Garry Kasparov will forever be remembered as perhaps the greatest chess player of all time, dominating the game for almost twenty years until his retirement in 2005. But ironically he may be best remembered for the match he failed to win twenty years ago in 1997 against IBM’s Deep Blue chess computer. That watershed moment – marking the point at which computers effectively surpassed humans in chess-playing ability – prompted much speculation and hand-wringing about the coming obsolescence of the human brain, now that a mere computer had been able to beat the best chess grandmaster in the world.
Since then, computers and chess software have only grown more powerful, to the point that a $50 commercial chess program (or even a mobile app) can beat most grandmasters easily. Faced with this, you might expect Kasparov and other top-flight players to have grown disillusioned with the game, or defensive about the encroachment of computers on their intellectual territory; but in fact the reverse is true.
Today’s chess grandmasters make extensive use of computers to practice, try out new strategies, and prepare for tournaments, in the process becoming a little more like the machines that outpaced them in 1997. Kasparov himself was instrumental in pioneering a new type of chess game, Advanced Chess, in which humans are allowed to consult with chess software as they play. In his new book, “Deep Thinking: Where Machine Intelligence Ends and Human Intelligence Begins”, Kasparov writes about an Advanced Chess match he played in 1998 against Veselin Topalov:
“Having a computer partner also meant never having to worry about making a tactical blunder. The computer could project the consequences of each move we considered, pointing out possible outcomes and countermoves we might otherwise have missed. With that taken care of for us, we could concentrate on strategic planning instead of spending so much time on calculations. Human creativity was even more paramount under these conditions.”
What Kasparov and his successors in the competitive chess-playing world have discovered was that, when it comes to chess, the strongest player is not man or machine, but man and machine. In fact, a new kind of chess tournament has sprung up, Freestyle Chess, in which teams of humans and computers compete against one another, each bringing their respective strengths to the game: creativity, strategy and intuition from the humans, and tactical outcome prediction from the computers.
And your point is?
You may be asking what relevance this has to digital marketing. In fact, there are strong similarities between chess and marketing (particularly digital marketing): they are both highly quantifiable pursuits with clear outcomes which have historically relied solely on human intuition and creativity for success.
As in chess, digital marketing relies upon a continuous reassessment of the ‘board’ (customer behaviors and history) in order to decide upon the next ‘move’ (a particular campaign communication aimed at a particular group of customers). Once the move has been made, the board needs to be reassessed before taking the next move.
Today’s digital marketer is much like the chess grandmaster of the early 1990s – they rely on their intuitive understanding of their audience’s makeup and preferences to decide what offers and messages they want to deliver, to which users, and in which channels. Of course, digital marketers understand that measuring campaign outcomes and audience response (using techniques like control groups and attribution analysis) is very important, but most still operate in a world where the humans make the decisions, and the computers merely provide the numbers to support the decision-making.
When Kasparov was asked in 1990 if a computer could beat a grandmaster before the year 2000, he quipped:
“No way – and if any grandmaster has difficulties playing computers, I would be happy to provide my advice.”
Today’s digital marketers can be forgiven for exhibiting some of the same skepticism. Ask them how they came up with a new idea for an ad, or how they know that a particular product will be just right for a particular audience, and they may not be able to answer – they will just know that their intuition is sound. As a result it can seem incredible that a computer can pick the right audience for a campaign, and match the appropriate offer and creative to that audience.
But the computers are coming. As I mentioned in my earlier post on bandit experimentation, companies like Amplero, Kahuna and Cerebri AI are pitching intelligent systems that claim to take a lot of this decision-making about creative choice, audience, channel and other campaign variables out of the hands of humans. But where does that leave the digital marketer?
We welcome our robot colleagues
The clue lies in the insights that Kasparov ultimately drew from his defeat. He realized that the strengths he brought were different and complementary to the strengths of the computer. The same holds true for digital marketing. Coming up with product value propositions, campaign messaging and creative are activities which computers are nowhere close to being good at, especially in the context of broader intangible brand attributes. On the other hand, audience selection and targeting, as well as creative optimization, are highly suited to automation, to the extent that computers can be expected to perform significantly better than their human counterparts, much as chess software outperforms human players.
Clearly humans and machines need to work together to create and execute the best performing campaigns, but exactly how this model will work is still being figured out.
Today, most digital marketers build campaign audiences by hand, identifying specific audience attributes (such as demographics or behavioral history) and applying filters to to those attributes to build segments. The more sophisticated the marketer attempts to be in selecting audience attributes for campaign segments, the more cost they incur in the setup of those campaigns, making the ROI equation harder to balance.
The emerging alternative approach is to provide an ML/AI system with a set of audience (and campaign) attributes, and let it figure out which combinations of audience and offer/creative deliver the best results by experimenting with different combinations of these attribut
es in outbound communications. But this raises some important questions:
- How to choose the attributes in the first place
- How to understand which attributes make a difference
- How to fit ML/AI-driven campaigns into a broader communications cadence & strategy
- How to use learnings from ML/AI-driven campaigns to develop new value propositions and creative executions
In other words, ML/AI-driven marketing systems cannot simply be ‘black boxes’ into which campaign objectives and creative are dumped, and then left to deliver clicks or conversions on the resulting campaign delivery. They need to inform and involve marketers as they do their work, so that the marketers can make their uniquely human contribution to the process of designing effective campaigns. The black box needs some knobs and dials, in other words.
The world of chess offers a further useful parallel here. Chess grandmasters make extensive use of specialized chess software like Fritz 15 or Shredder, which not only provide a comprehensive database of chess moves, but also training and analysis capabilities to help human players improve their chess and plan their games. These programs don’t simply play chess – they explain how they are making their recommendations, to enable their human counterparts to make their own decisions more effectively.
These are the kinds of systems that digital marketers need to transform their marketing with AI. In turn, marketers need to adjust the way they plan and define campaigns in the same way that chess grandmasters have dramatically changed the way they study, plan and play games of chess in the last twenty years, working alongside the computers before, during and after campaigns are run.
In 1997, it was far from clear how chess, and the people who played it , would react to the arrival of computers. Digital Marketing stands on a similar threshold today. Twenty years from now it will seem obvious how marketers’ roles would evolve, and how technology would adapt to support them. We’re in the fortunate position of getting to figure this out as it all unfolds, much as Kasparov did.