IBM and Getting Traffic

June 19, 2009

A happy quack to the reader who sent me a link to Bill Hunt, who is on the Board of Directors of the Search Engine Marketing Professionals Organization or SEMPO. The case study is about IBM. I have written about the frustration I encountered when using the IBM Web search system. This case study makes it clear that IBM’s focus is on traffic to its Web site, not making the Web site coherent to a human looking for information about a FRU or replacement power supply for an i series machine.

Mr. Hunt’s presentation can be obtained via Slideshare.net. You will be able to view the presentation in Slideshare’s microscopic font. To download the presentation, you will need to register, log in, and navigate to this page: http://www.slideshare.net/OMShare/enterprise-search-ibm-case-study-by-bill-hunt?nocache=8495.

Several comments:

  • IBM, according to Mr. Hunt, has “many cooks”. The site is dense, complex, and focuses not on providing information to customers, but upon selling new ideas and products in my opinion. Frankly, I don’t know where to look on the page. If you don’t know the official, current product nomenclature, it can be really tough to locate technical information.
  • Because the focus is on, according to the presentation, “driving increased visits and qualified traffic to IBM.com from external search engines,” I think the lousy search system shows where the Web site manager’s emphasis is: traffic to IBM from referrers. The idea that a person would navigate to IBM to look for something is not on the Web site manager’s radar. For that reason, I just don’t agree with this statement in Mr. Hunt’s presentation: Search is integrated throughout our programs as part of an end-to-end approach supported by back end systems. Web site search doesn’t work very well. SEO experts’ definitions of “search” are usually based on assumed understanding. Best to un-assume when the addled goose thinks about search.
  • Mr. Hunt provides data that shows how muddled the meaning of search is. A chart called “Building the Case” asserts that if your site does not appear high in a search engine result list, money goes elsewhere. IBM has to get “search” right – search means appearing in a results list, remember – or the $100 billion company loses some money. The chart on slide 5 uses a figure of $1,470,620. I am not sure how Mr. Hunt gets from 180,808 searches to $1,470,620, but the number looks precise, I suppose.
  • Mr. Hunt suggested on slide 7 that a “Site Wide Algorithm Compliance Audit” is a useful step. Sounds like a consulting gig to me. Slide 8 shows how templates can be twiddled to make them search engine friendly. I think the idea is that concrete nouns make content processing useful in handling user queries, but this type of stuff reminds me of English 101 injunctions.
  • Slide 11 shows a mind numbing amount of data. I have to confess, I have zero idea what a PLP is, what a Brand page is, and what is means to be right 17 percent of the time. This sounds like the surgery success rate of a doctor in the US in 1820.

If you are interested in what appears the steps IBM went through to rank high in Google results lists, this is the presentation for you. For me, I would be happy if IBM put a decent search engine on its Web site (maybe Exalead or Google Custom Search), standardized reference to product names, and refresh the side index so stuff the PR department pumps out is findable on the IBM.com Web site.

I went through the SEMPO SEO case with an open mind, but I am fearful that a quest for spoofing the Google has gotten in the way of making the site useful to humans who are or were IBM customers. Oh, FRU is field replaceable unit. If you can’t do it yourself, IBM tech support will roll for nearly $800 not including on site time to plug the FRU into a slot. You have to get to $100 billion any way you can.

Stephen Arnold, June 19

Comments

Comments are closed.

  • Archives

  • Recent Posts

  • Meta