Posts Tagged ‘data’

When I’m 90

Tuesday, February 26th, 2013

I will always remember an old professor colleague of mine. I had not seen him in twenty years. He was the last person coming off the plane in Boston at midnight and looked quite old. He was disheveled, slowly dragging his oversized suitcase up the jet way, holding a half-open shoulder bag full of flip charts. He was still wearing the same patched-at-the-elbows rumpled suit, and his shirt was stained by markers ink. He saw me waiting for him, and a large smile illuminated his face.

“I’m just back from the West Coast”, he shouted at me from twenty feet away. “Three-day-workshop with a bunch of kids managing a high-tech start-up. Not so bad for a ninety-year old guy who does not even use Facebook”.

This image of my old friend’s triumphant smile has stayed with me all these years. The ancestral anxiety of any educator is irrelevance. This is particularly true those of us who teach innovation. I want to look like my old friend when I’m 90. I want to come off the plane, physically exhausted, but exhilarated at the thought of having just learned about a new industry. When my consciousness starts waning on my hospital bed, I want someone to whisper in my ear how neural networks and 3D reservoir modeling algorithms are built.

I’m not quite 90 yet, so I’ve got a bit of a head start. The front-end discovery part of any new business is by far where I have the most fun (I feel like I’m being paid to go to school). Writing-up success stories based on consulting work I have done and teaching them on the lecture circuit is the second best part. Everything in-between is just hard work.

I’m learning about six industries right now and feel like a butterfly in the orchard of knowledge. Three of them are fairly easy to grasp at the beginner’s level: we can all visualize how a grocery store, a hotel or a movie theater chain work. Three others have a higher educational bar: few of us have an a priori knowledge of flow cytometry, software-designed networks, or smart grid optimization tools. And even the “simpler” industries become complex beyond the basic level: there’s nothing trivial about figuring out how to manage the supply chain of a grocery store, optimize the occupancy rate of a hotel chain, or improve the average spending per spectator in a movie theater chain. As my synapses fire at decreasing speed, I pray that the wisdom of my years and the presence of younger brains around me cover for my reduced mental agility.

This is where the power of a cross-industry framework helps. In the co-creation business, one sees opportunities for new connections everywhere. Some people’s experiences can be connected together to form “chains of empathy” (for example, suppliers, employees and customers can all help a grocery store figure out what its strategy ought to be). One can fairly easily visualize data flowing across previously disconnected individuals and companies, and new insights being generated between them (in the medical, telecommunication or energy worlds, for example). I am like the Haley Joel Osment character in the movie the Sixth Sense: I see dead people.  Unlike him, though, I arrogantly think I can make them come alive.

If you one day see an old disheveled guy dragging heavy luggage and coming last off an airplane at midnight, do not feel sorry for me. My dream is being fulfilled.



We don’t know squat about economics

Monday, July 30th, 2012


We think of weathermen and stockbrokers as the two “often in error, but never in doubt” professions. Let me nominate a third: economists.

If stimulus or austerity policies worked, we’d know by now.  If Friedman and Keynes were right, our governments would long have adopted their policies, and our economies would be roaring like Formula 1 cars in the Monaco grand prix. The predictability of tax cuts or stimulus spending on economic growth has the reliability of Paul the Octopus forecasting soccer game outcomes: sometimes it works, and most of the time it doesn’t. Yet politicians everywhere hang on to these disproven theories as economic gospel.

What’s wrong with economics? To paraphrase Mitt Romney in one of his awkward statements, economics is people. Instead, we think economics is policies. From government to universities, we teach economics as a massively aggregated database from which we extract insights, then policies, at the level of a state or a country. This leads to lame assertions about interest rates, monetary mass, jobs, trade deficit, and vague concepts of rational expectations reputedly anticipating economic behaviors. If we understood the true causes and effects in the economic system, our Presidents would not be sweating the job numbers every month: they would tell us beforehand what to expect.

Do you wake up in the morning thinking about interest rates, inflation and trade deficit? Do you actually decide to buy a car, a house, or go grocery shopping on the basis of interest rate and inflation? Do you look at your Turbotax statement to decide whether the rise in the marginal tax rate just passed by Congress will authorize you to go to Wholefoods and buy the fresh organic tomatoes that day, instead of going to the regular grocery store where the tomatoes are cheaper? Of course not. Yet this is the micro-level at which the economy works. Unless we can begin to comprehend decisions at this individual level, we have nothing of value.

If I were an economist, I’d start by diving deep into understanding how five or ten of my neighbors experience the economy. I’d try to build a model of their income statements and their balance sheet, and figure out how they decide to patronize five or ten local businesses (say, restaurants, grocery store, day care, etc.), or why they decide to save and for what. I’d try to understand the economics of the five or ten local businesses my neighbors buy from, why these small businesses decide to expand and hire, or why they scale or shut down. If there were one or two big businesses in my local area (corporate headquarters, big hospitals, etc.), I’d try to understand how the success of those large businesses contributes to the local economy through local taxes and jobs. I’d then try to model how the local township or municipality benefits from all this, and what impact local people and businesses have on the finances of my town (school, public funds, etc.). If we could just model this microcosm of economic interactions, we’d have data on a real, living ecosystem of actual people and entities and begin to understand how a local economy is co-created through their interactions.

Would this be representative of the economy as a whole? Of course, not. There would be massive biases linked to local industrial fabric and wealth levels. To roll up this data into a state, national, or global economy, one would have to empower people to build their own model at the local level. Providing the structure and platform that allows this local modeling would be a great role for government, instead of pretending that it owns the economy or creates jobs. The role of government in economic policy should not be to build top-down expert models of the economy as a whole, but to empower local folks to build models of their local market and learn from their interactions.

Even more importantly, if we began to understand causes and effects at the local economy level, the “economic agents” involved would be able to do something about the economy, rather than passively describe it. Individuals could change their relationship to local businesses, for example, by forming communities around them. Local businesses could mobilize those communities by setting up platforms that better connect them to their local customers (for example, I’d love to rally a few of my local friends to help a local hotel improve a few things in their menu and rooms, and we’d collectively bring them our out-of-town business). If local officials were to facilitate this dialogue, this would do more to create jobs and get them reelected than repeating hackneyed Republican or Democratic theories of austerity or stimulus.

Unfortunately, the scarcest commodity in economics is humility. Witness for example the recent Business Week article on the discussion between Paul Krugman and the President of Estonia and on the value of austerity in running a country. I will not venture an opinion about who’s right or wrong in this debate, but getting rid of the condescension conveyed in this dialogue seems to me to be job 1.

I, for one, would like to understand the economics of my village.

The role of personal data in co-creation

Friday, February 24th, 2012

I work with a lot of analytically-minded people. These people can be colleagues or client, often both. Because analytically-minded firms tend to hire analytical consultants, they often lock in a powerful embrace that makes it hard for either of them to view the world through the lens of co-creation. These are the workshops from hell. Their first objection is often that co-creation feels like a random process of human discovery void of any data. Real men don’t do co-creation, they say. They formulate a hypothesis, gather some analytical data through research that proves the hypothesis is right, may test the idea with some customers or other relevant stakeholders, and then go build whatever needs to be built, be it a product, a process, or a strategy. Data drives the hypothesis, and humans are there to validate the hypothesis.

Co-creation follows a different logic. In co-creation, the hypothesis is the result of the engagement process, and insights are in fact generated jointly by the company and the people it engages (this is the uncomfortable part, the “letting go” part). What is even harder for analytical people to see is that co-creation is equally data-driven as the more traditional company-centric process, product or strategy design, but the data is of a different nature because it comes from the people themselves.

In co-creation, what comes first is the platform, typically a very crude prototype in the early stages. The primary role of the platform is to generate data that can be analyzed and structured, which will then guide the next iteration of design, in effect making the early users of the crude platform into co-creators of the next iteration. When Mark Zuckerberg created the first Facebook site, he devised a blunt instrument that allowed obnoxious Harvard students to rate the attractiveness of female students at Harvard. Was it the ultimate design of the platform? Of course not (and mercifully so!). But what he did was engage a group of campus students on a topic that defined a community (Harvard male students) and a basic platform concept that could evolve from this humble beginning. The first draft of Facebook allowed Zuckerberg and his team to engage into a creative dialogue that generated subsequent interactions, leading to new features such as communicating one’s social status, or sharing photos.  Since then, the platform has continuously morphed as a result of a natural Darwinian process where users co-evolve the platform through use.

So what is the role of data in this process of co-creation? Simply put, data is everything. The platform is a data machine. It records who comes to the platform, how long people stay on it, what people do on the platform, and what features they utilize. Platforms can be of a physical nature, like a store, or virtual kind, like a web site, but the metrics tend to be comparable. We’re all by now familiar with the “eyeballs” and “stickiness” metrics of interactive sites. Since co-creation requires the development of scale and efficiency (this is one of the main differences between collaboration and co-creation), co-creative firms need to develop scale and efficiency measures for the platform itself, e.g., how many interactions take place, how quickly they unfold, how many of them complete successfully, etc…

The most important data on a co-creation platform, however, is the constant qualitative bending of the performance model. Users continuously stretch the limits of the platform, imagining new interactions that would be of value to them. It typically starts by offering unfiltered feedback in exchanges between users (the site becomes the market research department). In some cases, on electronic platforms, this evolution occurs through downright hacking. The development of the Lego Mindstorms operating system (Lego robots) was one such case of hacking (ultimately grandfathered by Lego designers), and the insertion of Google maps into the Nike + runners community site is another case where two technologies were “mashed up” by users, rather than orchestrated by Nike web designers.

There are however a few fundamental differences in the data required by a co-creation strategy vs. a classic strategy.  The first difference is that in co-creation, the data is not resident in some data base of trends or financial analysis of competitors which any analyst can access. This data is by definition much richer than any “study data” because it is only accessible to the company that curates the platform. It is original data generated by real people using the platform who find it valuable to share their data. This data is often quite intimate. Nike users will share weight and running patterns.  Patients on the Patients like Me site share personal health and effectiveness of treatment data on profoundly debilitating diseases such as cancer or depression, because they believe insights and potentially new treatments will emerge from the sharing of this data. (Of course, no doctor or health insurance company would have the right to share this data, but patients willingly contribute it because they have a vested interest in it). People on find such value on the platform that they give the platform access to a large amount of personal financial information, because offers them insights they cannot get anywhere else.

Even more importantly, the data on co-creation platforms is alive rather than static. It is associated with individuals who are interested in continuously interacting with the company on new topics of mutual interest. At any one time, the company or the individuals can initiate the development a new branch in the tree of co-creation. They can generate a new algorithm on how to look at the data itself, e.g., suggest a new way of looking at how to train for Nike, a new treatment for Patients Like Me, or a new financial strategy insight for Not only is the data co-created, but so is the analysis and the insights that come from this data.

The possibilities become endless when data and the development of associated insights are driven by the self-interest of passionate people combining with the professionalism of a co-creative staff at a company. The future of analytics in business is not company-created algorithms accumulated through CRM or other third-party data bases. It lies in the co-creation of both data and insights by willing individuals interested in working with companies of their choice to generate new experiences for themselves.