Indispensable to measure advertising effectiveness and to remunerate some digital players, attribution is a traditionally cookie-based approach. While cross-device / omnichannel travel and reconciling a user’s multiple identities was already a challenge in this area, the demise of third-party cookies, such as consent, is now greatly disrupting attribution and its traditional techniques. . What are the challenges, new rules and best practices for allocation in 2022?
Used in order to measure the impact of the different levers (contribution of new future buyers, purchase intentions and new conversions), the attribution allows to know the ROI or ROAS of each channel to better invest in the channels and in the customer journey. Other KPIs are also observed such as acquisition costs, CPA, contribution, web centric KPI (visits, transactions, CA) or media centric (impressions, CTR, CPC, CPM), as well as sales, increments (on and offline).
By allowing us to trace the different levers that influenced conversions, the customer journey analysis is at the center of the approach, and can also include the journey and offline sales, in the context of more in-depth analyzes aimed at taking into account the entire customer journey and channels (eg Marketing Mix Modeling, long-term multi-channel analysis to find out which channels contribute most to sales). At the same time, it also allows you to analyze the contribution of different channels.
The various existing attribution models are used on the one hand to obtain lessons (knowing how to attract new buyers, transform a visitor into a connoisseur, convert him and with what lever or media device), and achieve various objectives by giving greater importance to a particular channel or stage .
Among these models are: the model of the last click (conversion goals), the model first click (to advertise a product), the linear model (to see a user’s typical route or optimize the trip as a whole), depreciation and appreciation over timethe U-shaped position (gives as much importance to the first channel as to the intermediate channels) e the algorithmic model (data driven) which evolves over time based on the data collected.
The last click, is it really finished?
On the market, the end of the last click has been heralded for years. However, in reality, the last click model is still widely used within companies, and this is based on the need for granularity of the information needed by the various services.
“Today we continue to send last click reports to management because last click remains the model chosen by default on the market. However, in an ideal world, what we most recommend today is to get out of these single touch models to move to multitouch models more suited to the complexity of the Internet user’s behavior (multi-device, channels used, time to think). “ Hassen Hammeche, Web Analyst, Altesse (Les Georgettes di Altesse)
End of third-party cookies: a major obstacle to attribution
Attribution is probably one of the digital areas most affected by the end of third-party cookies. The attribution window in particular (period between the exposure of the Internet user to an ad (or click) and the moment it converted (post-view or post-click conversion), depends on the duration of the cookies. restrictions on third-party cookies began long before Google announced that third-party cookies would be blocked on Chrome by 2023. Apple had already implemented its restrictions by keeping third-party cookies for a period of only 24 hours (versus 30 days or more depending on the reflection period required to purchase the product).
Furthermore, in addition to making multi-device data consolidation more complex, the demise of third-party cookies is prompting players to review current attribution models, historically also based on cookies.
Consentless: the impact on the analysis of the user’s journey
Added to this is the application of the GDPR and the strengthening of the consent rules (CNIL) which represent a significant obstacle to monitoring the user’s path conditioned by the choices now available to the user regarding the tracking of his navigation: consent or minus (including both disallowed and unregistered traffic which is the worst case scenario for a marketer). In other words, the integral view of the customer journey may gradually disappear.
As in the context of ad blocking, this share of traffic using an unpaid medium results in a loss of revenue for both publishers and advertisers.
Marketing attribution, particularly affected by fraud
Bots, ad fraud and other bad practices … Being directly linked to remuneration issues, attribution / contribution is a very targeted area for fraud: last click fraud, cookie release, parameter hijacking, additions of impression pixels, etc. It represents a real deficit for digital advertising players.
Consent wall, login wall, cookieless ID: the future of the customer journey
Faced with these many obstacles today, marketers must find the best way to define new customer journeys that enable attribution to be sustainable. Logged on-site navigation is one of the solutions in particular.
“Publishers and advertisers are starting to think of solutions to convince users to access by subscribing to offers (example with Allociné:” access the site for € 2 including taxes without cookies or access the site for free by accepting cookies “), or through specific offers (eg discounts if you connect to the site). Today the access walls are developing and many publishers are evaluating the possibility of opting for attribution tools based on allowed and connected traffic, useful for maintaining part of the existing and attribution models ”. Pierre-André GAUTIER, Head of EU Strategic Partnerships at Liveramp
However, it is clear that unauthorized and unrecorded traffic will remain. A problem that could also favor technological innovations aimed at identifying people without cookies and being able to increase customer knowledge by exploiting the information provided through all touchpoints, including physical contact points (loyalty cards, e-tickets, cookieless universal IDs, etc. )
Pierre-André GAUTIER, Head of EU Strategic Partnerships at Liveramp: “Universal IDs (including the ATS) will allow you to keep the attribution tools but, however it happens, the web will be a 3-speed web: the non-consenting web (which will not be easily attributable), the consented contextual web and access authorized web. “
Attribution and data: deterministic methods and probabilistic methods
The various ecosystem transformations and data protection regulations have an impact on the volume and accuracy of the information collected and consequently an impact on marketing techniques. “We will have to get used to the fact that all attribution solutions and marketing mix modeling are based on probabilistic methods. Marketers will rely on extrapolated data (extrapolation methods: cross product, verification of the consensus rate according to the acquisition levers to adjust with compensation on this extrapolation method, etc .. “ Romain BAERT, CEO of Addingwell
At the same time, cookieless IDs or universal IDs will allow you to base attribution and marketing on deterministic methods.
Advertisers: first party and server side or how to regain control over tracking and data
The end of third-party cookies highlights two ecosystem problems: digital has been too much and for too long focused on the cookie, itself attached to the browser (client-side), while the advertiser has not sufficiently tampered with his tracking. By moving the tracking and solutions from the client side (browser side) to the server side (server side), the advertiser is able to regain control of their data in the first part by freeing themselves from the technical constraints imposed by browsers (Safari and ITP, Chrome ).
“An important technological change is taking place: we move from the cookie placed on the client side through a traditional tag management tool to the cookie placed through a server so that it is maintained and considered first party so as not to be blocked by ITP, Chrome or adblock , as it already happens today. ” Romain Baert, CEO, Addingwell
The emergence of data clean rooms
Data clean rooms provide a closed and secure environment in the cloud that can be leveraged for attribution. They represent a secure ecosystem for the transmission of anonymous data via a pseudonymous identifier. This allows you to associate multiple (multichannel) channels with different touchpoints and maintain attribution tools. However, the vision of the customer journey will be partial (end of mass attribution) and will consist in the analysis of groups, panels or cohorts (eg Privacy Sandbox).
Innovation in the service of attribution in the new world to come
Technologies will once again be at work to provide new solutions to maintain marketing attribution despite the end of third-party cookies and strengthening consent. Among these we find in particular:
- LiveRamp which provides a safe and authorized environment to work on the customer base and establish connections in the context of targeting and attribution without cookies, based on all offline touchpoints (under the prism of allowed and registered traffic)
- Addingwell which facilitates the installation and implementation of server-side tools, or server-side Google Tag Management
- Okube (Smart Traffik) which allows to measure the effectiveness of media investments (on and offline), to determine the attribution / contribution of each lever (through the algorithmic model) and to generate audience segments to help advertisers (retailers ) to be more efficient in their activations.
Between innovation efforts and model adaptation, the ecosystem appears to be able to support marketing attribution, while the arrival of web 3 and the metaverse could lead to the development of next-generation customer journeys.
File compiled by Stéphanie Silo
This topic was discussed on April 22, 2022 during the Topic of the Month webinar during the Ratecard Days webinar. The replay and a summary of the key points of the webinar are available here.
Next webinar Topic of the month
“Platformization of solutions: examples with programmatic and Retail media”, Friday 20 May 2022 on the occasion of the Ratecard Days. Inscriptions here.