While this type of customer segmentation was – and remains – important for engagements across traditional above-the-line engagements in mass media, digital marketing gives us the tools we need to target customers on a far more granular and personalised level. Where customer research gives us an indication of who the audience is, data can tell us exactly what they want and how they may behave.
The vast data captured through digital channels – crunched by powerful algorithms and machine learning tools – gives marketers the ability to understand and predict people’s behaviour more accurately and in more detail than any focus group or consumer research could ever allow.
These algorithms are constantly learning about people based on their actual behaviour and deciding what messages and ads they should see and how often. This approach is based on hard data and real-time engagement, rather than on a generalised set of assumptions derived from research that could have been conducted months or even years ago.
The machine can predict your online behaviour better than your friends or family
With deep learning and artificial intelligence (AI) at our fingertips, do we really need to define who our target customer is and what their likes and interests are? Based on the data, the machine will be able to give us a far more accurate answer about what the person is looking for at a given moment in time.
Take the example of a consumer researching options for a holiday, including flight and accommodation. After the Google search or visiting tagged travel sites, they will start to see banners and social media ads for deals to the destinations they were researching. A campaign manager might have a clear idea about who should be targeted: young, unmarried high-net income individuals, perhaps. But why exclude serious buyers who don’t fit this narrow segment?
Netflix is a master of talking to customers according to their behaviour and preferences. Unlike traditional pay TV providers, Netflix doesn’t segment its audience and offer channels tailored to different user profiles. Instead, it uses deep learning to decide which series and movies to feature for an individual user based on their viewing patterns.
Deep learning and AI do the same for online ads. Social media channels, search and programmatic display platforms, coupled with dynamic ad formats, all have the functionality to allow marketers to ‘atomise’ their audiences. They can target ads to an audience of one, customised to their exact need at that precise moment in time.
The Google Display Network, for example, combines three optimisation technologies to automatically:
- Determine the most appropriate bid price to achieve the target cost per acquisition, based on the likelihood a person will convert;
- Show ads to the person most likely to respond, based on their historical online activity; and
- Present dynamic content tailored to an individual’s interests, with headlines and images customised to the user.
The Google Display Network offers the option to switch to manual bid adjustment if the automated platform can’t achieve the targeted cost per acquisition (CPA).
Now is the time to look at atoms, rather than segments
Despite the maturity and power of this technology, we still see many experienced marketers applying traditional media tactics in a digital world. But to restrict campaigns to predefined target audience segments is to lose the real value of digital media. The true power of digital marketing lies in using data and algorithms to prospect for customers based on their real behaviour, rather than based on a simplistic persona that may exclude some serious customers from seeing your ads.
Once a person who has viewed the ad engages with a prospect-focused campaign via a social media engagement, video view, website visit, or lead form, we can treat their information as first-party data. Under South African data regulations, we can continually market to these prospects as they move through the sales funnel from awareness to conversion to loyal, repeat customers.
As a side note, understanding the target customer remains important for your communications strategy, especially when it comes to brand-building campaigns. You need to define your audience to craft relevant messages, but once you have done so, you can use data and analytics to reach the right people with the right ad at the right time.
For example, you can craft a range of ad messages and use dynamic ads in performance-focused campaigns that aim for conversions. Brand-building campaigns will more closely resemble traditional mass media campaigns and marketers should still focus in reaching the ideal customer in the right place with a message that will resonate for them.
Let’s wrap up with a few tips about how you can get started on using behavioural data to engage better with your prospects and customers:
- Clearly define the steps to a final conversion and ensure appropriate tags are in place to track each step and to inform the algorithm for automatic optimisation at each step.
- Don’t predefine your audience – rather set your target goal and trust the machine to use the tag data to speak to the right people at the right time and in the right context.
- Create as many custom audiences as possible using segmented databases, specific page visitors, social media engagements, video views, etc.
- If there is not enough data or if traffic volume is not sufficient to inform the algorithm to optimise for your end goal (whether that is a lead or a sale), then set the campaign to optimise for earlier steps in the funnel such as “add to basket” or even a “landing page view”.
- Boosting posts from a Facebook page does not use the platform’s full capability. Create an ad account linked to business manager, which allows for effective conversion optimised campaigns across Facebook, Instagram and their audience network on third-party sites.