The ‘holy grail’ of marketing is to get personalisation through effective use of data, and understanding how to do that in a way that has privacy and consumer experience at its heart is one of the biggest challenges marketers face.
In a year that’s accelerated the speed of change like no other, collecting, reading and understanding the signals that allow businesses to make better decisions is critical. For many brands, data has the power to be a competitive advantage, particularly when the opportunity is connected to strategic and creative thinking.
To get under the hood of this important challenge, WPP set four of its APAC agency leaders the task of presenting their view on the future of data in the region at an event this week.
The Drum has distilled the top takeaways from the event but you can see the session as catch up.
How AI and machine learning can improve creative and experience
Thomas Brauch, Chief Data Officer, Wunderman Thompson
According to Brauch, one of the biggest opportunities to improve creativity is through the use of data to realign how and what we are measuring. He believes that we’ve been focusing on engagement instead of reaction and by switching to understanding the ‘why’ instead of the ‘who and how’ brands will unlock more effective creative.
Brauch says the true measure of creative efficacy may not be number of clicks or views but how it made consumers feel, the level and type of emotive response.
To do this brands should be looking at neuromarketing and Vision AI as tools to optimise creativity and strengthen a brand’s ability to make human connections.
The nature of these tools and the insights enable agencies and brands to develop creative work that is more likely to capture and carry consumers attention and leave a longer lasting and positive impression. Studies have shown that ads with an above average emotional response from consumers caused a 23% sales increase as opposed to average ads.
In terms of whether there are biases in AI and ML Vision, Brauch says, “yes and no, there are cases where we see differences, for example, in the level of emotive response. You could compare the reaction from people in Italy and South Korea and see baseline differences of emotive reactions. The AI does a good job either way of identifying the level of emotive reaction and the type of emotive encoding around the reactions. So there's no structural bias but cultural differences that we always need to be aware. I would watch out for potential gaps in image annotation and logo or brand detection. The AI tends to do better with Western brands, so that’s one to watch for, a little bias towards identify logos for Western brands.
How AI is powering business outcomes
Deepika Nikhilender, senior vice president, Xaxis APAC
While AI had been a much-hyped new topic for marketing for some time, Xaxis’ Nikhilender believes that the technology is now synonymous with marketing, with many marketers reliant on it for executing digital marketing strategies. Indeed, she says some 76% of marketers say they use data and analytics to drive key decisions.
However, the key to truly driving more value from data and AI, according to Nikhilender, is to make sure brands are optimising to the right outcomes and challenges. She says not all brands can achieve their unique business challenges with standard or similar outcomes, “it is key to customize outcomes for your specific challenge and data set”.
She used the example of Xaxis’ Copilot tool, which uses AI & Machine learning to deliver the brands’ custom outcomes via programmatic strategies. She explained that it addresses strategies much faster than humans and creates 10X more granular strategies than traders, allowing you to leverage the data volume.
When asked if Copilot replaces a brand’s DSP, she says, “Copilot doesn’t replace DSPs, in fact, there’s no conflict relationship between the two and it’s actually complimentary. This is because typically DSPs work as platforms that bid at the inventory level, in order to deliver an ad to the audience. While doing so they are built to tackle the most common denominators of the problem itself, they never provide anything bespoke or can seldom be set up to provide anything bespoke for the problem or type of outcome that a client wants and that’s purely because that’s not scalable for them and not their business model.”
Ultimately, however, Nikhilender believes that AI will not replace humans. Instead she says humans and AI go hand-in-hand, and they need to learn how to tango together.
How to manage addressable audiences in a cookie-less world
Sudipta Roy, chief transformation officer, Mindshare APAC; chairman WPP Team Unilever- Asia, Africa, Russia
Firefox, Apple and Google have announced the depreciation of the third-party cookie, with Google Chrome being the last to act on this with its deadline of 2022. According to Mindshare’s Roy, this means we are heading into one of the most disruptive and dynamic phases the ad tech and mar tech industries have ever seen.
The end of the cookie-era has an impact on the very fabric of the internet today, he says. The web works on the back of an invisible value exchange mechanism and cookies were fundamental to that value exchange. Cookies helped to connect the right people to the right content, in a way which allowed billions of websites to receive ads and therefore help consumers enjoy non-subscribed content without consumers paying for every article. In essence consumers exchange data for content.
Advertisers pay for the ads on websites using this data. Websites use ad revenue to cover part of all of their costs. The world is searching for a solution to replace this model – where the value exchange is preserved without sacrificing consumer privacy.
All is not lost, however, as Roy believes that brands that prepare will come out strong. He referenced industry attempts to address this issue, such as the building of universal ID using email single-sign-ons that will be used as open-source answers to identifying consumers anonymously within a single view. Likewise, the walled gardens of the internet, which own a lot of data on consumer’s identities will have a stronger position. Among some of the more radical solutions being designed , are federated-learning systems which use machine learning in the browsers themselves which help create consumer cohorts without any private data leaving a consumer’s device. Some large clients are also exploring the route of privacy sandboxes with the likes of Google, which Roy believes could give an alternative to “willing partners” but also create a less competitive, less healthy marketplace.
He adds, “Our expectation is that almost every client that has to reach out to wide sections of audiences will get affected. If you are a client with a very specific customer set e.g. a B2B marketer targeting CTOs and CMOs, you don’t depend so much on the third party audience data and you won’t get affected much but if you are a client who depends on reaching a wide set of audiences in the upper funnel, to generate reach and awareness, you will be impacted. This is because your ability to describe and find those audiences by those descriptions, which the cookies and the device IDs provide, will be impacted.”
In terms of the steps brands should be taking, Roy says the first step is reviewing your identity partners, strengthening their first-party data strategies and really focussing on rewiring some of the marketing models to drive richer value exchanges.
How brands can use AI and content to unlock growth
Rina Simon, content lead, GroupM Digital Indonesia
According to GroupM’s Simon, content data is the missing piece that marketers must invest in amidst the hunger for big data in the industry. She believes the use of content data platforms cuts across marketing business functions, allowing brands to make smarter business decisions for their marketing.
In terms of how it works, she says data processing of content allows us to identify which content really matters and is impactful in driving results for your brands.
To really understand the power of content data, she talks about the content genome, which is the multi-layered intelligence that can be gleaned from content. The genome consists of four pillars of data, including profile data, audience data, content data and outcome data.
She says, if brands decode the content genome and invest in a content data platform, it allows us to unleash the full potential of a piece of content. It also paves the way for applying AI and automation to the content process, by using automated content planning, such as GroupM’s Content Trainer. Similarly, the data can be used to match brands with content creators, to optimise and create more content that will generate the right business outcome.
Linking to the topic of cookie depreciation, she says this adds even more reason for brands to invest in this technology.
“One of the biggest shifts is that we are moving into a cookie-less world so never has it been more important for brands to invest in their own first party data. Secondly, given Covid-19 and what we are facing, we foresee a content overdrive, so the only way for us to make sense and drive reason when it comes to content measurement and effectiveness is to build your own content data platform as a marketer. You are not looking at measuring it short-term and tactical but investing in the long-term,” says Simon.
This article was produced in partnership with WPP for the Future of Data in APAC event.