At its core, Carwow functions as a dynamic marketplace where dealers supply cars, and customers seek to buy or sell vehicles. For a platform like this, customer acquisition is vital, with Carwow using a blend of organic and paid channels to drive growth. Among these, paid channels such as Google Ads are pivotal in attracting and connecting users with dealerships. However, the initial structure of Carwow’s Google Ads campaigns revealed several critical challenges that impeded their ability to scale effectively.

 

In their original setup prior to the project laid out below, Carwow’s campaigns operated under a Target CPA (Cost Per Acquisition) model. The goal was to acquire signups at a set cost per conversion, encouraging users to begin the car-buying or selling process. 

 

Yet this didn’t account for the full customer lifecycle, as revenue is tied to completed car purchases – a process that often spans weeks or months. This lag left a significant gap between conversion reporting in marketing platforms and actual realised revenue for the business.

 

 

The extended conversion timeline therefore posed a challenge for the marketing data team, who were keen to predict the potential future value of signups. Without immediate revenue feedback, they relied on predictive models and historical data to make decisions, introducing uncertainty and limiting their ability to act in real-time. 

 

Optimising campaigns became a complex endeavour, requiring frequent manual adjustments to ensure alignment between ad spend and eventual returns. To further complicate matters:

 

• • Campaign Fragmentation: Carwow managed a highly segmented campaign structure, with campaigns broken down by car models. Each segment exhibited unique performance metrics and conversion rates, necessitating individual CPA targets. This fragmentation demanded constant monitoring and updates, stretching resources thinly.
  • Manual Adjustments: Given the delayed feedback loop in Google Ads, the media team often found themselves making frequent, manual adjustments to CPA targets. These adjustments were necessary to account for shifts in customer behaviour, market trends, and dealership demands, but they consumed valuable time and resources.
  • Misaligned Optimisation Goals: Google’s algorithms optimised for signups (the top-of-funnel consumer metric for Carwow) rather than events that generated actual revenue. This misalignment meant that while campaigns might achieve high signup volumes, they were not necessarily driving the downstream revenue outcomes critical to the business.

 

The cumulative effect of these challenges was an inefficient and labour-intensive process. The marketing team had to spend a significant amount of time managing campaign settings, reconciling discrepancies and predicting future performance, thus reducing their capacity to invest in strategic ideation. The delayed conversion cycle also left potential revenue untapped, as decisions were being made based on incomplete data.

 

In that setup, optimising campaigns felt like navigating a maze with a blindfold on. The lack of real-time feedback made it difficult to adapt swiftly to changing conditions or seize growth opportunities. The limitations of the model not only restricted Carwow’s ability to scale but also hindered their paid campaigns from achieving their full potential.

 

The solution? A transformative overhaul of their revenue model, alongside a reimagined approach to data tracking, attribution, and integration within their advertising ecosystem.

Carwow and WeDiscover came together with a goal of creating a data-driven and automated approach that addressed the core issues strangling growth, a process which they saw as involving a foundational re-building of the approach to data usage within the account.

 

 

 

1. Perfecting Session Tracking

 

The first step was constructing a comprehensive data infrastructure. Using Amplitude, Carwow meticulously captured every user interaction across their website, from clicks and page views to UTM parameters (and many more). This highly granular level of tracking allowed the team to build a detailed map of the customer journey, uncovering actionable insights at every stage of the funnel.

 

The potential challenges with session tracking were numerous, and included missing UTM parameters and ad blockers or browsers blocking certain aspects of tracking, leading to the collection of incorrect information needing to be untangled in the data layer to then provide accurate data.

 

Another significant challenge was the size and volume of information that needed to be dealt with. The Carwow site sees a significant amount of traffic, with a large quantity of information that needs to be quickly processed. Given the scale of the operation, efficient data layer structuring had the potential to drive significant cost savings related to data processing. 

 

Solving all these challenges took over 12 months. A new structure was designed with efficiency and flexibility in mind, developed in DBT using incremental models wherever possible. The first layer processed web events and built out user sessions. Here, an event represented any action occurring on the website, while a session combined multiple events to capture a single user’s complete visit.

 

The transformation is computationally intensive, and for large companies like Carwow, which handles millions of visits annually, it can become very costly to run. Therefore, efficiency and incremental data design were essential.

 

The second layer enhanced Amplitude’s data with marketing information, such as mapping out attributes like campaign IDs to their latest names. The third layer then built out the final table, processing the information from all previous layers and enhancing it with channel classification logic and session-to-user mappings to ensure that all known information was accurately tracked. 

 

By organising data with precision, Carwow created a crystal-clear picture of user behaviour, enabling more informed decision-making and laying the groundwork for effective attribution and optimisation.

 

2. Cracking Attribution with LinDDA

 

A pivotal component of the transformation was developing Carwow’s proprietary attribution model, known as “LinDDA” (Linear + Data-Driven Attribution). LinDDa is a hybrid model that’s powered by a linear click-based internal attribution model, and enhanced by Google Ads’ own logic. This allows for platform-led data-driven re-allocation of credit, using the touchpoints known to Google.

 

Click-based attribution models alone reduce credit to YouTube and Display’s impact in the customer journey, so allowing Google to redistribute credit by utilising its own in-platform signals around visitors gave an opportunity to improve overall performance. 

 

Unfortunately, this hybrid approach is not available within all publishers, so most other channels could not benefit from having the complete LinDDA redistribution applied to them, instead using internal linear attribution. 

 

If a customer was driven to site by a Google search ad, an organic visit, or a YouTube display ad however, LinDDA ensured that each interaction received its fair share of revenue attribution. This approach provided a balanced view of campaign performance in-platform, and improved the value of multi-channel interactions.

 

 

 

3. Synchronising with Google Ads

 

Uploading attributed revenue data back into Google’s system via offline conversion uploads then enabled Google to have exactly the same goals as the business. This – paired with the ability to use DDA to reassign revenue based on Google’s powerful algorithms – allowed Google to not just optimise for what mattered to the business, but also for it to use every available ad format seen to have impact to achieve its goals. This created a seamless feedback loop wherein Carwow’s in-depth insights worked in tandem with Google’s optimisation capabilities.

 

The use of DDA allowed Google to move attributed revenue from late-stage ad categories (brand) to early-stage formats (YouTube, Display, generic search), thus moving budgets from ad categories that are not easily influenced to the ones that are. It was a model cognizant of the fact that brand activity has a natural ceiling – generally being a space near-completely owned by the brand itself – while the other ad categories could often benefit from more budget – if being seen to be driving value. 

 

The process evolved through several iterations of tooling for synchronising the data with Google Ads, including the use of custom solutions, finally landing with Hightouch, which was seen to deliver a high-quality service with minimal downtime. 

 

4. Optimising for Target ROAS

 

With accurate revenue data and attribution in place, Carwow was then ready to safely transition from fragmented CPA targets to a single, unified Target ROAS (Return on Ad Spend) model. With in-platform information having alignment with internal objectives, requirements on the team around making regular manual adjustments to targets were largely mitigated, with Google Ads campaigns instead being automatically optimised at scale. 

 

Freed from those routine optimisations, time was opened up for the marketing team to focus on strategy, innovation, and scaling initiatives. With that time, the team has since: launched Google Vehicle Ads, integrated into the operation 2 different acquisitions, launched activity in new markets, integrated new products, and experimented with integrating the most advanced Google Ads solutions into their operation.

 

Prior to the project, Carwow’s media team spent a significant portion of their time manually adjusting campaigns, predicting conversion values, and managing a vast array of Target CPA figures. The scale of it resulted in low capacity for creative strategising and proactive scaling.

 

The new system was therefore able to drive a fundamental change. With advanced tracking, the integration of LinDDA, and bidding straight to revenue generating events, the account transitioned to a single, static Target ROAS (Return on Ad Spend) model. This shift allowed Google Ads’ algorithms to optimise campaigns automatically, eliminating the need for manual intervention, and stabilising Google’s observed objective. 

 

With the reduction of time spent on reactive adjustments, the team was then able to increase their focus on strategic initiatives and innovation. As a result of the more efficient campaign management, attention then moved to actions that made iterative improvements to areas like data quality. The streamlined system provided the foundation for consistent performance and scalable growth.

The impact of the changes has been profound, delivering measurable improvements across key performance metrics:

 

• • Increasing Revenue Contribution: With campaigns now running more smoothly and profitably, Carwow has dramatically increased the revenue contribution from Google Ads, driving unprecedented growth.
  • Bolstering Efficiency: ROAS increased by 80%, even alongside increases to total spend.
  • Empowering Creative Innovation: Freed from time-consuming manual campaign optimisations, the marketing team has shifted their focus to building new campaigns, testing creative ideas, and scaling successful strategies.

 

 

The project therefore not only enhanced operational efficiency – it also sparked a cultural shift within the team that they had been striving for: moving from investing a high proportion of time and being overwhelmed by manual adjustments across a vast array of activity, to instead working on more innovative projects designed to drive further business growth. By eliminating repetitive tasks, Carwow unlocked the full potential of its marketing talent and technology.

The remarkable results achieved were driven by four critical factors:

 

• • Aligning Goals: By syncing their internal data with Google Ads, Carwow ensured that both the team and Google’s algorithms worked collaboratively towards the same revenue-focused objectives, eliminating the inefficiencies previously present as a result of misaligned optimisation goals.
  • Investing in Data: Accurate session tracking and a scalable data infrastructure formed the foundation for smarter, data-driven decisions. Tools like Amplitude provided actionable insights that empowered the team to optimise campaigns effectively.
  • Trusting Technology: Leveraging Google’s advanced attribution and optimisation tools allowed Carwow to unlock performance gains that manual adjustments alone could not previously achieve. The synergy between technology and strategy was key to achieving sustainable growth.
  • A bold vision: Overhauling the revenue model, redesigning the internal attribution system, and reconfiguring Carwow’s largest acquisition channel required assertive and confident decision-making. All stakeholders had to trust in the team that the substantial changes being made would ultimately yield results – a vision that paid off with transformative outcomes.

Carwow’s impressive results exemplify the transformative power of alignment, automation, and trust in data-driven decision-making. For businesses looking to scale their own activity, Carwow’s experience offers a roadmap:

 

• Shorten Feedback Loops: Accelerate decision-making by improving data collection and refining your revenue model.
• Invest in Data Infrastructure: Adopting tools and working on projects designed to develop robust attribution models can turn fragmented data into actionable insights.
• Leverage Platforms: Ensure alignment between platform optimisation goals and your business objectives to maximise impact.

 

Armed with robust data, seamless automation, and a cohesive strategy, Carwow is primed to continue testing new strategies, scale effective campaigns, and maintain their leadership in a competitive marketplace. With their improved clarity and efficiency, the team is hungry for continued growth and innovation; ready to capitalise on the opportunities they’ve created to continue navigating future challenges and opportunities.