Microsoft Fast Arrow Electronics Parametric Search
April 24, 2009
In April 2008, BNet reported here that Arrow Electronics signed on with Fast Search & Transfer for the deployment of the Fast ESP (enterprise search platform). Today, an observant reader sent me a link to a story dated April 23, 2009, that appeared in 4G Wireless Evolution here. The title was “Arrow Electronics Launches New Features to Online Search Engine.” The newly enhanced system offers:
new online features providing greater access to product data and simplifying the search and ordering process within its expansive electronic components database. Building on the strength of FAST (Microsoft), Arrow’s new parts search engine, the enhancements represent the next step to provide greater tools and information via Arrow’s online resources.
The 4GWE story added:
In fall of 2008, Arrow launched FAST, offering enhanced site functionality, a greater range of user options and improved search speed and accuracy. Since the launch, search effectiveness on components.arrow.com has increased by 75 percent – giving customers faster, easier access to Arrow’s expansive parts database with readily available product and inventory data, enhanced filtering and cross-referencing capabilities.
Arrow appears to have indexed content about electrical products from about 800 suppliers and 120,000 original equipment manufacturers. There is scant information about the size of the content indexed. I navigated to the Arrow site here and ran some test queries. My initial reaction was that the system seemed snappy. As I clicked through the result pages, I saw output like that shown in the screenshot below for the query capacitors:
I clicked on the PDF logo for the first result, viewed that document, and tried to enter the following phrase “Monolithic Ceramic Capacitors”. I entered that phrase in the search box. What I discovered in that the search box only accommodated a portion of the phrase, 25 characters to be exact. This type of query constraint has been common to parametric search systems for decades, but I was surprised to encounter that hard stop.
I then trimmed the query to “monolithic ceramic”. The good news is that the first hit was to a PDF containing that phrase. What surprised me was that the full result set returned 87 items, and I had insufficient data to determine the features of each capacitor without downloading the PDF. I could click the hyperlink in the result set and get the pricing and a function to buy the product in multiples of 3,000 pieces.
For a buyer with deep familiarity with specific electronic products, the system will probably handle most look up and ordering needs. But as a buyer asked to procure components for an unfamiliar application, the set up of the system requires quite a bit of downloading, reading, and requerying.
The good news is that the system works and has been upgraded at least once in the last 12 months. The problem is that compared to the “mashup” type of presentations that are beginning to make their appearance on such sites as Kosmix.com and Navgle.com, the Arrow approach is dated in my opinion.
I was expecting to see a richer presentation of the data. Presumably the content resides in a blend of structured and unstructured tables. Hyperlinks are already in place because I can click to see a PDF or click to see a price. The problem is that I have to do those clicks at all.
I wanted to know more about Arrow. I clicked on its Investor Relations tab and I discovered a page of information but no search box. I was surprised that the Microsoft FAST ESP technology either had not been deployed against unstructured content on the firm’s content management system or its Web site. Here’s the page that begs for a search box:
This portion of the Web site has received more design attention which leads me to believe that the search function is disconnected from the “soft” information on the Arrow Web site. The visual difference is jarring to me, but the omission of a search function is a design and usability mistake.
Congratulations to Arrow and the Microsoft FAST team. I anticipate more work will be done on the Arrow site. I will check back to see if the 25 character limit is addressed, the interface issues that put the burden of assembling needed data on the user, and the omission of the search box on a text heavy page. I have no information about cost, hardware infrastructure, and the size of the index and its refresh cycle. If you have some data to share, use the Comments section of this Web log to illuminate the dark corners of these topics.
Stephen Arnold, April 24, 2009
Comments
One Response to “Microsoft Fast Arrow Electronics Parametric Search”
-
- Subscribe to Beyond Search
Feature archive
News archive
-
Steve,
Very interesting analysis and good points on limits and awkwardness of having to open all those PDFs to see the actual pages that contain the query string.
A couple of weeks ago, I was briefed on a Finnish product, Documill, that appears to solve just that problem. It sits on top of any search engine and delivers an enhanced version of search results. It presents a visual display of the actual pages from the source documents that contain the query string. Perhaps you already know about it but your readers might find it an interesting solution. It might significantly boost the usefulness of the Arrow application. Lynda