Original article was published by Glenn Gow on Artificial Intelligence on Medium
The AI Threat: Winner-Takes-All
I’m not a very good poker player. (That’s why my friends always invite me to play.) They like to play until the winner takes all of the winnings, which ultimately includes all my money. I’m OK with this because it’s only a small game, and for me, it’s worth the price of admission.
Now shift your thinking to the business world. Imagine running a business where your main competitor has the dominant market share, and you are in second place. You have been struggling for years to overtake your primary competitor, but they have advantages in product, in costs, and in marketing that you can’t match.
You are improving, but your competitor is improving at the same rate. You are stuck in a perpetual second place.
Now imagine a technology that can help you improve faster than your competitor. A technology that can enable you to surpass your competitor and eventually take an insurmountable lead.
That technology is AI. AI is different from every other technology disruption because it continually gets better every day with little or no human interaction. In some cases, it gets better every day all by itself.
We see technological progress all the time. For example, Apple comes out with a new iPhone just about every year. However, the advances in technology like the iPhone, are driven primarily by humans. With AI, the progress is continuous, and it can happen without human intervention. AI-based progress happens automatically and much faster than human-driven technological advances.
Imagine your competitors are getting better and better along the straight line above. They are focused on process improvements, on cost reduction, on more effective outreach to customers. They get better every day, but at a linear growth rate.
Now, imagine that you have embraced AI. Once your AI-based advances kick in, your growth rate will be higher than your competitor’s, and the rate of growth will accelerate. The speed at which AI improves means you can get to a growth rate where your competitor can never catch you.
I’m going to use a fictional example to show how AI can help one competitor overtake another. (My apologies in advance to anyone who works at Procter & Gamble or Kimberly-Clark. I am making up a scenario to make a point, which does not reflect reality in any way.)
Let’s say you work at P&G, the makers of Pampers diapers. The Pampers brand manager decides that investing in a new marketing campaign is more important than investment in AI.
Over at Kimberly-Clark, the brand manager of Huggies diapers decides the only way they are going to win against the Pampers juggernaut is to invest in AI.
The Huggies brand manager decides on two objectives:
1) Acquire customers at a faster pace and a lower cost than Pampers, and
2) Sell related products to the same target market.
Huggies begins by collecting (and cleaning) all the data they have about their current customers. The first thing they apply AI to is customer acquisition. They apply machine learning to market-mix modeling so they can understand the optimal mix of advertising, coupons, retail discounts, in-store placements, and so forth. Using AI enables them to run simulations and then continually test and refine what is working and what’s not working to rapidly improve the rate of customer acquisition while reducing the cost of customer acquisition.
For example, the AI may predict that offering a 15% discount on a large package of diapers one month before a mother’s estimated due date will result in high response rates. The brand manager approves of the experiment and likes the results.
As Huggies gains more customers, it gains more data about customers and feeds that data into the machine learning models. As more data is fed into the models, these models learn and improve in a continuous cycle of improvement. Huggies is now learning about how to acquire customers faster and less expensively than Pampers is.
Huggies has reached an inflection point where their growth rate is higher than Pampers. Once Huggies surpasses Pampers in market share, all things being equal, Pampers will never catch up!
Now the Kimberly-Clark team wants to sell related products to the same customers. They apply AI to their data about their customers to discover patterns of behavior and preferences that are not obvious to humans. This pattern-matching is something AI is very good at.
For example, the AI may discover that families purchasing diapers eventually start buying band-aids and wet-wipes. But since other brands within the company offer those products, there’s no obvious connection.
The AI being used by the Huggies brand manager suggests that not only will Huggies sales increase, but the overall share of wallet for Kimberly-Clark will increase if Kimberly-Clark ties specific promotions together.
Once again, as Kimberly-Clark gains more data about families’ behavior with babies, the more it learns about the wants and needs and buying preferences of these families. Their AI gets smarter about how to increase share-of-wallet in a virtuous, continuous cycle.
The brand manager for Pampers is left wondering what happened. P&G’’s learning curve has stayed the same, while Kimberly-Clark’s learning curve — driven by AI — has dramatically accelerated.
Going back to my poker analogy, if I use AI to become a smarter poker player on every hand, I will improve at such a rate that eventually no one can compete with me, and I will keep winning everything. (No wonder I’m not getting invited to my friends’ poker games anymore!)