Blog Post
by Dr. Christopher Dixon, GIFTED Faculty Member at the Chulalongkorn School of Integrated Innovation (ScII).
While I was a student in the UK, I worked part-time in a restaurant. As with all restaurants, the customer experience is of prime importance. For every complaint made by a customer, there are twenty silent, unhappy (possibly former) customers. Short of harassing customers with a questionnaire when they leave – a practice usually not recommended, it is difficult to identify the difference between the ‘silent happy’ and the ‘silent unhappy’. In essence, this is a problem of being unable to determine whether the customer experience was positive or negative.
Similar problems are also apparent in various engineering fields. For example, it is difficult to detect the growth rate of biological agents used in the production of pharmaceuticals, or ascertain the efficiency of an engine, or determine the extent of corrosion in a structural beam that could affect the structural integrity of a building. The problems described above have all been solved by engineers and scientists with considerable expertise in how these processes work.
As an example, it is difficult to create a sensor that would be able to quickly detect the concentration of bacteria within a solution – and it is even more difficult to develop an online sensor for this purpose. It is, however, much easier to detect the temperature of a solution, the amount of ‘bacteria-food’ entering this mixture and the amount of bacteria seeding the production. By combining these measurements with some basic understanding of how bacteria grow, we are able to create an “indirect” method to measure the amount of bacteria in a sample.
This idea of combining know-how with easy-to-access sensors goes by several names depending on who creates it. This idea is called a soft sensor – combining software and different sensors to sense something else indirectly. Still, depending on the field, this idea could be called a virtual sensor, a state observer estimator, and in some cases a smart sensor.
So how could restaurants use soft sensors to measure the customer experience? A low-tech (and arguably unintentional) soft sensor that restaurants have been using for decades is the loyalty card; a returning customer indicates a happy one; counting the number of redeemed loyalty cards could be an indirect indicator of customer satisfaction. A soft sensor should be unique to the process it is attempting to measure, so understanding that process is crucial. Rather than lamenting what cannot be measured directly (or is difficult to measure), let’s use what we have intelligently.
READ ALSO: Our Previous Blog Post: Does a Smart City Need a Smart Mayor?
