Rules-based automation cannot handle the intricacies of the Big Content era: complex customer journeys, constantly evolving user profiles and myriad pieces of content that need to be categorised and structured before being served across multiple-platforms.
It’s time for marketing automation to get algorithmic…
Rules are useful for giving (and maintaining) structure and order in a messy world.
“Keep off the grass”
“Knock before entering”
“Do not park here”
Rules work well in rigid systems and societies where people are predictable and pliant. Yet we know from experience that life is complex and humans rarely act in rational or consistent ways.
We are all subject to a multiplicity of influences – both interior and exterior – at any given moment of the day which govern whether we obey or circumnavigate a rule. Sure, I might not park on that double-red line under normal circumstances but today I’m in a rush and I’ll only be in the shop for a few minutes so….
People rarely follow rules. Why do you expect them to follow yours?
The role of rules is particularly acute in marketing automation and campaign management tools – all of which rely on pre-defined business rules (“If this X happens then do Y”, “if X does not happen, then do Z”) to architect marketing campaigns.
Many marketing automation platforms still base their design and application on the purchasing funnel developed by E. St. Elmo Lewis in 1898. Lewis’s four stages of the purchase process have been used by marketing and sales folk alike for over a century: Awareness, Interest, Desire and Action. But the consumer buying processes have changed drastically since the purchase funnel’s inception. Consider the Internet, online reviews and social media and their impact on how we made purchasing decisions in the ’80s, ’90s and today. We know that in a ZMOT world, content increasingly influences an individual’s process and not necessarily the brand’s content alone. It’s safe to say the funnel is outdated. People don’t follow your campaign process, they follow their own!
DON’T BELIEVE THE HYPE – Marketing automation has not caught up
Marketing automation (particularly in a B2C environment) intrinsically struggles to recognise that people (and let’s take time to appreciate that these are people – with complex feelings, needs and interests – and not nondescript ‘leads’/’fans’) are not working their way through campaigns in isolation. Every customer’s journey with your brand is different, perhaps similar, but never the same as another’s path. Assuming a campaign can dictate how customers interact with your brand is shortsighted and dangerous.
Furthermore, marketing automation ignores that people are continually evolving in their interests, needs and motives. Whilst marketing automation can recognise if a person is stuck in a particular campaign phase – perhaps they have recently opened a second brand newsletter but did not click on the CTA button at the bottom, meaning that a business rule will say “resend the email again in three days” – your MA tools will not be able to tell you that the person has had a bereavement, lost their job or gone on holiday. Context is everything; marketing automation in B2C needs to get better at understanding individual context – particularly if marketers are going to get better at automating content marketing.
The smoke and mirrors of content personalisation in the B2C space
Over the past few years, you will have no doubt seen a considerable amount of activity across marketing automation which has been stimulated by bigger vendors realising that simply flogging their B2B stacks will not work for brands that needs to communicate with millions of individual customers – all of whom at different stages of the ‘customer lifecycle’ (ugh).
As a response to this, enterprise vendors who are making a rush for the new frontiers of B2C automation have made a flurry of acquisitions to supposedly power better content personalisation – both marketing automation platform companies such as Eloqua (acquired by Oracle), Aprimo (acquired by Teradata), Unica (acquired by IBM), Marketo, HubSpot, and Pardot (acquired by ExactTarget, which was then acquired by Salesforce.com and web content management companies such as Day Software (acquired by Adobe), Autonomy (acquired by HP), Ekton, SDL, Sitecore, and many others. There’s been a lot of consolidation here, and likely more to come.
Whatever moniker these new mega-systems are carrying – all of them are sold with tantalising promises of intuitive, real-time personalisation.
But don’t believe the hype: a deep-dive into many of these systems shows that it’s the same-old ‘if this-then that’ rule-fixing – now with a CRM or DMP bolted-on meaning you now have hundreds (perhaps thousands!) more customer segments to plan around (yaaay!).
The devil is in the detail.
Content marketing automation: rules vs. algorithms
Rules-based personalisation engines are pre-defined, static and have a limited effectiveness. There are drawbacks to this approach: creating, implementing and maintaining “rules” is arduous, time-consuming and, ultimately, very costly. Rules simply identify groups and categorise people; as such their intelligence is finite and ineffective when dealing with big data sets.
Rules bucket people into similar categories which is fine with small and narrowly focused websites or email marketing segments. But in a B2C context where customer records as long as a piece of string, increasing the amount of customers pushed whilst maintaining the limited number of segments or ‘buyer personas’ beloved of B2B, is a recipe for marketing irrelevance. Essentially you start to tar everyone with the same brush and your marketing messages become less relevant to a larger group of people.
By contrast, algorithmic personalisation analyses and creates potentially a limitless amount of machine generated micro-segments. More importantly, rather than being rigid and constrictive rules-based automation, the algorithm learns, adapts and evolves from these experiences.
With algorithmic personalisation, an algorithm monitors user behaviour patterns. A model is developed and personalised individual content is presented. The algorithm learns from these experiences and as the individual’s likes and dislikes change over time these changes are reflected, in real time, in the content displayed. The algorithm is constantly evolving, learning and adapting in real-time.
In 2014, as we move into a Big Content era – customers expect a ‘big content’ experience; 1-2-1 communications that understand their context and whims perfectly. This requires not only a deep understanding of the customer but also large volumes of categorised and available content. This is well beyond what a legion of campaign managers, CRM experts and content creators can handle – the ‘heavy-lifting’ has to be done algorithmically.
Moral of the story: if you have a large user base with diverse interests and a large (and growing) database of content – a real-time algorithmic approach is the only way to go.
With this in mind, as you go about investing in marketing automation systems this year – make sure that it comes with a layer of content intelligence.
