Will AI be the saviour of multi-touch attribution?
- Thursday, March 14th, 2019
- Share this article:
In light of Apple’s recent move to cap the lifespan of first-party cookies on Safari, Carl Erik Kjærsgaard, CEO and founder of Blackwood Seven, looks at the alternatives to cookie-based tracking.
Apple recently announced plans to cap the lifespan of first-party cookies on Safari to a maximum of seven days, leaving global marketers wondering if this is the cookie’s final crumble and if a post-cookie world is on the horizon.
The introduction of ITP 2.0 (Intelligent Tracking Prevention) last September at WWDC was the first step – an upgrade which meant that the tech giant could cap all ‘third party cookies’ across Safari browsers. Apple’s privacy enhancement drive did not stop here. The roll out of a stricter upgrade (ITP 2.1) announced in February is creating a big headache for the marketing community.
In an industry obsessed with capturing user data for delivering tailored messages, the anti-tracking update has left the marketing community wondering if multi-touch attribution (MTA) is still an appropriate tool for the measurement of campaigns. The upgrade also puts a big question mark on the survival of publishers and brand networks who rely on cookies for tracking visitor behaviour and for other marketing purposes.
Alternative measurement model
Perhaps it’s time for marketers to change their approach and ditch cookie-led tracking. If you listen closely, global marketers are clamouring for an alternative measurement model, something that can produce genuine marketing results with maximum precision. Although MTA has some strengths and is far more effective than last click attribution, it isn’t really fit for purpose, as it doesn’t account for external factors such as the impact of seasonality. Plus, it’s far too slow to react to real world events.
The secret to building a comprehensive marketing experience is to eliminate uncertainty and create market intelligence in as close to real time as possible. While the attribution model has allowed marketers to capture data during every single consumer interaction, it has also created a ‘data paradox’, where we now have both too much and too little data.
AI is no longer confined to the realms of cinematic fantasy, and although we certainly aren’t yet in a world where the plot of ‘The Matrix’ is possible, a lot has changed in the last 20 years since the film was released. AI has progressed to become a reality in modern businesses, helping them automate their activity and personalise the products and services they offer. Take a look at Sky, which has implemented a machine-learning model to recommend content according to the viewer’s mood. This is the level of personalisation that AI can bring with it in business. Even search engines are starting to take a ‘predictive and contextual’ approach, serving content that is of interest to the user.
For brands that are having to make decisions, in an increasingly complex environment, AI allows marketers to understand exactly what consumers are thinking, saying, and feeling about a brand – all in real time. It can also help businesses to tap into the consumer data hidden in keyword searches, social profiles, and other online data, for smarter and more effective digital ads.
Real-time marketing intelligence
This poses a conundrum for marketers, as the promise of AI suggests that it could solve the problem of getting real-time marketing intelligence. But, due to the data paradox outlined previously, most marketers simply do not have enough data to inform the AI, and in some cases, such as with new media channels, never will. This is where the combination of human intelligence with machine automation can play a role.
Arguably, the role of human knowledge and intelligence has been left behind in marketing’s rush to embrace data, but it should not be discounted. Instead, businesses need to understand that the expertise of marketers should be used in combination with data. Increasingly, we will see solutions that marry these two sources of information to give marketers a full picture.
With more being demanded from marketing departments in terms of accountability, tools which accurately measure the effectiveness and ROI of marketing efforts are in increased demand. This pressure, alongside GDPR, means that marketers need to find tools capable of fixing this conundrum. As we approach the end of cookie-based tracking, ignoring AI because you don’t have enough data to inform the machine is no longer a valid strategy. Instead, slow but steady moves towards an AI-first approach that combines AI and human intelligence will allow brands and marketers to prosper.