Tag Archives: data mining

Continuous improvement and personal development: the role of competition

Continuous improvement (CI) is often heard as the answer to moving forward in organizations that want to do better. Much of it is rooted in Toyota’s Production System (TPS) and targets the company’s faulty processes in iterative fashion.  In this post, I will try to briefly discuss a more personal aspect of CI.

The PI/CI approach

Organizations interested in improving performance eventually come to an understanding that this is a journey, not an overnight task.  They are willing to transform, and for this they engage the help of outside consultants, they train their workforce, they put processes in place where none existed or they re-engineer current ones, they strive to meet regulatory benchmarks, and so on.

This is all well and good as far as collectives go. And yet, the question I have is how to transform ourselves by using the same principles?  Is it possible to continuously improve as individuals?  Is there a clear progression?  How do we advance on this path?

At first sight, the obvious answer may be in getting more education.  One can never know enough, so the old saw goes.  One buys books, watches videos, travels to conferences when financially feasible, and, yes, one keeps taking tests and obtaining certifications.

Is this enough, or can we do more?

Motorcycle racing to the rescue

I will digress now, but only apparently, by recounting what somebody I know says. This person races motorcycles in what is known as ‘hare scrambles’. These are competitions for dirt bikes, dating back many years, in which riders race off-road and across open pastures, woods, and narrow, boggy trails at full speed, while negotiating a variety of obstacles.  This very successful rider has uploaded a number of action videos to YouTube, which many watch and follow. He races an older, heavier bike, which is also under-powered when compared to more modern ones used by the competition. Regardless, he performs quite well and places consistently.

When people witness success, they become curious as to how to replicate it and ask questions. More often than not they tend to focus on tools, as if something they may purchase can magically level the playing field or even give them a much needed edge.  Predictably, this rider was asked questions about modifications made to his bike.  These questions came so often that at some point he felt compelled to make and upload a video where he clearly went over all the ‘mods’, which added up to basically none.  One can only imagine the disappointment felt by those who were hoping for a quick fix to their real or perceived riding deficiencies in the form of a souped-up engine, a better and costly suspension, a lighter exhaust, and so on. Unfortunately for them, as i said, no significant mods had been made to the bike in question, which remained ‘standard’ as bought.

This was enlightening to many, or should have been. The message is clear: one can do well, indeed very well, even with sub-par tools.  This also begs the question, what is the secret sauce then? How can this be? Is it raw talent or is there something else at work?

The answer came in a reply by the rider to someone’s comment on one of the videos. He said: ‘Race with those faster than you, and you will get faster.  This is quite the pearl of wisdom, and if I were giving a talk in the context of PI and CI it would bear repeating: competing with your betters makes you better, and competing frequently ensures you are on a continuous path to improvement.

The role and benefits of competition

When it comes to PI and CI, can we improve our skills in continuous fashion by ‘competing’ with those who are more experienced or talented?  I believe much of our personal development does not occur in a competitive enough environment, and therefore we self-develop as described earlier up to a point, and then plateau.  We should keep in mind that competitions are not new, and are certainly not limited to the physical.

For example, contests are being held constantly in the data science arena by Kaggle and other organizations.  Individuals and teams from the world over strive to squeeze every last drop of performance out of existing algorithms to solve the data analysis problem posed.  The problem is typically defined by an organization, which then posts anonymized data sets for the contestants to analyze. The organization also determines the financial incentive to the winners.

Note that this is a welcome influx of money to the top scorers, but is an even bigger win for the company, because it manages to focus the best minds on a tough question at a far more modest cost than via in-house resources and with a much higher likelihood of success, given the breadth and depth of the talent pool. The main difference from the real world is that the data furnished are already ‘clean’, something that does not occur in real life when facing a new situation.

There is little doubt that taking part in these events tests and improves the skills of participants. Repeated efforts at solving a score of problems tunes those skills continuously.  Some of the benefits of competing in this fashion include:

  • collaboration (within your team, or via advice from others) and focused pooling of strengths
  • learning ‘by example’ and in hands-on fashion what works and what doesn’t
  • having to be ‘up to speed’ in the ever-changing world of machine learning
  • stretching the mind to solve problems outside one’s own comfort zone
  • receiving immediate feedback in terms of a score that depends solely on how well one’s algorithmic solution does relative to others in improving the current best solution, thus leaving politics out of the equation.

Complacency and stagnation are not good things.  It may be time for people and organizations involved in CI to think along similar lines to the above and try to develop more and better skills in people in a manner that exemplifies continuous improvement on a personal level.



Process mining vs. data mining

In this post, I will try to highlight key differences between process mining and data mining, and explain why we should be paying greater attention than we currently are to the former.

Continue reading


PPDM: people, process, data, mobility — a four-legged platform for change

In this post, I discuss why in my opinion process and data ‘go together’ as enablers for sustainable change. I leave the people and mobility parts for a later post.

Continue reading