Will Goodhand, Innovation Director at Kantar TNS, looks at what programmatic can do to move away from the relentless targeting that ultimately only alienates consumers, and become better targeted and tailored messaging welcomed by its audience.

Of all the developments we've seen over the past year or so, perhaps least surprising has been the continued rise of programmatic advertising. Some $39 billion was spent globally - expected to rise to $64 billion in 2018. 'No wonder', you might think, given the prospect of the Holy Grail: tailored and highly resonant messaging, delivered in an automated fashion to its precise target audience.

Yet a number of indicators suggest programmatic's promise is yet to be fully realised, and until it is, growing numbers of consumers will continue to be alienated by consistently poor experiences. Thirty-four percent of the world's online population feel they are 'constantly followed' by brand advertising online, as revealed by Kantar TNS's global Connected Life study of online trends, behaviours and attitudes across fifty-eight markets. This is resulting in increased use of adblocking technology, now standing at 27% in the UK and 25% in the US.

Thirty-four per cent of the world's online population feel they are 'constantly followed' by brand advertising online

This article looks at the problematic side of programmatic, then explores some solutions to the issues raised, using real-life and award-winning case studies - including tailoring messaging to different audiences in order to achieve campaign objectives 

As we were picking up considerable disaffection with today's digital advertising, we decided to conduct some qualitative work to dig into what was driving this. Even among Leaders (the most socially connected and digitally savvy segment that we identify in Connected Life), we saw considerable disaffection with programmatic. Much consumer disgruntlement tends to centre upon what Caspar Schlickum, EMEA CEO of Xaxis, condemns as 'irresponsible retargeting' - essentially, the phenomenon of serving advertising back at people from websites and for items/offers where they have shown interest (through an initial visit or click). 'Relentless retargeting' might be just as good a term, reflecting as it does that the nature of the advertising is not the sole aggravation but rather that it seems ever-present and ever-persistent. After the sixth viewing of such digital ads, brand impact actually turns negative - not just wasted impressions, literally damaging ones.

The Data Exchange Economy was one of the key trends we identified in Connected Life 2016: the notion that people are happy to trade their personal data provided they are receiving value in return. Fitness apps which monitor and facilitate refinement of personal exercise regimes are a demonstrably successful example of this trade-off but by no means the only one.

But can this philosophy extend to advertising? Some respondents felt there were a few brands doing it right, proving that it is possible to harness data and target in ways that will drive better and more fruitful engagement with consumers than ever before. So how can we achieve this?

The solution to much of the challenge lies in the data management platforms (DMPs) which now exist in various forms - essentially agglomerations of the data from the cookies capturing people's behaviour as they use the internet. This is where the rubber of all that talk about Big Data really hits the road. For a career market researcher like myself, this is an incredible boon. If I had a pound for every time I'd heard industry professionals talk about the need for research to achieve impact, then I could afford to buy products over and over again whenever irresponsibly retargeted ads implored me to do so.

Tweet thisThe world of DMPs opens up a whole field of opportunity for Mindset Targeting - true insight-driven programmatic. So long as, say, the panellists in online market research have consented for their cookie data to be harnessed, we have the scope to identify a target segment (or segments) in a survey and then to find matching cookie profiles on a much larger scale via lookalike modelling in the DMP. These lookalikes can then be reached by programmatic advertising - that is, the ad is only served on recognising that a user landing on the page has the qualifying cookie profile. The next step is realised by applying the insights from the diagnostics in the survey to shape the message which features in the advertising: the Holy Grail of targeted advertising delivering a tailored and personally resonant message, giving you something in return for your data.

Fascinatingly and excitingly, surveys need not be the input to such a process. Take passive mobile metering for example, as currently practised at scale by Kantar in the USA and UK, with capability rapidly growing in other countries. This ability to access the realities of people's behaviour, rather than what they merely claim, has already revealed remarkable statistics, like the fact that US smartphone users on average unlock their phones eighty-three times a day (the Leader segment 120 times or more).

Yet the capability can also form the basis of lookalike modelling in a similar way to survey. How so? Let's say we are able to identify the profile of people who respond to an advertisement served to them on their mobile device; we also know the moment in time that they responded, their approximate location and other data points, all of which enable us to refine targeting. We can even use our group of ad responders' cookies to find lookalikes in the DMP, profiled to be likely to respond to the very messaging we will be serving to the lookalike audience. At Kantar TNS, we are working on stepping away from devices altogether, and experimenting with identifying nano-segments on social media platforms and using these as the input for lookalike modelling and subsequent targeting. The scope can only grow, since even data from devices connected to the Internet of Things should be able to form the basis of lookalike modelling.

So, we're all agreed these are exciting times. But what of the real-world applications so far?

Our first real-life validated case comes from the US, with a Mindset Targeting approach deployed by Kantar TNS, working with the KBM Group's DMP, for Holiday Inn. A survey was conducted, using TNS's Conversion Model approach to identify a segment of consumers who were positive to the Holiday Inn brand yet who for various reasons did not manifest their positivity in their actual behaviour (perhaps they bought another brand more often, had a broad repertoire, etc.). This segment was modelled in the KBM DMP and the lookalikes were targeted at scale with programmatic digital advertising. In order to independently assess the impact of this campaign, Millward Brown was brought in by the client to carry out a validation study. In comparison to a control cell matched in every regard except exposure to the Mindset Targeted advertising, the exposed cell showed a 514% uplift in Holiday Inn site sales. Clearly, this is a remarkable number.

The world of DMPs opens up a whole field of opportunity for Mindset Targeting - true insight-driven programmatic.To understand how such a thing can be achieved, we need to go back to the underpinnings of the philosophy. We need to remember how so much programmatic advertising works. How much value does an advertiser receive for its dollars if an advertisement for a flight or a hotel room is served repeatedly to someone who has just visited the site in question and already transacted on the offer being advertised? Data from our Connected Life study tells us something else too: globally, people tell us that if they are in a pressured or 'routine purchase' state of mind online then they tend to have already made up their mind which brand they are going to purchase (65% for Pressured, 59% for Routine). In an exploration state of mind, however, only 33% have made up their mind. By targeting a Mindset profile, we are not just reaching the right people more often, we are also maximising the chances of reaching them when they are in an Exploration frame of mind. Thus, we start to understand how a Mindset Targeting approach can deliver such impact.

We wanted to demonstrate the effectiveness of the Mindset Targeting approach in the UK and explore how tailored messaging can be delivered programmatically to effectively achieve campaign objectives. Accordingly, I reached out to colleagues at Lightspeed and at Xaxis (the data agency which has its own DMP, namely Turbine). Each of us committed to work pro bono to create an insight-driven programmatic campaign for the Rwandan genocide survivors' charity SURF Survivors Fund. Kantar TNS developed a survey which really explored people's attitudes in the broadest sense - certainly their disposition to charitable giving and the nature of that, but also looking at attitudes to the world and global issues - in order to give us plenty of building blocks for a responsive target audience.

It's worth noting that the SURF Survivors Fund had just launched a new website selling beautiful jewellery handmade by widows (www.rwandanbeauty.com), and so the campaign's focus was to promote the website, which also provides visitors the opportunity to make donations. Hence, we also included questions in the survey about behaviour and attitudes towards 'good gifting'.

From the results of the survey, Kantar TNS developed a segmentation, and then using the cookie profiles of the segments offering greatest promise, Xaxis was able to conduct lookalike modelling to create target audiences. Validations were conducted by GMI to verify the accuracy of audience targeting.

One of the most actionable findings was that, of the various possible inputs, it was the lookalikes of people expressing a positive brand-level response to SURF Survivors Fund who were most responsive to the eventual digital advertising; in other words, targeting an equivalent Mindset to that in our Holiday Inn work. Of course, the beauty of the Mindset Targeting approach is that any of the survey questions can be used as the basis for lookalike modelling: for a new product launch we can model against those who 'definitely would buy' the product (to ensure the level of claimed commitment is actually delivered in market); we can even isolate those responding positively in a test of the creative and model against that group. Indeed, this latter audience was overall third strongest in terms of in-market response after SURF Brand Positives, and a segment open to Overseas Charities (even though they had not specifically been SURF-positive).

To fully optimise the tailoring of messaging to target, two initial creative concepts were built: ads showing the handcrafted jewellery and ads showing pictures representing the charity and the people it helps.

Responses to the ads in the initial survey enabled refined targeting so that each segment would be served ads to which they would be most responsive. To this, Xaxis added Smart Creative technology, meaning these creatives were dynamically optimised based on in-market responses to the campaign once live. In total, there were 1,664 variations of the original creatives, each optimised to generate the highest engagement with specific segments.

Highly refined programmatic technology is particularly good for charitable causes because efficiency is crucial for these campaigns to be successful, as the resources are often limited. NGOs and charities just cannot afford to waste resources in reaching users who will not engage at all with their cause. Although in this case all nine million impressions were donated, the precise targeting made a huge success of the campaign and the personalisation of the creatives drove engagement.

The results of the campaign were impressive - so much so that our team won The Drum magazine's award for Best Digital Strategy or Campaign 2016. The use of Mindset Targeting, along with Xaxis' Smart Creative, increased brand awareness by 66% against regular display advertising. Similarly, likelihood of donations increased by 29% against regular ads and we also observed a 33% increase in the likelihood to purchase the jewellery among the users who saw the optimised creative.

So, what to conclude from what we have discussed? All the indications are that programmatic's influence will continue to increase. What is more, as our case studies have shown, clear solutions exist to the challenges which programmatic presents. The future does not have to be one of brands alienating their consumers through relentless irresponsible retargeting.Tweet thisHarnessing the latest technologies in a smart way points to unlocking programmatic's true promise, delivering accurately targeted and tailored messaging welcomed by its target audience.


Will Goodhand is Innovation director at Kantar TNS. He drives clients' growth through the moments that matter most: employing social media analytics, insight-driven programmatic and other new solutions.