Back in the late 90s I had the absolute pleasure of working for Video Games developer, Maxis, and the Living Legend that is Wil Wright (he was seriously called that on his business card), creator of SimCity, the Sims and a whole range of other Sim games.

At that time, I was running the Maxis European office and responsible for the launch of SimCity 2000. For those of you who do not know (where have you been!) SimCity is one of the greatest video game franchises of all time and SimCity 2000 was the Terminator 2 equivalent of the franchise. As Kotaku put it, “[It’s] the game that struck the perfect balance between accessibility and the finer tuning of city-management.”

The launch of SimCity 2000 coincided with Ken Livingstone’s appointment as Mayor of London and we came up with the idea of sending Ken a copy of SimCity 2000 and grabbing a photo op to promote the game launch.

We went down to his office armed with a copy of the game, installed it onto his PC and gave him the basic tips and tricks to get started. What I remember about the whole encounter was; aside from how tall Ken Livingstone actually is – he had his desk jacked up to give it more height, and that he was a lover of hedgehogs – he had just come from a meeting at the Tiggywinkles hedgehog charity; how excited he was to get playing and prove out all his theories of how a city should be managed. We left him with it, got our photo of him playing the game and off we went.

However, it didn’t quite work out as expected. A week after our visit I got a call from a rather puzzled Mayor of London. He told me that he had built his city, put in loads of public transportation, schools and other amenities and lots of other fantastic benefits for his residents. In order to pay for them he hiked up the taxes but to his dismay, no matter how much he gave his simcitizens, they left his city in droves and he kept on failing.

I had to explain to a rather irate Red Ken that the AI which influenced the outcomes in SimCity 2000 was based on the theories of Systems Scientist Jay Wright Forrester.

Forrester argued that the use of computerized system models to inform social policy is far superior to simple debate, both in generating insight into the root causes of problems and in understanding the likely effects of proposed solutions, rather than just doing what felt right – as often what felt right, resulted in the wrong outcome.

In SimCity 2000 Wil Wright implemented Forrester’s theories to calculate values such as the city’s education, unemployment, and growth rates, and these figures in turn determined whether the city prospered or not. In doing so Wil Wright codified Forrester’s theories as the consistent and underlying set of rules onto which the entire game was built.

But this allowed for no further intelligence or evolution, which was where Ken went wrong. Instead users had to develop their city gradually, understanding what works and iterate. That way the underlying rules could be slowly worked within and around. Through testing and learning, players could learn what rules could be bent, worked around and broken to their advantage.

In a similar way, this human AND machine combination that is constantly testing, failing/winning, scaling and learning is at the heart of what we do to service our clients at Hearts & Science.

AI and Machine Learning-based models help us generate insights to understand the root cause of why a customer has behaved in a certain way, as well as to understand the likely effects of our proposed media solutions for the future, then we iterate to optimise the plans.

And thanks to modern computers and data storage systems we can continually build up datasets that describe the world around us from which we can observe the real rules. Crucially though as we grow our datasets, we modify our knowledge to account for new insights. If we stick to a static set of rules as in a game, then we’re destined to fail. Instead our rules have to constantly be poked, prodded and tested by our agency team so that they evolve and get smarter all the time.

The ability to rapidly add data and having a flexible approach to insights and analytics is at the epicentre of what is different at Hearts.  We are learning all the time, not just mindlessly following the same media rules of the past. We can see how TV has influenced sales, how Social Media has expanded our customer base, how one creative has worked but another has failed.

We also don’t need to construct a complex set of rules and then build out simulations to see how things may pan out. We are far more scientific about it; we observe, then analyse to identify the rules for a period of time, then we design tests to take advantage of these new rules for the future.

As one SimCity 2000 player, who went on to become a real town planner put it, ‘some of the game’s most popular hacks don’t work at all in real life. Putting a coal plant in the corner of your map protects a SimCity from much of its pollution, but in the real world, that would mean the smoke just covered a neighbouring town instead.’

Cities and their policies cannot be designed in vacuums and neither can media strategies – it is imperative to combine the science with the heart and as the name suggests, at Hearts & Science we have some of the best ‘real life’ people working with some of the best AI and Machine Learning, combining the two to help our clients grow their businesses.

 

Jo Cooke, Chief Development Officer, Hearts & Science UK