.
Building High-Value Audiences: Targeted Automotive Lead Generation
Our process begins by utilizing a rich set of consumer data points, including demographics, interests, personas, and past car buying behavior. This allows us to understand the characteristics that define successful car sales for similar vehicles and utilize lookalike modeling to identify a broader pool of potential buyers who closely resemble the brand’s ideal customer profile. By analyzing a list of quality leads provided by the automotive brand, we focus on high-value audiences to maximize their campaign’s lead generation potential.
How to Build a look-alike Model
At smrtr, we leverage advanced machine learning techniques to refine our custom lookalike model for our automotive partner. This ensures we select the most relevant audience for their campaign. We employed:
- Feature Engineering: To extract the most relevant insights from the rich consumer data set we possess, including demographics, interests, buyer personas and past car buying behavior.
- Data Optimization: We ensure its quality and usefulness by filling in missing data points, adjusting outliers, grouping similar data types and determine the most appropriate historical data range
- Data Balancing: Utilize a method called SMOTE to address potential imbalances within their sample data, ensuring the model doesn’t get biased towards specific subgroups.
- Model Selection: Once the data is optimized, we select the most effective prediction model for building the most accurate lookalike model (LightGBM, in case you were wondering), fine-tune its settings and rigorously test its performance.
Measuring Campaign Success
After supplying our automotive partner with hundreds of thousands of targeted, permissioned and compliant leads for their Facebook campaigns, we were all keen to see how the audiences had performed. In August 2023, smrtr securely matched campaign audience with actual vehicle sales data provided by our partner, Pentana. This privacy-compliant matching process utilized encrypted identifiers to ensure data security while measuring the campaign’s effectiveness.
Out of approximately 650 luxury SUV sales reported across Australia, Pentana identified 597 transactions with complete customer details. From this set, 277 individuals matched with those identified by smrtr’s models, and a remarkable 120 purchases were attributed to consumers targeted by smrtr’s audiences.
The custom SUV audience segment delivered the highest contribution, followed by the general SUV lookalike audience. Interestingly, another 11 purchases originated from individuals identified by other smrtr audience segments. To put this into perspective, if we randomly identify 1 Million potential clients, we would expect 14 buyers. This highlights the significant advantage of data-driven audience targeting.
| Audience | sales | % of sales |
| General SUV look-alike audiences | 24 | 17.9% |
| Custom SUV look-alike audiences | 85 | 63.4% |
| other smrtr audiences | 11 | 8.2% |
| random file of 1m individuals (not used for campaign) | 14 | 10.4% |
Quantifying Uplift Through Control Groups
By comparing sales figures against a randomly selected control group of 1 million individuals, we were able to effectively measure the campaign uplift delivered by each audience segment. The results were impressive, with the custom lookalike audience showcasing a lift of over 500% compared to the control group, highlighting the effectiveness of a highly targeted approach.
Key Takeaways
- Targeted Approach Drives Sales: This campaign demonstrates the power of targeted marketing in reaching the right audience and driving sales.
- Predictive Modeling Success: smrtr’s machine learning models accurately predicted customer behavior, with the bespoke lookalike model generating the most significant uplift.
- Measurable Sales Impact: By targeting the identified lookalike audiences, the campaign contributed to 120 new luxury SUV sales, bolstering the brand’s market presence in Australia.
Conclusion
The strategic collaboration between the automotive brand, their agencies, smrtr, and Pentana, leveraged sophisticated data modeling and analysis to ensure a successful launch for the new luxury SUV. By focusing on the most relevant audience segments, the campaign significantly amplified sales. This case study reinforces the significant impact of integrating advanced technology and data with marketing strategies to achieve exceptional outcomes in any sales-driven campaign.
