PLM Search Engines – A Present Possibility or Future Hope?October 3, 2015
One challenge that PLM users – decision makers, designers, business analyst, engineers – face regularly is – “How to reuse existing part / drawing to create a new one?” How to find such a part or drawing in our PLM system? This is exactly where the trouble starts. To filter out specific data from probably millions of data your PLM application has. Most of us who have used some kind of PLM application will be well versed how a typical search performs. It is no cake walk. Most of the search is heavily dependent on Part or Drawing Number and in more advance cases on a few common attributes or properties. How many times have we lamented—if only someone built a PLM search engine like google? Let’s analyze a few things first. With google web search you are not looking for an exact match. You are probably looking for best match. This is contrary to our situation where we are looking for an exact match. PLM enterprise search engines are not hosted on a server with infinite capacity and across geolocations. They are as per industry needs. Let’s summarize in the challenges in the table below.
|Web Search Engine||Enterprise PLM Search Engine|
|Users ok with best match||Users need exact match|
|Servers hosted across geographies||Servers hosted on limited locations|
|Super high capacity servers||Servers are generally enterprise application hosting grade with lower capacity.|
With the above challenges it is clear that an enterprise grade search engine has much more work to do with limited number of resources. What exactly should we expect from an ideal PLM search engine? To understand this we should first analyze why do we need to find the precise part or drawing in our PLM system? Designers are probably looking to re-engineer some part, regulators are trying to compare two similar parts, trying to launch a new product leveraging an old product – The possibilities are limitless. Hence search should not be limited to basic properties like name, type, version number, description etc. A more ideal search could be based on geometrical configurations or geographical parameters or purpose based parameters or a mixture of all of these. This way users will be able to related more to the search results. Is it really possible? Well let’s not blame search results entirely on the search engines. As users we are responsible for bad search results as well. In the golden words “What comes out depends on what gets in“. We cannot fill inaccurate information in the application and expect the search to be accurate. Trying to search a bolt when it has been entered as a screw will give us bad results. Search engines cannot make decisions. Well, the PLM Enterprise search engine journey has a lot of milestones to cover but has kicked off with high pace and some very interesting players are out there. Will write about them in my next post.