Aggregation and Approximation

Fallacy of aggregation. One of the key things we need to remember which is very unique about the skills space is that we cannot generalise and focus on aggregation. In most industries we tend to look for data slicing and dicing on percentages on different variables.
However, in the skills area every element matters even if they are in ones and twos.
For example, in a company there may be 1000 employees and yet, in the whole company there is only one systems administrator or one IT security analyst.
If we aggregate and look at the skills SysAdmin skill will be just .1%. This may lose significance numerically but for the company this is of high impact. The roles of each of these ones and two skills profiles are very critical.

We cannot, for example, draw a data visualization to show where percentages of people are occupied, hinting at where people should consider getting occupied, like we would say for stocks, for example. There are so many factors as to why somebody chooses the occupation they are in. If we need to help people we need to get down specific skills areas and see what is happening in that narrow cluster.

Approximation does not work. Most job recommendations that we get from job portals are based on matching at a very high level. Just because I have mentioned that I am a CEO, I often get recommendations for CEO related job posts. There is a good level of filtering on locations (which is a simple filter), but since we are not able to get granular information, recommendations are made on approximations. This is quite frustrating particularly in an era where other fields are getting more and more exact or individualist or personalised.

The problem is that probability is not enough and exactness is required. That is, it is fine to show a product to 1000 people even if we know it may only be relevant for 10% of them, because we show it to 1000 people to find out who the 10% are. In the case of talent, this can be extremely irritating. I would rather have two jobs that are close to me being shown rather than have 10 jobs of which only one is useful to me.

Also, when people are looking for specifics, alternatives may not work in the case of skills if the alternatives are not proper.

PreviousIssues in Treating Skills as Discrete Entities NextOwnership of skills

Last updated 3 years ago

Was this helpful?

Aggregation and Approximation

Fallacy of aggregation. One of the key things we need to remember which is very unique about the skills space is that we cannot generalise and focus on aggregation. In most industries we tend to look for data slicing and dicing on percentages on different variables.
However, in the skills area every element matters even if they are in ones and twos.
For example, in a company there may be 1000 employees and yet, in the whole company there is only one systems administrator or one IT security analyst.
If we aggregate and look at the skills SysAdmin skill will be just .1%. This may lose significance numerically but for the company this is of high impact. The roles of each of these ones and two skills profiles are very critical.

Approximation does not work. Most job recommendations that we get from job portals are based on matching at a very high level. Just because I have mentioned that I am a CEO, I often get recommendations for CEO related job posts. There is a good level of filtering on locations (which is a simple filter), but since we are not able to get granular information, recommendations are made on approximations. This is quite frustrating particularly in an era where other fields are getting more and more exact or individualist or personalised.

Also, when people are looking for specifics, alternatives may not work in the case of skills if the alternatives are not proper.

PreviousIssues in Treating Skills as Discrete Entities NextOwnership of skills

Last updated 3 years ago

Was this helpful?