This is what real personalisation looks like

Increasingly, media personalisation is talked about as a silver bullet for many issues facing digital media today. It can deliver measurable rather than attributed performance and overcome issues with attribution models that preference high volume low-quality media buying - an issue that's also driving adblocking.

Many recent acquisitions show large global businesses reaching out to this new market. Verizon buying AOL, and Dentsu Aegis' acquisition of Merkle, for example. The former is an example of a company building its ability to target person-first media at scale, one of the prerequisites of personalisation; while the latter is a case of a media business obtaining a single customer view – essential to retain a personalised profile over time.

Both deals allow the companies involved to target individual consumers rather than the devices they own – a market trend started by businesses such as Conversant several years ago. The Merkle and AOL purchases are still recent, meaning these capabilities are some way off.

True one-to-one personalisation at scale

Few businesses can deliver true one-to-one personalisation at scale. This is why businesses talk about anything even remotely targeted or customised and claim it is personalisation.

To explain the difference, if you and I visited the same product page on a website, non-personalised but dynamic, segmented creative would show us the same product-led ads. There'd be no difference in creative or frequency, with that frequency being a rather high rate of ad delivery.

Using the same example, truly personalised media would show us a completely different set of creative, in a completely different order and frequency. Perhaps not even referencing the product, depending on our engagement on and offline with the brand, as well as our own individual media habits and consumption.

As personalisation is a key market that businesses are moving into, it’s important to be able to differentiate between aspiration and fact. True one-to-one marketing is impossible without the alignment of three key elements, but once these are present, it is possible to deliver measurable incremental growth to the top-line of global businesses. The three key elements are:

1. Brands must understand and speak to consumers as individuals. Not as cookies, devices or segments.

Recognition is the all-important first step in being able to deliver person-first media. This is the ability to identify an individual and all the devices assigned to that individual. It requires multiple inputs that may leverage, but are not limited to cookies. 

"Recognition is the all-important first step in being able to deliver person-first media"


Accurate person-first media requires a vendor to have access to deterministic data. In order of relevance, deterministic joins include: second-party email, publisher login data, consumer login and data on commerce such as credit cards.

The closer to a real person, the better the join. An email address generated by a competition is not as accurate as a join matched from a customer’s credit card, where, for example, a user may have many low-use email addresses but only one credit card.

2. Brands need to have real ongoing conversations with consumers to run true personalisation – the dialogue should be proactive and persistent over time. 

Data persistence is rarely raised by vendors of personalised marketing opportunities in discussions about the market, but it is absolutely vital to have an ongoing dialogue with a customer. Personalisation can’t occur if a brand can remember only the past 30 days of a conversation, or if a brand can recognise only the consumer on one of their devices. The alternative is a peculiar stop-start. Imagine you keep greeting old friends during the day and forget them completely 30 days later.

This is why brands need to leverage their own CRM data to have a chance of creating personalised experiences, but few have enough data to make personalisation worthwhile. Vendors must show what a customer is doing when they are not engaged with an advertiser’s website or can enrich their CRM-based insights with deterministic data of their own.

3. The success of a personalised media campaign should be measured against incremental return over time, because attributed conversions within a window are irrelevant to the lifetime brand engagement of a customer or prospect.

Up until now, digital media’s dependence on the cookie has led to prioritising direct response (DR) retargeted messaging, focusing on the last click or linear attribution models. But this results in high volumes of low quality, product-led DR creative being pushed out, delivered in short windows to maximise attributed return when a customer makes a purchase. There is widespread discussion as to whether this leads to a negative brand effect for the advertiser and the growth of adblocking.

In this instance, attribution is not equal to understanding the value being created, as it can be built up over many months. I am unaware of any attribution model that is capable of also deduping existing brand value.

"Truly personalised media needs to be tracked through AB testing and incremental measurement over time"

Truly personalised media needs to be tracked through AB testing and incremental measurement over time – which can only happen if profiles are persistent over time to keep control groups clean.

This model of measurement will give an exact figure for incremental return on investment over time. It will also offer a true measurement of the effect of personalisation that any marketer would feel comfortable putting in front of a senior leader.

Once measurement is focused on incremental return; frequency of advertising can be controlled at the level of the individual, and a creative mix, including brand, loyalty and DR can be delivered. This will improve the customer experience of engaging with a brand.

The net result of proper personalisation is better brand engagement, happier consumers and a greater return for advertisers over time.