Rob Webster is Founder of Canton Marketing Solutions. He’s worked in the adtech industry since 2001 and is NDA’s monthly columnist.
Thinking Fast and Slow by Daniel Kahneman is one of my favourite books of the last 10 years. It shows how the brain essentially thinks at two speeds, the instinctive and automatic mind and the much slower conscious mind. This can be a great thing, the automatic mind is needed, essential even when speed is of the essence – when jumping out of the way of a car for instance.
The conscious mind is crucial for more considered decisions, making sensible purchase decisions for example. The book highlights the inherent conflict between these ways of thinking and how by being aware of it we can make better decisions.
A similar dichotomy applies to marketing, with marketers either making decisions very quickly or quite slowly. These decision-making processes are a consequence of how marketing has developed. Like the examples of the mind in Daniel’s book this is both a requirement for success but also without examination can lead to many of the failures in modern marketing. To improve how brands execute they must challenge the pace of their decision making.
The Planning Process
The main difference between digital media and traditional media is speed. The same fundamentals apply but the speed at which execution and measurement occur is much faster.
Traditional media planning is a process that can take months. It was born at a time when building creative, booking media and measuring results all took weeks and months. This long lead time can be an advantage, the very best brand campaigns have all had significant care on their creation. The side of marketing which is an art should not be unduly rushed.
Digital media is so much faster and can be booked and measured in hours or minutes and with the help of algorithms faster still. Dynamic creative optimisation can be just as fast.
Real world events can be responded to in a way impossible with traditional media, opening up new opportunities. Like a retailer responding to its product appearing in a popular TV show, or a travel company responding to international events like coronavirus.
This dichotomy in speed of planning approaches manifests as two common problems.
- Long term planning does not accommodate the speed of digital media. With budgets locked down there is reduced opportunity to invest in short-term options which would not have been foreseen at the campaign’s inception. Money is left on the table that could have been effectively spent on marketing. This is the mistake made by many brands favouring traditional planning routines.
- A focus on short-term tactical options can result in creative that lacks bite. It can result in execution that has not had time to consider the big picture, focussing too closely on the bottom of the funnel. Ultimately it can be advertising for advertising’s sake rather than activity furthering the interest of the brand.
The solution is a multi-tiered approach. Mid to long term planned campaigns along with bottom up planned direct response and budget left over for contingency. The exact layout of this will be different for brands in different sectors. Retailers may have more need for the short term, luxury brands more long-term planning. However all brands need to evaluate the risks and benefits of thinking fast and slow.
Algorithms v humans – tactics v strategy
The fastest way to make execution decisions is through an algorithmic approach. Algorithms can change media (bid) prices based on predicted conversion rate across a multitude of variables in real time. They allow more decisions to be made faster than any human and when used correctly this can drive significant performance improvements.
If you want to think fast, use an algorithm. And machine learning is having somewhat of a renaissance in marketing. No longer is it locked behind a black box and progressive brands are using machine learning modules from the various cloud computing companies. Some CDP systems and other specialist companies are democratising the machine learning process.
Yet algorithms only can respond to data that they know about. A prominent data scientist at a leading machine learning company in adtech once said to me “We learnt these techniques in finance. In finance, if you are evaluating human-led trades rather than machine learning, always back the machine.”
But in marketing, the marketeer has access to knowledge about their market impossible in finance. For example, the marketeer can know that they will launch a last minute sale on the weekend, or the impact of a retail or sports event.
What this all boils down to is tactics vs strategy. When it comes to strategy marketing is still too complex and subjective for machines to make the decisions. As we will see below, in a world where machines have yet to nail attribution never mind understanding the concepts of awareness and consideration, strategy is still very much in the domain of humans (thinking slow).
For tactics however, particularly in programmatic and biddable media as well as personalisaton, thinking fast through AI and machine learning is the best bet for the future.
Attribution and measurement
Looking at daily, even hourly results has become standard for ecommerce companies. Day on day, week on week and year on year results are constantly analysed. This fast thinking can be vital to survival because at key periods (think Black Friday in retail) there simply isn’t the time to lose if results are not going well.
Whilst this urgency can be helpful, essential even, it can cause big problems particularly when you look to atomise the data. Top line sales and revenue per day is a matter for the bookings system.
However when you look to go deeper in many modern systems, key errors can occur that are normally missed when thinking fast. For example, attribution and analytics has been massively impacted by the changes to cookie technology, notably Apples Safari IPT.
Naming conventions can be inconsistent month from month, year from year. Simple conditions can also be missed such as when Easter falls, the weather or the periods of a sale. These changes can dramatically impact the interpretation of the numbers. What may look like the failing of a given marketing channel or campaign may in fact be down to changes in underlying tracking conditions.
This is only going to accelerate as legislation and technology turns upside down measurement and attribution practice over the next two years.
So what’s to be done? Businesses must not lose the speed of action made available by fast trading. However they must use quieter periods to analyse and contextualise changes in numbers. Indeed brands need to invest more in attribution and analytics to understand the why behind the daily sales figures. Econometric modelling (otherwise known as predictive modelling) was (and is) the pinnacle of measurement for traditional marketing.
Yet such models rarely update faster than once a quarter, far too slow for digital channels. Multitouch attribution is much faster but going through a technological transition right now and it contains notable holes. Expect a way to be found to combine the best of these two disciplines whilst respecting privacy in the coming years. Then we will truly have the information to make more detailed decisions fast.
Creative
Last but by no means least is the world of creative. Creative more than any other benefits from thinking slow. Authenticity is such a vital part of modern brand building and that cannot flip moment to moment. The greatest marketing campaigns all came from care taken over the strategy and most importantly in that creative and messaging.
Yet creative execution must speed up. Social media thrives on fresh content. Creative that is right for Instagram often is not right for Facebook never mind Amazon, Snap or TikTok never mind traditional media.
What is needed, as with planning, is a layered approach. Time and care taken on the major messaging and creative themes. That setup needs to feed the machine of fresh content and messaging at speed. This means creatives need to learn much much more about online execution and DCO whilst not forgetting the core skills.
Final thoughts
Disruption in marketing has been prevalent for years. This has played merry havoc to the pacing that marketing needs to succeed. Brands and marketers need to look at the pace with which they make decisions and critically evaluate that process. Most brands need to change their practices to deliver better work.
Much of this boils down to education though. How many traditional marketers understand biddable and programmatic? How many online practitioners get the principles of marketing?
In my own experience both are at critically low levels. Yet the problems go further. Finance teams, IT teams and particularly in a privacy-by-design world, data and legal teams are all intimately involved and have their own biases on speed.
For brands to succeed, all of these need to learn to think fast and slow – and when to do each.
