One of my friends in the FMCG industry said something the other day that led to this article.
She said, “we hesitate to test (our advertising or products) with consumers because of FOFO – fear of finding out that we were wrong after all.”
As strategic insights go, this hit the bullseye – she didn’t just get me to see things in a new way, she also used the least number of words to do it.
(Bonus points for packaging her insight into a sticky acronym I could place next to my FOMOs and YOLOs.)
But there is another more interesting reason why this is a perfect example of strategic insight – it fertilised new connections for me – technology has made marketing faster. But has it also made it easier for us choose shallow learning over deep learning? Is FOFO therefore a defence mechanism to protect our egos because deep learning is by nature, painful?
In short,
When technology has changed not just the way we market, but also how we learn, are our learning muscles atrophying because of FOFO?
Grab your coffee and dive in.
Choose. Shallow Learning or Deep Learning?
Pre-social media, AI, programmatic media and performance marketing, most marketing development looks like this,
Meet Consumers and Develop Hypothesis → Generate and Test Ideas → Brief Agency → Judge Ideas and Finalise one → Test it → Go Live
There is a back and forth involved here – test, tweak, retest. Because of this back and forth, you end up with enough conviction behind one big idea.
But without realising it, you also learn more about the consumer along the way.
Now you can choose to learn in a shallow way, or in a deep way.
Chris Argyris and Donald Schön have researched this and they tell us that single loop learning is shallow learning because here, you focus on symptoms and so, stay at a surface level. It’s like taking Dolo to reduce fever.
Instead, double loop learning dives deeper below the waterline and questions your assumptions – before prescribing a medicine, you ask, what’s the root cause of the fever?
This is what it looks like in marketing –
Let’s say one of your hypothesis is, “yellow communicates freshness”.
If you are wired for shallow learning, you go all out and try 10 different colours, centre align the logo, and make the fonts scream “freshness.”
If you are wired for double loop learning, at some point you stop and ask, ‘what if my assumptions were wrong? Do people even want freshness?’
When you shallow learn, you might still improve your consumer rating. But you won’t know why at a deeper, insight level.
No one starts their day with, ‘today I choose shallow learning’. But does technology create conditions that make shallow learning easier?
A technology led marketing process might look like this –
An intern (or AI) generates 50 ideas → Algorithms run all 50 → We invest behind the 3 that get the most clickthroughs.
Fast, efficient, maybe even more affordable than the previous one.
Whereas the previous process gave us one big idea, this one focusses on volume over precision – much like a VC, you spread risk and a few breakthrough ideas pay for the rest.
The non-tech process does not automatically make us deeper thinkers. But it gives us more occasions to pause and question our assumptions.
The tech-led process does not automatically make us shallow thinkers. But it makes it easier for us to delegate the thinking to machines and still feel like we are learning.
I have been in meetings where the media team shares a 50 row sheet. Each row captures a campaign and its CTRs. The discussion revolves around moving money from lower performing channels and campaigns into better performing ones.
A bot could perform such deterministic tasks better than humans.
But 8 of us spent 2 hours daily on said sheet during campaign time.
In my experience, tech-led marketing is making it easier to shallow learn in three ways.
1. Tech-led marketing keeps us busy. which gives us the illusion that we are learning.
Tech-led marketing is fast and high volume, so it makes us feel productive. Each time we tweak a campaign and engagement % rises, we get a dopamine rush.
This is the learning equivalent of Instagram scrolling. We spend months scrolling campaigns, and can’t explain what the consumer wants and why some things work while others don’t.
2. One failed idea that we own hurts more than 47 we don’t own.
No one wants to feel like shit everyday when they can feel great everyday. Deep learning is frustrating. We are forced to live with ambiguity and self-doubt until we get the right answer.
In pre-tech processes, we back one big idea. This idea becomes our identity and we stake our reputation on it. If it fails, it hurts. It shakes our confidence.
But when an algorithm creates and tests 50 ideas and 3 win, we feel we’re hot stuff. The 47 that failed don’t hurt because we created them at arm’s length. The failure doesn’t feel personal.
3. We value better performing ads more than better learning.
Ad performance is visible. Learning is invisible.
When algorithms make creatives and machines buy media, we no longer need to stretch our brains to dig for that one perfect idea that will work. We become addicted to algorithmic success.
And even though our judgement muscles atrophy, we still feel like we’re learning.
Are we becoming highly productive managers of machines?
Our culture has become one where being wrong today equals being incompetent forever. In the race to look and sound smart today, are we postponing difficult discussions to when it might be too late?
We cannot keep learning at a surface level and expect the same playbook to work tomorrow also, because we are in the middle of the Red Queen race, where the landscape itself is shape shifting.
Which team do you think will be better equipped for such an environment? The single loop learners or the double loop ones?
Our job is not to be just execution machines, but to be learning ones.
We need both – speed and deliberation. We need thinkers who can read the machine data as well as a market. We need leaders who use machines as tools, not crutches.
When calculators arrived, teachers worried that students might stop learning. Turns out, the calculator took on the load of solving arithmetic. But the job of choosing which arithmetic equation to solve, whether the answer made sense and explain what it meant in the larger context remained with the students.
Companies pay big bucks to leaders who can outthink others.
To leaders who have pattern recognition and well-honed judgement about the market.
In short, leaders who have insight.
If today we operate with FOFO, we avoid double loop learning and our learning muscles atrophy over time.
This will lead to more meeting rooms where 8 marketers will spend 2-hours doing the work of bots while they delegate the discernment to machines.