For several years now, Messe Frankfurt has been working together with numerous internal and external experts to develop online business matching portals based on semantic technology. Henric Uherek, an internet expert at Messe Frankfurt who is heavily involved in these projects, explains the situation.
The idea of business matching sounds so simple. Is the technical implementation particularly challenging?
It depends on the approach that is taken. If one is satisfied with simply matching a trumpet manufacturer with a trumpet retailer for example, then the technical side is not very sophisticated. But if you want to imply and serve other interests based on the interest in trumpets, such as an interest in wind instruments or brass instruments in general, for which there are accessories like dampers, sound recorders, music racks and specialized music publishers, then you quickly reach the limits of normal search engines.
And is this where “semantic search technology” plays a role?
Yes, because semantics makes it possible to formulate fuzzy searches that include data, which doesn’t necessarily directly match the search criteria, but still may be interesting.
So semantics help computers think logically?
Semantics consists of teaching computers rules that they can then use to derive logical conclusions based on relationship structures. So for example if “Tom is Tim’s father” and “Jan is Tom’s brother”, then it is logical that Jan is Tim’s uncle even though the two are only joined by the following rule: “The brother of a father is the uncle of the father’s son”.
The advantage of this type of rule is that it can be used on any number of situations. However, this requires a structured environment that allows for this type of logic. In other words, it requires structured profiles that the system can use to determine which suppliers match a search with what percentage priority by way of logical conclusions. 100 percentage points are awarded for each full match in the relevant profile aspects, whereas less compatibility results in lower percentage values. At the end of the process, the accumulated values for high levels of relevancy result in a match. The conclusions and rules that we have stored in the system are the heart of the technology and therefore top secret.
So semantics requires a tremendous amount of preliminary work behind the scenes?
Absolutely. We have to intensively examine each industry and its peculiarities and anticipate how supply and demand work and communicate precisely there. For example, we work closely with experts from textile and the consumer goods industries. The logical rules form a foundation that is continually being developed further based on participant feedback and analysis of usage behaviour. The business processes used here generally remain hidden from organizers of actual trade fairs. We can learn a lot from our customers in this regard.
What role does participant engagement play?
It is important that the participant profiles be filled out to the greatest extent possible. Checkboxes and selection lists make this much quicker and more convenient than entering lengthy text to ensure precise matching results.
You are already using the semantic technology on several business matching portals. Will these serve to replace normal trade fair business in the future?
No, certainly not. The online world can only serve as a sensible enhancement to the offline activities. The internet is perfect for purely obtaining information. This is the case when making an initial selection of potential new contacts. It also saves time and money to be able to carry this type of research out from the comfort of one’s own desk. But it is like in other areas of life: it is also important to get out and personally meet the people behind the profiles. For this purpose, our business matching portal i-tex Texprocess, for example, provide an easy way to make appointments for personal talks at the trade fair event including features for booking a room on the trade fair grounds.