Search Engine Results

Notes from the email on search engine ranking from Emma to the IASC.

On the old site in the "search results" you were essentially querying for the file NAME stored on disk, and not performing an actual search of the content of the documents.

This means when you were searching for: http://www.humanitarianinfo.org/iasc/pageloader.aspx?page=content-search-spfastsearch&query=IASC%20TOR you wanted the results to be displayed as IASC TOR - 11Feb2014? This is actually the name of the file that's saved to disk, and we display the administrative TITLE for the document (so that you don't end up with weird names that no one could guess to search for except the original author and "No title" results... although the file NAME does show up in the text snippet description, so it's not completely missing). HOWEVER, by implementing a real search engine, which actually reads the contents of the documents, we don't display the name of the file in the same way, we use the TITLE of the file as it was stored in the admin area of the site (http://webapps.humanitarianinfo.org/IASCadmin/pageloader.aspx?page=document_edit&documentid=4491). In this case if we use the administrative TITLE (IASC Terms of Reference 2014) it comes up as #3 in the search results after the bias adjustments -- this is also what will display on the node page where you could add a description with more keywords to help improve the ranking for this particular node (more about this in the next section).

If administrators want to query for a file name, please use the following administrative view: http://interagencystandingcommittee.org/admin/content/file

Creating high value content for the search engine

Now that you're using a real search engine, you're going to need to start thinking like Google does. (Except with a much smaller budget.) For example: "Holmes" is no longer an important person, but just a word in a website. So we need to see what would be an example of "high value content" for the word "Holmes". Currently we assume that meetings and documents are important. Are meetings more important than documents, or the other way around? Unfortunately the answer can't be "it depends!". So for #5: what would be an example of four pieces of content which should appear in the positions #1-4 on the new site? We can use this information to adjust the biases...

For IASC Principals: showing up in the #2 slot is excellent. No adjustments here to be made. Again, this isn't a database query, but rather a complex search algorithm. Currently IASC Transformative Agenda uses the phrase "IASC Principals" more than the IASC Principals page does *AND* the Transformative Agenda puts that keyword into HEADINGS! (mega important to the search engine as it usually denotes important information; we toned down their importance, but please be aware of how they can impact results). So if you want the Principals page to appear before the TA page, you'll need to adjust your content accordingly. (Now you see why SEO is such a big industry! You have to be tricksy.)

The same concept applies for the minutes of the Working Group. What *should* the search engine have found? Do those documents contain the word "minutes" in the title, or a number of times throughout the document to make them "high value" for those terms? Are they even referred to *as* minutes? In a quick gander through the Working Group meetings it seems like the term "Summary Record" or "Agenda" is used, but I don't see "Minutes" used anywhere....sample URL: http://interagencystandingcommittee.org/node/124