Google’s Murtaza Lukmani is a leading expert on marketing analytics and attribution. He works with large companies across Europe, the Middle East, and Africa to help them improve their operations by updating their attribution strategies.
Everyone has those things they know they should do, but keep putting off — personal or professional tasks that are too big and too daunting. Sometimes, the crucial first step just isn’t clear.
For companies, the challenge is the same. In marketing departments, in particular, overhauling a brand’s outdated marketing attribution model is one of those tasks that many firms put off.
A large proportion of companies still rely on the last-click attribution model, even though people understand its major limitations in reflecting real-life buying behaviour. It can distort budget allocations and media-buying decisions, leading brands to place far too much budget on channels favoured by the last-click model, while overlooking other channels that can drive incremental results.
We’re now at an inflection point: privacy changes and our increasingly digital lifestyles mean it’s become a business necessity to overhaul our approach to attribution. Additionally, improved technology and tools make it simpler and more effective to adopt future-proof attribution models.
While various marketing attribution models exist, data-driven attribution can deliver results with precision that other methods lack. Data-driven attribution uses machine learning to decipher how different touchpoints throughout the customer purchase journey impact conversion outcomes, and the model distributes credit accordingly. The model has been available in Google Ads since 2014, but is now being incorporated into the new Google Analytics 4 platform as well. The model works across devices such as smartphones and tablets, and across channels such as Search, social, apps, Display and YouTube. Brands can also now use a relatively small sampling of conversion data to train the system: just 3,000 ad interactions and 300 conversions over the past month is enough to launch successfully. That’s a fraction of the data that was previously required, and Google is working to keep reducing these data requirements over the coming months.
Euro-Assurance: A new view of YouTube
French insurance broker Euro-Assurance paused spending on YouTube’s TrueView for action ads in late 2020, assuming Search was more efficient based on its last-click attribution model. The company wanted to focus only on the most effective channels to drive profitability.
But a Google Ads attribution report led to a complete marketing rethink. The report gave insights into the complex customer research and purchase journey for a typical Euro-Assurance client, and indicated that YouTube ads contributed to twice as many conversions as previously assumed.
“We never would have known about this with our last-click model,” said Vincent Couchellou, head of digital acquisition at Euro-Assurance. “The report opened up new opportunities to understand our customers and their purchasing behaviours.”
These insights led Euro-Assurance to switch from pausing YouTube spending to activating an always-on campaign in 2021 and shifting 2.5X more budget towards the video platform. The company is set to begin using a data-driven attribution model in the near future to further optimise their marketing and bidding strategies.
A range of brands across Europe, the Middle East, and Africa have recently begun successfully transitioning to the data-driven attribution model within the Google Ads and Google Analytics platforms. It can seem like a daunting process at first, but these brands are now reaping the rewards. Here are their stories and the lessons they learned along the way:
Geox: Making strides in social integration
Italian footwear firm Geox made great strides this year by bringing together fragmented data sets to gain a better understanding of how each marketing channel contributed to business growth. In particular, the company wanted to understand how paid social campaigns were contributing to sales, and then optimise.
The firm worked with agency Webranking to pull together vast amounts of its own raw social data and fed this into Google’s Search Ads 360 platform. The approach to consolidate data, and the subsequent move to use a data-driven attribution model, helped improve reporting and high-level bidding decisions.
Geox was able to see how paid social campaigns contributed to conversions across devices, and then it worked on a cross-promotion strategy: people that saw the paid social campaigns were also shown paid Search ads, and people who saw Search ads were then served paid social posts. This strategy led to a 6% increase in return on ad spend (ROAS). Plus, the streamlined system helped reduce the time spent on campaign management by 30%.
“The social integration provided us with a better picture of cross-channel marketing data impact and allowed us to open up new audiences across Search and social,” says Giulio Salvucci, global digital business and innovation director at Geox. “Thanks to this approach, we increased purchase loyalty and reactivation.”
Apo-rot and DocMorris: Finding the right words
German online pharmacy Apo-rot — which was recently acquired by DocMorris — switched from the last-click model to the data-driven attribution model to gain a holistic view of its customers’ purchase journeys. The ultimate goal was reaching new individuals and boosting sales. Apo-rot knew a large portion of its customers were from older generations, and they often used their smartphones to research health symptoms and then made their final purchases on desktop computers. These customers didn’t immediately seek out name-brand remedies as they conducted their research, but instead used generic health-related search terms.
Data-driven attribution offered a new way for Apo-rot to test Search keywords that had previously been considered inefficient from a last-click perspective. The brand deployed and evaluated 3,000 new, generic keywords, such as "ginger capsules", and found that placing ads against these search terms drove customer consideration and sales. Performance improved dramatically, with a 40% increase in mobile shopping conversions, and an 18% reduction in cost per order.
“The data-driven attribution model helped us see that many of our customers were researching on mobile, then buying on desktop. This helped us make the decision to reallocate budget to mobile,” said Lara Harter, head of online marketing for DocMorris. “We’ve also rolled out data-driven attribution to our other brands, DocMorris and Medpex, and we’ve been very pleased with the results.”
It’s time to stop putting things off until tomorrow. Brands that mobilise now will gain deeper insights into the consumer purchase journey and reach customers with the right messages at the right times. Plus, the data-driven attribution model is designed to help brands grow and evolve within a privacy-first digital ecosystem. With that in mind, here are three key tips to make the most of data-driven attribution:
1. Set clear goals
Decide on what you want to achieve before selecting a new attribution model. Ask yourself: are you aiming to reach new customers? Drive more sales? Improve your ROAS? Or re-evaluate the efficiency of a specific channel? Your team should set clear targets from the start.
2. Go beyond the reports
The data-driven attribution model provides insights and reports about the customer journey, but you can take it a step further. You can trial new keywords or channels in your campaigns, upgrade your bidding strategies, and assess effectiveness along the way. This attribution model is the base on which you can drive business results and incremental conversions.
3. Test, learn, and scale
Changes in approach don’t have to be all or nothing. You can start with pilot tests on a single brand, account, or market to prove value to internal stakeholders. Then, once everyone is comfortable, you can expand your approach knowing your team is behind you.