Predictive algorithms as assets

One thing I'm most passionate about is that the world would be a much better place if we used more predictions and if we made better predictions. We'd better manage our economies, companies, natural resources, traffic, pollution, our machines, relationships, careers… actually the list of what wouldn't potentially benefit would probably be shorter, partly because I can't think of anything right now.

I don’t mean this statement as in "if only we could make better predictions"… I mean we can make better predictions right now, but we're not investing enough of the right people, time and technology to achieve them. We're grossly underachieving, and the world is missing out, in some cases very badly.

It is of course increasingly widely known that combining the right data, domain understanding, technology and algorithms can enable the generation of new and better predictions. Domain understanding is being somewhat overlooked at present as people realise value from combing more and better data with better algorithms and technologies; remarkably, that can enable you to predict something better without necessarily improving your understanding of why you can predict it better. However that build up and use of understanding can also enable improved predictions and should be formally developed as part of any on-going effort to improve predictions.

All this leaves me predicting the next phase for prediction in society. From advances in tech came advances in the availability of data, as also came advances in our ability to analyse those data and extract insights and predictive abilities. Now it's passé to talk about data as having intrinsic value and being the new oil. As expected, with the proliferation of data and technology and algorithms, we're seeing the proliferation of organisations that specialise in mixing those things together and producing new insights, new predictions, new information. All great! I think next is the proliferation in predictive algorithms as a more widely available asset and service.

Right now the actual process of prediction is hugely hidden away. It so much depends on the data, technology, domain and those who did the magic bringing them all together. However just as now we're getting institutes for open data, for data sharing, for domain-specific data hubs, I think in the future we'll next be getting that for predictive algorithms. For example some sort of open institute for predictive algorithms would provide access to algorithms in a way that enabled the generation of new predictions to be made so much faster and better than before.

I think predictive algorithms need to be represented and recognised as an asset class at least as important as the data, technology and domain understanding used to construct them.