Balancing The Possibilities & Pitfalls of Personalised Marketing #DigitalSense

Individually personalised marketing has felt like the promised nirvana of digital advertising for at least the past decade, but marketers who’ve attempted to realise this dream are often floundered by costs, practicalities, creative constraints and data shortages. We need to be using data and targeting to be relevant to more people, not visible to fewer.

Whilst personalisation discussions tend to rapidly lean into first party data, advanced programmatic approaches and complex multi-creative development there’s a huge opportunity most brands leave on the table - to truly embrace context or better reflect DEI in their creative and start personalisation by talking more relevantly to these huge sub groups of consumers.

Setting the Scene for Personalisation

One of the earliest industry mentions of Personalisation at Scale is a 2015 Guardian article which boldly claims it ‘is the next big thing in digital’. 

Almost a decade on and I think it’s fair to say it hasn’t quite become the successful & universally adopted approach that the author anticipated - and I should know, after all I wrote it. The personalisation concept no doubt emerged from multiple directions but in one of its first iterations it came out of the joint work of Mondelēz, Facebook and Carat that I was leading at the time.

We had landed a global ‘Storytelling at Scale’ approach which pushed people to abandon engagement-chasing, small-scale social activity and embrace the power of such channels as mass scale mediums with meaningful budgets. Ultimately that felt a bit too blunt for such data-rich tools, and so exploring the right balance of personalisation and segmentation as a way of bringing more nuance to the mass reach channels of digital began. We called it Personal Connections at Scale to hold onto some of the idea that this personalised targeting would lead to deeper engagements and conversations, even if our evidence suggested this wasn’t notably a business driver.

Over the years there has been some growing doubt that these sort of personalised approaches can work at all - Byron Sharp has certainly questioned the entire endeavour, and I also recently wrote a joint article with P&G’s Gerry D’Angelo questioning how wise the constant pursuit of data & programmatic hyper targeting is.

This isn’t for lack of trying of course. Whether they call it personalisation at scale or just espouse the critical importance of first party data & targeting for their business, many big global companies have talked publicly about their ambitions in this space. That includes some of those I’ve had the pleasure of consulting on recently. Perhaps even more companies and media professionals have questioned these approaches behind the scenes, though usually with a heavy note of caution & secrecy.

Most definitely there are some meaningful success stories in this space, but there are also most certainly a long series of stalled efforts or projects which were doomed from the start. I thought I’d take a moment to consider how you can navigate the pitfalls of a personalisation strategy to untap the possibilities it can bring.

 Ultimately personalisation is a spectrum - everywhere from a single universal creative to a uniquely individualised message for every single consumer. Starting down that journey makes a lot of sense, but jumping to the extreme far end of it will be a stretch for most brands.

The Possibilities of Personalisation

Better Relevance & Impact

Most of the arguments for personalisation ultimately come back to the opportunity to have more cut through and impact if you can perfectly tailor your messaging to the audience receiving it. Think the classic programmatic slogan of ‘right person, right place, right time’ and you’re on the right grounds. Sometimes this can be nuance around copy or highlighted SKUs, or personalisation can tap more into unique audience qualities or passion points.

It’s almost certainly true that consumers will pay more attention to adverts that map to their interests - think of a football fan spending a few more seconds looking at your advert if it happens to feature the sport or one of their favourite players. It can be harder to believe that some of the more nuanced adaptations of end frames, copy etc resonate dramatically differently with different audiences, but sometimes they do seem to help shift numbers slightly.

Customer Development & Loyalty

There’s a specific subset of relevance which comes if you truly understand someone’s experience and purchase history with your products. In fact a large amount of personalisation seems to focus on acquiring the first party data to better understand this and build upon it.

Whilst penetration is the ultimate driver of brand growth there are certainly categories in which digging into this loyalty and looking to upsell make a decent amount of sense. There’s also an argument for exploring personalisation amongst the look-a-likes of these people to find other low hanging fruit - though personally I tend to be a bit cynical about how truly alike such expanded audiences are.

Growing into New Segments

Personalisation is at its heart a segmentation strategy - it’s about taking your overall audience and dividing it up into individual segments, possibly as small as one person, to better be able to communicate with them. A rich personalisation approach doesn’t have to just start with your current audience and divide it up however, it can look for new or underserved audiences that you haven’t truly been connecting with to date.

This could mean unusual passion points or use cases, but quite often the simplest example of this is through DEI media and personalising to unique cultural moments and occasions. This moves personalisation at scale from just being about a better contact with your existing audience and allows it to push you into new growth spaces.

Lower Waste

One of the overall claimed benefits of digital targeting in general, and personalisation within that, is that you eliminate the waste of reaching consumers outside of your perfect target audience. In some situations it’s obviously very important to specifically talk to the consumers who might be in the market for a specific product or service, that could be wildly irrelevant for others.

You do have to be careful that our defined marketing targets are often narrower than the real audiences that buy us, and that building brand recognition amongst people who might have no current interest in your category can pay back very well in the longer term. There’s also a practical caution that if you spend too much in narrowing down your audience you could have reached the same people, and many others too, through broader reach approaches.

The Pitfalls of Personalisation

From 0 to 60 in 2 seconds (Over Complicating Things)

A lot of brands have historically only had one mass version of their creative yet frequently, when they’ve started on this personalisation journey, they suddenly want to have 100s with micro differences. In most cases there you aren’t even doing personalisation at all; you’re just doing a large scale a/b test on the exact tweaks to your creative that might deliver slightly better impact or attention.

Before you go anywhere near that level of hyper targeting have you tried having 2,3,4 or even 5 different versions of your creative? Rather than throwing everything into a programmatic soup have you considered strategically identifying key occasions or audiences you could target. Do you have a different creative for football fans, or for Eurovision week, or for young mothers, or for men, or for different minorities? 

Some of this moves you closer to the world of ‘reactive’ and event social media work, but that level of creative thought can bring you much closer to real personalisation. If your goal is to be more personalised and relevant you are much more likely to achieve that by playing to people’s passion points, life moments or individual diversity than you are by trying subtly different end frames on them.

The Loss of Creativity

The above point is particularly important when we think about the creative itself, second only to scale/reach in driving marketing effectiveness. If you are trying to produce dozens or hundreds of different executions, you can be almost certain you’re heading down some sort of formulaic creative approach with subtle copy or design tweaks.

In many examples I’ve seen hyper personalised campaigns barely end up with a creative idea at all - it just becomes Campaign Slogan + Some Thing Think Might Be Relevant To You + Something Local + End Frame. It’s quite the gamble to assume that having an advert which is ever so slightly tailored to someone is going to cut through more than a brilliant generic advert designed for everyone, and it’s usually a bet you’ll lose.

Despite all the wonders of personalisation, good storytelling & clear communication trumps it every time. Of course, the nirvana we’re aiming for is both but that requires a big creative idea which has personalisation opportunities by design, and then a personalisation approach which leans into that. Again, you’ll be much more likely to be delivering a high quality & impact creative if you’re trying to adapt for half a dozen personalised situations than if you’re making hundreds.

Data Assumptions and Taking the Hardest Route

Most personalisation approaches start and finish with data. They look at what data you have and what you can acquire (through partnerships or data capture) and then build a targeting and creative strategy around them. 

This in many ways is like trying to target a football fan by hiring a private detective to look through their bins and see if there’s any evidence that they’ve recently bought any football merchandise or tickets. You’d probably find it a lot easier to identify football fans by standing outside a football ground or pub on game day.

Thus, whilst data can clearly play a role in personalisation, and especially if you get into the nuances of trying to personalise based on different customer types or consumption, it’s an odd and hard starting point for most brands. Context is not only a lot easier to identify, but it also tends to be cheaper to buy and there’s good reason to believe it will work harder as well. Having identified that football fan through your detective work, if you then use that data to target them in the middle of their working day you might still be irrelevant - better to show up, talking about football and your brand, right when they’re exploring that passion point.

The Cost of Over Targeting & Personalisation

I’ve hinted at this earlier but when you acknowledge that brand growth comes from driving reach/penetration, often to disengaged future consumers, you have to be a little bit cautious about any approach that starts narrowing down your reach to a very tight identified audience.

When done well personalisation can slice these broad audiences up and find relevant passion/category entry points to grab people’s attention and bring them in. When done badly you’re just using data to be visible to fewer people.

The absolute kicker that we don’t talk about in all of this nearly enough is the cost. Creating an individually personalised advert for every single person in the country and individually delivering it to them might be some brilliant ideal, but the costs of achieving that are clearly prohibitive. When you dig into personalisation even mid-scale schemes can fall apart at a true cost evaluation.

Have you considered the costs of creating different assets, of acquiring the data & audiences you need, and even of the programmatic machine itself where often only 30c in a dollar makes it through to a media owner. Your personalisation needs to work incredibly hard to push through all of that and still end up delivering better results than a good advert served to a broader mass audience.

In Hope of a Better, Personalised Marketing Future

Which isn’t to say it cannot be done. In fact 10 years later, I still believe personalisation is the future of digital marketing, and we’ve seen some sectors (perhaps most notably politicians) embrace it effectively. Generally well done right, especially in that example, it’s about telling distinct stories that resonate with different audiences, not about tiny programmatic tweaks to the colour of your background or the end frame.

If I was a brand starting on this journey, I’d probably identify 5-10 distinct growth audiences based on occasions, passions or consumer identity and make mini sets of campaign assets properly tailored to tell different stories & resonate with each. Then I’d look first for dedicated media channels or contextual opportunities that allow you to bring that message to that audience first.

Perhaps the richest place of all to start a personalisation at scale journey is around DEI - there are huge underserved communities of different ages, sexualities, abilities, ethnicities and more who would respond hugely well to adverts which better speak to and represent them. This is a massive open door with a broad range of existing partners available that many advertisers completely overlook as they jump straight to the most advanced programmatic approaches which attempt to personalise based on what they think people are for breakfast that day.

Alongside that I’d run some of the ads fairly broadly in mass media because it doesn’t hurt for unexpected people to see them, and if I had a really strong data set I’d layer a bit of that on as the icing on the cake. Social platforms with their inbuilt broad knowledge about consumer interests & identity also have a strong role to play in a media mix focussed on meaningful & impactful personalisation.

I suspect for most brands really hyper-targeted, programmatically driven, hundreds of creatives-led campaigns are something to consider only when you’ve nailed these first steps of meaningful, high quality personalisation. Unless you really are just talking about a sophisticated A/B test of content variations to drive better impact & attention which is a different story.

Certainly it could be possible to target a football loving, African American, lesbian with a passion for cookies and chocolate but I wouldn't jump straight to trying to find data to target that person before I'd exhausted richer apporaches around partnership, sponsorship & community engagement to get to that personalised storytelling.

If I worked in a very high consideration and value category, I’d focus a little more on building up that truly individual and first party data connection, bearing in mind that if you really want to sell luxury to rich bankers the FT’s HTSI is still right there in front of you. Consumer data is an important part of the future of marketing, but there are also consumer & legislative reasons for believing the future will see less of it used, not more.

 Jerry Daykin is a WFA Diversity Ambassador, Media Partnerships Consultant & Fractional Chief Strategy Officer at Adfidence. He has previously led global or regional media teams at brands including Diageo, Mondelēz, GSK Consumer Healthcare and Beam Suntory. In 2022, he literally wrote the book on “Inclusive Marketing.”