How I Fell in Love with Concentric Market
This is a love story. Not a romantic one, but a love story of a parent to a child. The parent is me, Dejan, the CTO of Concentric, and the child is Concentric Market. You may call Concentric Market a software application. I see it as a helpful, efficient, intelligent, albeit artificial member of our Concentric family.
Concentric Market is a simulator, a machine unlike any other that answers questions about consumer markets in minutes. I am also not the only parent, nor the most important one. Concentric Market has dozens of parents: all my teammates, our users, board members and friends of Concentric. It takes a village to raise a child after all, even if it is an artificial one. But this is my love story.
Today is Concentric Market’s eighth birthday. Like all software applications, this one does not have birthdays on the same day each year. We celebrate whenever a new major milestone propels it to a release, in this case, version 8.0. This version of Concentric Market is easier to work with, faster when providing answers, and more collaborative.
I wanted to tell you this story because I think you should meet Concentric Market. As one of the early “mind augmenters,” software that extends the power of human decision-making, it has a bright future. There will likely be much more to come. Humanity has built a complex world outside of nature. Our biological brains are not suited to always make good decisions in economies with million moving parts. In such complex environments, Concentric Market and others like it will help us consider more aspects of our decisions. The mind augmenters will analyze more trade-offs than we are able on our own. And they will help us avoid unintended consequences in our lives, communities, and societies.
To tell you what Concentric Market is capable today, let me start near the present day.
A few days ago, it was late at night at Concentric in Cambridge, MA. I was on a call with a user, on the other side of the World. I felt tired. It was release week. But she had a set of urgent questions that her bosses needed answered. She did not know it was possible to answer them. She was nervous. She went through the list:
- how much should we emphasize brand versus product in our campaign?
- how do metrics like perceptions and consideration affect sales?
- should we invest in the new channel that others seem to be flocking to?
- what if our competitor does what we think they will do next year?
- how much should we expect to invest to get to the target sales number?
- what if there is another economic downturn?
The questions stopped for a pause of a few beats. I asked, a bit too soon: “Is that all?” This came across very cocky, let me tell you. These are not simple questions, as you know. This person and others like her have been facing similar vexing conundrums without relief for decades. And here I am asking if that is all she’s got. I did not mean to be condescending. The social skills of a software developer deteriorate as a release date nears. Yet, I did know that these questions are a trifle for Concentric Market. There was a laugh on the other end of the phone and a challenge: “I appreciate your confidence, but show me how we solve this!”
Now, forgive for a moment a prideful parent when describing his child. In a half hour, I demoed how Concentric Market helps someone like her. I showed how to navigate the interface to define her question, integrate her data, validate a model, evaluate the past, and experiment with dozens of scenarios in the future. Her next question:
“This seems too easy. What is the catch?”
Of course, there is a catch. It’s a big one. Massive catch. Have you ever seen a child play the piano, and play it well? Have you seen young Messi on YouTube dribble the ball? Have you ever been around a toddler who has just started speaking? Those wondrous moments of seamless achievement are a product of countless hours of the child and the parents toiling away in a cycle of the mundane and the repeated. Day-in and day-out a whole village’s worth of changing diapers, making food, babysitting, tying shoelaces, kicking the ball, and teaching, all towards the singular goal of human development. Combined with natural predilections, you get a young Mozart, a Messi, and your very own precious chatty Cathy.
The catch is the same for Concentric Market. Dozens of us, the parents, have spent the past decade to get to this level of ability on Concentric Market’s eighth birthday. We’ve spent a combined few centuries of what you would call FTE (full time equivalent) time. The catch is on us, not on those of you, who may be meeting Concentric Market for the first time. The existing members of this village have gone all in, often putting lives, finances, and relationships at risk. We could not help it. This artificial child mesmerized us all. Also, Concentric Market, being made of code, has some strong natural predilections, too. It can do math in its head like it’s no one else’s business. It can search through hundreds of parameters and get to the correct answer in minutes. It makes no unforced errors. And most importantly, it learns as it goes.
So now that you know what Concentric Market is and what it can do today, let me share how my love of it evolved in three stages.
Stage 1: Falling in love with the science
I learned about the DNA of Concentric Market before it existed. I spent my years in college and grad school enthralled by artificial intelligence and political science. In the summer of 2005 I was looking for a place where I would combine them through social simulation. I wanted to connect the efficiency of computer science with the topic of human decision-making. Being young and enthusiastic, I wanted to work towards answering the tough, non-trivial questions of our World. I still do, when I take a break from all the metaphorical diaper-changing.
I found what I was looking for at Icosystem, a consultancy composed of scientists and coders. The Concentric village was already under construction. Eric Bonabeau, Paolo Gaudiano, Pablo Funes, and Andrey Anev were already working on the key algorithms. They were perfecting machine learning, agent-based modeling, network science, and evolutionary computation to solve thorny problems for international organizations and corporations.
Having studied behavioral economics, I added those ideas to the mix. Prototypes showed promise for a restaurant chain, a shampoo maker, an automotive maker, and a health insurance company. No one was sure you could simulate human populations with precision, but as it turned out, you could.
One day, a senior scientist on the project mentioned to me a tantalizing possibility. He said that in the future we are likely to combine the algorithms we are working on into a single software application and likely a separate company. The application will be flexible enough to handle any market. One app to rule them all! This would make the science of social simulation broadly accessible. I could not wait to be a part of that future.
In 2008 we met Greg Silverman, now the CEO of Concentric and our village chief. Greg taught us marketing science and how to understand our potential users. He also taught us how to build a company around what had already become a unified code-base named Concentric Market. It was a clunky, unsteady, buggy artificial toddler, but like a young Messi, you could tell: the kid had potential. In 2010 Concentric, Inc. was born. Concentric Market had a home.
Stage 2: Falling in love with the applications
By the time Concentric started, there was no question any more whether Concentric Market would be useful. It could show results by consumer segment, by metric, over time, aggregated or on an individual-level. It had the power to answer all kinds of questions faster and better than its alternatives. The goal at hand was to prove that is was valid. With the help of a whole set of new parents, who put their careers at stake, the cases started to mount. A few early adopters provided a testing ground to demonstrate that Concentric Market could be applied to multiple markets and multiple industries.
It wasn’t always successful, and it was never easy. Each new application would uncover new limits. There were other kids on the playground, all with protective parents on their own. Sometimes the old kids on the block punched. There were dogmas about how business questions got answered and distrust in a gifted newcomer. There was a recession. Sponsors moved jobs. Legal departments broke contracts. Yet, we, a gang of wide-eyed parents continued to expand the village, write code, build support materials, and evangelize.
Despite the challenges, it was intoxicating to see how a simple piece of code could be so versatile. Parents often say that watching a child do something for the first time awes them. “Six months ago, she could not walk and now she is arguing about putting a jacket on!”
It was amazing seeing Concentric Market for the first time pass a hold-out test with 2% weekly mean absolute percentage error (MAPE), a standard time-series forecasting metric. Who thought it that?! Companies employ an army of data scientists cannot achieve that level of precision. It must have been at daycare. My heart would swell with joy.
Sometimes Concentric Market got into trouble. It could not forecast with the desired accuracy, could not integrate a data set, would not be flexible enough. In those times the village would come together and find a solution. Sometimes we would do by hand the things Concentric Market had not automated yet. All of these new functionalities eventually made their way into the core code. In this way, Concentric Market continued to grow and help people make decisions.
From 2010 to 2015, Concentric Market ended up accomplishing plenty. It helped people meet sales targets, grow subscribers, choose retailers, advertise financial services, build brands, raise money, and win elections. With dozens of case studies, partnerships with some of the largest global marketing organizations, and dozens of references, the validity of Concentric Market was no longer in question.
Stage 3: Falling in love with collaboration
By the fall of 2015 hundreds of companies had worked with Concentric Market with users on five continents. Yet, as I interacted with users and observed our support team help them out, I saw a problem. The child had become difficult. It’s a painful thing for a parent to see. Concentric Market had added so much muscle as features, that it was not easy to communicate with it any more. Confused users would get stuck, and often needed to call the parents.
While Concentric Market was clearly useful and valid, it was not simple and easy to socialize. It was not enough to be smart, it needed to also be social. We needed to re-organize the user interface and re-define the user experience. Concentric Market needed to speak the language of its users. And it needed to follow an intuitive workflow.
In the meantime, the Concentric villagers ventured out and learned more about how decisions happen in large organizations. We saw the inner workings of more companies than most of us had ever wanted to. We discovered the political, bureaucratic, and emotional maze of the modern corporation. Here was a challenge equal to simulating social dynamics: How to make an organization get to better decisions, faster?
The biggest obstacle to the user on the phone a few days ago was that data and people within her organization did not communicate well. They cannot answer whether to emphasize the brand or the product, because separate teams are in charge of branding and product. Each team evaluates separate outcomes. Data and people do not connect. In that environment, it is tough for decision-makers to analyze trade-offs between initiatives. It is often unclear what happened, much less what should happen next.
Because this user’s problem is typical, we made working with Concentric Market easier on one’s own or across teams. Each step of the workflow speaks to a particular persona. An account manager or a product lead works in the Define section. A data gatherer works in the Integrate section, who in turn sends it to an analyst, who tests the model. The final two steps, Evaluate and Experiment are where anyone can ask any question of Concentric Market and get an answer in minutes.
And Concentric Market became a connector of our users to our village. Most of our thinking is now part of extensive help topics located next to each workflow section. Hints show up next to most fields. An automated screenshot goes to the support team whenever a user reports an issue.
Our demos went from a couple of hour-long meetings to a single half hour. Today, on most calls, I hear a “wow!” Users share a sense of relief that such powerful intelligence is accessible in simple, natural ways. I am a proud parent again. We are not dealing with a moody teenager, but a gifted young adult. The design of version 8.0 began in December 2015 and now, a full year later, we are baking the cake and putting eight candles on it. The village is getting together to celebrate another birthday of our pride and joy.
But before I end this story, I want to share one more thing. I said at the beginning that I am not the most important parent. Do you know who it is?
It is you. Without you, your shares, likes, comments, demo requests, feedback, and trial of Concentric Market, the village would stagnate. Your continued support (even simply sharing our blog) is the life-blood of our company. So, give us a shout-out, send us your thoughts, tell us your vexing conundrums. Spend some time with Concentric Market and see how it can be a force for good in your environment. On this eighth birthday, Concentric Market is still young. We keep track of a hefty roadmap of work ahead of us to unleash its full potential. Join us on that road.
And come over and visit. We may still have some leftover birthday cake for you.