How marketers are navigating complexities of decoding consumers' digital footprints

In the maze of modern marketing, understanding the twists and turns of consumer behaviour and journey is like navigating a bustling city without a map. Every click, scroll, and purchase leaves a trail, but piecing together the full picture remains a challenge.

Marketers wield magnifying glasses to decode consumers’ digital footprints, unveiling the secret patterns that guide their shopping whims. It's a quest akin to solving a mystery, with each clue shedding light on the journey from click to purchase.

In a panel discussion titled "Efficient Attribution Systems to Decipher Consumer Behaviour and Journey Touchpoints" at Ad:tech 2024, Ashish Tiwari, CMO of Home Credit, Vipul Kedia, Chief Data and Platforms Officer at Affle, Kapil Bonde, Head of Insights at Insea, Adjust, and Chandan Mukherji, Director and Senior VP of Strategy, Marketing, and Communication, Nestle, discussed the challenges associated with constructing a 360-degree view of the customer.

They also delved into the latest attribution systems aimed at optimising consumer behaviour and improving efficiency in media buying, among other topics.

Tiwari asked the panellists to provide an example of when they have effectively delivered attribution to business designs and another where it wasn't as successful.

Responding to his question, Mukherji said, “We operate a coffee machine business globally, encompassing brands like Nespresso, Nescafe, and Dolce Gusto. Our offerings include machine sales, pod sales, and more. We have a deep understanding of our consumers, backed by substantial first-party data. Through the consumer journey, from initial awareness to loyalty building, we employ a comprehensive customer database to track and analyse their behaviour.”

“However, when it comes to the fast-moving consumer goods division, which constitutes the majority of our business, our understanding is limited. This lack of insight makes it challenging to accurately attribute value or credit to specific inputs at various stages of the consumer journey. While digital marketing holds the promise of data-driven insights, our attribution system remains slow and fragmented due to disparate data silos. As a result, we only obtain fragmented insights rather than a holistic understanding of our operations,” he added.

Furthermore, Mukherji emphasised that marketers still have a long way to go for attribution to truly benefit them and add significant value. Additionally, digital marketing also has room for growth to achieve its next significant advancement.

Moving on, Bonde said that for them, initially, the primary metric was driving installations. Subsequently, with the evolution of Mobile Marketing Technology (MMT), the emphasis shifted to Cost Per Action (CPA), tracking outcomes based on specific conversion Key Performance Indicators (KPIs) such as KYC conversions. 

“This transition has been facilitated by last touch attribution. While last touch has enabled app marketers to scale effectively over the past 15 years, it also exhibits some limitations. For instance, during performance evaluations, relying solely on last touch attribution would mean evaluating based solely on, say, the previous month's performance. However, this approach is flawed as it neglects the contributions from previous months,” Bonde said.

“As a company, we assist app marketers in comprehending various touchpoints and gaining a comprehensive understanding of how different channels interact. This allows them to view the complete picture rather than relying solely on the last touch. With this evolution, app marketers are also learning and devising various attribution models to better interpret data. Overall, we observe an industry-wide evolution where individuals are gaining better insights from attribution data,” he added.

Adding on to that, Kedia emphasised that attribution holds as much weight in the industry as media or marketing channels do, as it ultimately determines our revenue generation. Last touch attribution simplifies the process but overlooks the diverse roles various channels play throughout the marketing funnel.

“So, essentially, what last touch attribution does is give credit to the person who received the last click from the user. However, it only attributes the conversion to where the user last clicked, which is problematic because throughout the user's journey, there could have been various platforms and touchpoints across different channels, both online and offline, that contributed,” Kedia said.  

Furthermore, he stated that unfortunately, the mechanisms to credit each channel adequately are lacking in our industry. This deficiency often leads to instances of ad fraud, as there's a rush to claim the last touch of attribution. This issue persists across both programmatic and non-programmatic channels, with entities resorting to fake clicks and impressions just to secure that final attribution. This poses a significant challenge in our industry today.

On being asked about the favourite attribution model, Mukherji said that they are interested in exploring multi-touch attribution with a broader model approach, but there seem to be several shortcomings. The potential of multi-touch attribution rates is about 8 or 9 out of 10, but the actual delivery falls more around 3 or 4. Similarly, Kedia also suggested multi-touch attribution and cross-screen attribution.

Adding on to it, Bonde said, “The selection of the attribution model varies from one business to another. For instance, while some businesses prioritise the first touchpoint, others may consider the middle touchpoint crucial. However, this choice is highly dependent on the specific needs of each business. Personally, I still favour the last touch attribution model, as it has been successful for the industry over the past 15 years.”

After analysing the responses, Tiwari stated that our wishlist is a multi-touch and more complex model. However, the reality is a single touch and probably a couple of additional features here and there.

He also asked the panellists, “In the age of data privacy, how do you think these attribution models will evolve? How does data privacy impact them, and how do marketers tackle the challenge of data privacy along with attribution?”

Responding to the question, Kedia said that data privacy will play a crucial role in shaping the evolution of attribution models.

“We have witnessed this shift within the Apple ecosystem, where Apple transitioned to their own internal attribution method, ensuring compliance with data privacy regulations. Similar changes are occurring within the Android and web ecosystems,” Kedia said.

“This transition entails a move from deterministic attribution models to more probabilistic ones. However, it raises the challenge of accurately tracking multi-touch interactions and attributing them to final conversions. Thus, the primary change observed is the shift from deterministic to probabilistic attribution models,” he added.

Bonde highlighted that advertising and privacy are fundamentally at odds with each other. However, there is a way to navigate this dichotomy while still adhering to privacy standards.

This is where Apple made a significant contribution to the privacy movement by introducing the App Tracking Transparency (ATT) framework. For those unfamiliar, the ATT framework operates within iOS. When installing an app, users are presented with a prompt asking if they consent to allowing the app to track their usage. Users then have the option to accept or decline this permission, he explained.

Furthermore, he mentioned that in the gaming space, approximately 50% of people consent to having their usage tracked. However, in non-gaming contexts, this percentage varies. The optimal consent rate is around 30%, but it can drop as low as 3%. This indicates a significant lack of trust among users who choose not to give consent.

"So, how do you still measure them? That's where the overall measurement time frame is actually changing. It's transitioning from measuring user-level data or device-level data to a marketer's ability to measure based on aggregate-level data. Now, for those users who have not given you consent, how do you measure that? That's where Apple has introduced an aggregated framework for measuring those audiences,” Bonde said.

“Even Adjust is moving in this direction. Imagine if users don't provide any consent tomorrow. How can app marketers still track them? Hence, we are introducing a new product called Incrementality and Marketplace Modeling. Incrementality simplifies tracking by providing three metrics to determine if a campaign delivered incremental impact, the same impact, or affected organics, even with only aggregate data. This product helps understand campaign impact and optimise accordingly, even without user or device-level data,” he added.

He also mentioned that, as an organisation, they collect vast amounts of data, including first-party, third-party, and zero-party data. However, the challenge lies in utilising this data effectively. Hence, they have introduced Media Mix Modelling, an algorithm that analyses first-party datasets to offer meaningful recommendations.

“By running data through this product, you can determine the optimal allocation for future spending across channels, predicting the impact of each dollar spent,” Bonde added.

Moving on, Mukherji said that there is still significant value in aggregated data. It hasn't disappeared entirely. Aggregated data is valuable for marketing mix modelling and media mixes, providing a holistic view from end to end.

“However, privacy and content are crucial; without them, we've seen various issues arise. Marketers and platforms must earn trust by respecting privacy and ensuring data is used appropriately. Personalisation, when done right, can enhance the user experience and add value. Unfortunately, much data is misused, leading to spam, fraud, and other problems,” Mukherji said.

“Therefore, consumers will resist until institutional trust increases, which is challenging to achieve. The industry must consider this issue. Without trust, obtaining accurate data and effectively serving consumers is difficult. Thus, trust is crucial for proper data utilisation and consumer service. Moving towards aggregate models and integrating various data sources, including digital platforms and offline data, is valuable. Overemphasising last-click attribution distorts the picture and undervalues certain aspects. As these occur in separate silos, connecting them is essential to grasping their full value and improving consumer satisfaction, which is our goal,” he added.

Info@BestMediaInfo.com