How can success be measured, what do we know about the connection between performance and recognition? To answer the question, Albert-László Barabási and his colleagues examined the works of scientists, artists and sportsmen. They found that „whereas performance depends on you, recognition depends on us”: while human performance is limited, success can be virtually unlimited. We talked with Albert-László Barabási during the Social Futuring Conference in March.

You mentioned in your lecture that success is a notion that is based on perception. How is this influenced by receptiveness? We have heard several stories about people who could not achieve success in their own times and the real value of their work was only discovered later.
That is true and actually that is why we say that it is better to be the last one to discover something rather than to be the first. Being the last to discover means that there is no point in discovering it again. Interestingly, many discoveries were made several times, but the moment comes when the community recognizes the fact. This happened in my life as well. It transpired that one of the mechanisms of scale-free networks, the so-called preferential selection had been published in scholarly literature fifty years ago. More precisely, György Pólya was one of the scholars who described it in mathematics. Although it did not appear in the context of networks, but it did exist and I was unaware of it. Preferential selection cannot be discovered anymore as it has been sufficiently discovered.

It is worth clarifying what exactly we are talking about in this context. When we think about success we should very carefully distinguish between performance, i.e. what the individual does and what is usually very concretely assignable to an individual or to a piece of work. And then there is success, which is the appreciation of the same performance by a certain community. These are important to distinguish because performance cannot always be measured precisely, but can always be assigned to an individual. Success, on the other hand, is a collective phenomenon, consequently has more available data points. Owing to the fact that it reflects the activity of not one, but hundreds, thousands and sometimes even millions of individuals, it is easier to measure and to forecast. Every time we talk about the activity of several millions of individuals, these tend to average, unify and render a mathematically calculable quantity. Success therefore is easier to calculate and to handle than performance.

What is the role of social networks in all this? Does our position within the network influence receptiveness to our performance?
Networks play a very important, actually a key role when performance is hard to measure. If performance can be measured precisely, like in the case of runners, it will define success unidimensionally. At the other extreme is art where it is extremely difficult to measure performance. The importance of the work and the artist depends on who else considers it important, which institutions, which curators deem it good enough. It is the institutional and curatorial network that creates the value. We are working on this too and can very accurately forecast the artists’ future career. This is because performance cannot be measured, and as a rule only the network counts. It is the network that mediates the success which can already be measured quite accurately, thus giving us considerable predicting abilities.

What about science, how well can it be measured?
Science is about halfway between art and sports. A large number of network effects influence what is important, what is worth researching, what the public see sas an achievement that is to be recognized. If, however, a result has been achieved, there are some sufficiently objective measures - at least in exact sciences - to help us decide whether it is true or not. If for example you and I both have a formula for the same phenomenon, sooner or later an experimental measurement or an empirical test will be carried out to decide for instance that your formula is valid, mine is not. And then you will have the success and not me. In fact, there is a degree of objectivity which measures the results after they have been achieved. Nevertheless, there is a very strong network effect in what we are researching, in which institutions and how we access the tools necessary for conducting the research.

In your book entitled „Bursts” you wrote that although there were some breaking points, the behaviour of individuals can be fairly well predicted on the basis of previous data. Can this be handled at the social level? Could we predict the behaviour of society?
It is essential to identify what phenomena we are talking about. Our findings presented in Bursts show that if we follow the movements of a man, we can collect enough information to be able to predict with an accuracy of 98% where he will be tomorrow at three. What made our prediction so successful? It was so successful and accurate because the movement of humans is very repetitive and the physical, spatial and temporal limits of where one can be at a certain time are given. We don’t go to the bank to withdraw cash at 2 am, as only ATMs are accessible at this hour.

If, however, someone moves to another location, chooses a new home or a new job, graduates from school etc., there are breaking points. Breaking points can also be brought about by technological developments, for instance a new underground line is constructed, I will opt for a different means of transport to get to my office. Or I will ride as a new bikeway was built. As a result of biking, many more places become accessible than before when I took the metro and flashed past underground. These breaking points are there and cannot be really predicted. We are looking at behavioural patterns that seem to be stable. That is, if someone changes his job or moves otherwise, will he return to a former behavioural pattern. There is a fairly strong returning effect.

At the same time it is very well known that the patterns observed at the level of the individual can hardly be applied to another individual or the same individual in another period of his life. That is why Asimov in his famous book Foundation froze scientific development in order to be able to make predictions. In fact, scientific development changes patterns of behaviour dramatically by creating possibilities that had not existed before.

Thus, it is not always possible to make predictions, but by collecting a sufficient quantity of data, we can decide how far the phenomenon in question can be predicted. There are phenomena, for instance human movement, where predictability is around 95% and there are phenomena where predictability is low. Irrespective of whether predictability is high or low, it can be quantified and handled with the tools of mathematics. This, however, does not mean that my tools and data are certain to predict what you will be doing tomorrow.

What are you working on currently, what is your upcoming project?
Two things are worth mentioning. One is that I am about to finish my book entitled The Formula- The Science of Success. It formulates in more general terms what I presented today, that is the difference between success and performance and how the former can be quantified. Writing the book has taken up much of my time in the past months, but I trust it will be ready in a few weeks. The Hungarian version is planned to be published in September in Budapest. The other thing that we are focusing on in the laboratory is to understand the connections between networks and diseases. Recently we have also been trying to grasp the effect of dietary habits on cellular processes and ultimately, our health. Many of us in the lab are and I guess will continue to be engaged in this area.

Baksa Máté