0.1 The four ages of marketing

The scientific method goes something like this. An academic observes data and creates a theory. The academic will then attempt to prove themselves wrong and find evidence for their theory being false. If they’re unsuccessful, they’ll try again. And probably again, using a different method each time. If they reach a point where they are unable to think of a new way to prove themselves wrong, they’ll publish their findings to a community of peers. Those peers will then jump on the publication and a number of them will then try to devise further ways of proving the theory wrong. And when they’ve exhausted their efforts, a new cohort will jump on the theory, sometimes many years later.

Eventually, when enough testing of the theory has been completed and it’s proven to hold up, it will be accepted and new theories will be built upon it. But, should new evidence arrive at any future time to contradict its findings, the theory will be revisited, reviewed and, most likely rejected.

Let’s contrast this to what I will provisionally call ‘the marketing method’. When a marketer has a new idea for how to create business value they will attempt to share it to as many of their peers and colleagues as possible. Ideally, they will give it a snappy name to make it memorable.

This idea will be backed up with some compelling results. With a bit of luck, this new idea will pick up some traction and others will attempt to repeat the formula, sometimes achieving similar results, sometimes missing the mark. No one likes to be associated with failure so any follow-up results that don’t meet success criteria get swept under the rug. The marketers achieving these superb results will be promoted for their work, may be offered speaking opportunities to share their experience with other marketers and will further increase the recognition of the original idea. And at some point a new idea will come along and the cycle will repeat.

Many of the imperfections of the marketing world can be explained by the importance of demonstrating value. Marketing is, after all, a commercial environment. It is easier to get promoted by complementing the Emperor than by pointing out he has no clothes.

Sciences, though no by means perfect, generally advance because of a collaborative quest for deeper understanding. Marketing has generally been advanced by a desire for personal gain. Neither method is without flaws, neither is without merit. Arguably the sciences are also invaded by the need to prove value – a lot of investment is driven by the private sector – but the output is still motivated by a need for insight. In the sciences, people’s names are tied to their work forever as it is published and peer reviewed. In Marketing, people move roles and organisations and their experienced is mainly measured by their current job title.

Perhaps because of this self-interest, Marketing has an exceptional diversity of technologies, techniques and tactics. A simple look at a logo map of marketing technology vendors should be enough to prove as much. “Whenever I look at the list of martech providers out there I know I’ll always be safe for a job,” one digital marketer once told me. “There are so many technologies we could be using, I know there will always be things we can do more than we are now.”

For years the goal of enterprise marketing has been the same: perfect segmentation, personalisation and activation at scale. The term ‘ephemeralisation’ was coined by Richard Buckminster Fuller, the American architect, systems theorist, writer, designer, inventor, philosopher and futurist. The automobile was Buckminster Fuller’s example: though the first cars were really not many years old, Henry Ford’s production line had driven (if you’ll excuse the pun) the rapid development of better and better cars, with lower and lower costs. The theory was that with no upper boundary on the potential of the car, cars could be produced for lower cost and greater utility.

Of course, the world rarely works that way. Marketing as much as any business function has marched forwards on the back of the promise of new technologies and the realities of unplanned consequences. How many of these issues do you recognise?

• Speed: inability to respond quickly and at scale to external pressures

• Insight: Inability to unify all data sources into an action aligned measurement framework

• Communication: Silo’d teams running their own campaigns independent of strategic initiatives

• Impact: Inability to understand discrete effectiveness of marketing activities against business measures

• Experience: Difficulty unifying customer experiences into a holistic engagements

• Cost: Understanding what truly drives value and prioritising accordingly

So how did we get to where we are today?

The first age: the brand and broadcast age – from the dawn of time until the mid-1990s

The first age of marketing ran broadly from whatever point people saw benefit in branding their products or services until the early 1990s. In 1389, King Richard II of England decreed that landlords must put signs outside their inns, so that they could easily be identified – so from at least then. Either way, it was quite a long time ago. Marketing budgets were mainly focused on creating brand value. Sales owned most if not all of the engagement with a buyer and the sales environment was built on personal networks.

But there was a challenge: a rolodex is only so big. Most models of marketing effectiveness were measured by sales results. Marketing teams would run TV, radio, print or other common broadcast channels, sometimes using coupons or discounts as the call to action or sometimes pushing lifestyle and brand. To grow revenue you need to grow the sales team. This model only really scales linearly with sales people, though, and good sales people aren’t infinite. The organisation that can do more with less will steal a march.

The second age: the data age – mid-1990s to late 2000s

With the dawn of CRM ,a marketing database became the must-have tool. Data – the acquisition and the management of it – became a strategic differentiator. Marketing teams moved from being predominantly creative- and relationship-based to purchasers of technology solutions and managers of technically skilled teams to execute campaigns. Marketing CRM systems – both the creation and population – were the ‘it’ thing in the 1990s and early 2000s. Marketing Operations developed much more power as a team that could provide critical differentiation to an organisation. Integration of marketing CRM with sales CRM was a strategic challenge for many organisations.

The legacy of this era persisted for a long time in some organisation’s technical architecture. Sales databases disconnected from marketing databases with integration that was mixed at best between revenue billing, sales management, renewals and marketing, making it near impossible to get an holistic customer view. For marketers, this type of technical debt made pipeline attribution a manual and imprecise process. The opportunities you directly created were about the only effective measure of true business value, but that was a start. Organisations that grew up as digital marketing natives had a major advantage. Not weighed down by a legacy of dysfunctional integrations, they were able to architect their sales and marketing reporting to the needs of early-2000s “modern marketing.”

The third age: the automation age – late 2000s to 2020

The end of the second age of marketing was as inevitable. This was driven by the same imperatives: systems of data management and application that were superseded by new technical solutions. Data is a gateway drug to harder things, and once you’ve started managing data in a CRM system there can never be enough data, and it can never be clean enough. If data is not maintained, it ages and declines in quality. Manually maintaining data quality can be an expensive and complex process when done as a discrete activity.

Around the mid-2000s two types of companies had a transformative effect upon marketing. The first was the marketing automation company. Marketing Automation sidestepped the issue of managing data quality as a discrete exercise by allowing for the automation of communications at much greater scale. This essentially made campaigning the new way of maintaining data quality. With marketing automation, it was no longer about simply acquiring more and more data; the opportunity was how you made use of that data. Nurture, which was a manual job of sending emails or picking up the phone previously, could now be automated. Sources of information such as clicks and opens on emails could now be automated and scored. When a prospect hit a certain score, it could prompt a trigger for an Account Rep to call the prospect. It changed everything. This had the net effect of removing some of the burden of manual data cleaning by allowing campaigning and customer action to take the majority of the data manual work.

The other type of resource that emerged was the digital advertiser. With the likes of Google, Facebook and LinkedIn, closely followed by the network advertisers, marketers could target their core segments with greater accuracy than ever before without actually owning customer data. Tracking, and the potential it enabled, exploded. Account- and contact-based insights were now available to any marketer, and along with these insights came new tech companies that could extend these capabilities further and in more discrete ways.

To support these systems, new headcount was needed, and along with the headcount new budgets. Marketing teams lapped up increasingly consumerised IT applications that offered new takes on the same problem: how do I engage my customers in the most personalised ways. Anyone that has looked at most marketing automation tools will realise that under the surface, these are not simple tools. Marketing operations is now integral to the success of an organisation. Implementation, configuration and ongoing management of these systems requires skilled people, experience and investment. But the ephemeralisation of marketing – achieving that Nirvana of perfect personalisation for low cost – still requires another leap forward.

The fourth age: the intelligence age – 2020 to present

An inevitable challenge quickly appeared during the third age and remains a problem to this day: how to make all of these data sources useful. Data is only useful when it is connected. The insights on an account’s actions identified in one system are of only limited use if they either aren’t connected to similar actions in another system, or are missed entirely because there are too many systems to be effectively used. Usefully centralising data and the interconnection of marketing tools through APIs into core marketing dashboards remains a pressing barrier to many marketing teams.

Which is where we arrive at the current state of marketing: AI.

Artificial Intelligence is now being built into many marketing platforms as a way of making the trove of data that has now been accumulated more actionable. How to nurture, who to nurture, what all of these data points mean – the race is on to create or become the AI platform for marketing. For those that achieve first-mover status successfully, the rewards are likely to be many, but there will be a lot of cost sunk until that point with no guarantees of winning. And then, of course, there will probably be a “Fifth Age.” But we’re not there yet.

AI, unlike other solutions that focus mostly on collecting data, requires data to be fed to it to produce useful outputs. That means a lot of data. AI without broad context will be limited and come to incorrect decisions. Even AI with a ton of data can make poor decisions. Look at how much data Deep Blue, the machine that IBM taught to play chess, required. Or look at some of the disastrous examples of chatbots that have been trained on data from social media platforms and how quickly they move to extremist positions. In 2016, Microsoft released a Twitter bot called Tay, which was meant to interact as an American teenager, learning as it went. Instead, it learned to share radically inappropriate tweets. Google had to remove the ability to search for gorillas on its AI software after results retrieved images of Black people instead. Apple had to pull its much vaunted notification summaries when the BBC complained it had entirely misrepresented its reporting.

AI, like a baby, needs to learn. Is your marketing data really that comprehensive, complete and correct enough that you expect to get accurate results from it?