Using Search Analytics in SharePoint 2010

  • 4,859 views
Uploaded on

This slidedeck is from our surfray webinar on Search Analytics in SharePoint 2010. The presentation contains some search theory and an introduction to search analytics reports in Sharepoint 2010. It …

This slidedeck is from our surfray webinar on Search Analytics in SharePoint 2010. The presentation contains some search theory and an introduction to search analytics reports in Sharepoint 2010. It also covers simple techniques for improving search based on the analytics.

More in: Technology , Design
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Be the first to comment
No Downloads

Views

Total Views
4,859
On Slideshare
0
From Embeds
0
Number of Embeds
0

Actions

Shares
Downloads
53
Comments
0
Likes
3

Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide
  • Hi, Welcome to our webinar on Search Analytics in SharePoint 2010. My name is Robert Piddocke and I am author of Pro SharePoint 2010 Search. I am also working on a book on FAST for SharePoint and have been working in the enterprise search business for over a decade. Today I’ll be presenting a little on what makes search good and some theory about search and then digging into Search Analytics in SharePoint 2010. I’ll also show you how to action those analytics and give some best practices for improving search.
  • Your search engine is one of the few truly interactive tools on your site. Most people look at search as a black box but it really can tell a great deal about the users expectations and their experience in SharePoint. There is a clear difference between browsing for content and searching for it and many users will use search as their main navigational tool. The challenge they face is, of course, finding information. That information can be things they know exist in SharePoint, information they are currently working with or information they themselves have created. The search engine is a tool people hope will respond to direct questions for content in SharePoint. It is important therefore to make sure that the conversation is not completely one sided. This is where search analytics come into the picture.Another important thing to note is that SharePoint search is not Google – AND WITH REASON! I’m sure we’ve all heard the request to make SharePoint search more like Google search. Well, global search and local search are different because the users expectations are different. Users mistakenly want google in the organization without realizing that their expectations for google are drastically different than their expectations for SharePoint search. Google users are looking for one of many items. If they search for ‘shoes’ on google, they will be happy to get 1000 results of 100,000,000 possible pages about shoes. The search is usually not that specific and any set of results is more than enough. Within the organization users want to find the one unique document or piece of information in a document out of thousands or millions in the organization. Finding that single document requires a much more surgical approach.
  • Documents all have a meaning and a reason for being created in the organization. I decide that we need a document about a particular topic or I am asked or told by my boss or colleagues to make a document about a specific topic. I do this by creating a document with particular themes in it that relate to that topic. Those themes are described by terms I decide to use. Sometimes those term choices are also guided by company policy. Users, in the same way, try to find the content they want by searching for terms they think match those themes that will surface the right document. Often there is a disconnect between what people are writing in documents and what other people are using to find them. This can be because of different terms or jargon or just simple spelling mistakes.One of the biggest problems with search is that there is no monitoring or reaction to this disconnect. This is where search analytics comes into play.
  • There are 4 basic behaviors of search users. These behaviors can be matched with certain levels of success for the search engine. Users search for a document or piece of information they are looking for, get a result set, choose a result they like (usually on the top of the result page) and it is the document they are looking for. They check it out, edit it, copy some information from it, or maybe forward it to a colleague. Partial Success – users search for a term they think will get the document they want, they get a result page, they click on a result link that looks good but don’t find what they want. They go back, search again, paginate, add terms or abandon the search. Unsuccessful with results – the users search, get a result set but nothing in the result set looks good so they search again, paginate or abandon the search. Sometimes they can use filters or refiners to dig into the result set and improve it.No results – users search for a term they expect to find a document with and get nothing. This behavior is more common than you might expect. These are the 4 behaviors we can see with search analytics, the two bottom being the most dire and requiring action.
  • Most of us have heard of the long tail of search proposed by Chris Andersen where the cumulative amount of searches on sites such as Amazon that are for unique items are more than the most popular searches. Luckily for the enterprise, the tail is not very long. In contrast, we usually see a fat head on this graph. What that means is that most users search for few terms repeatedly. This may vary over time but at any point, there are likely 20 or so terms that are very common and popular in SharePoint search. This makes it easy for us to action those terms and improve search greatly.
  • Typical problems for search that lie outside of this but that we can often identify with search anlaytics are listed here. The largest two being garbage and noise on the site. If you have tons of unnecessary content indexed from file shares or other sources that are not actively being used or required, the search results can be filled with noise that will hamper good search. Poor metadata also will deter good search. There are other factors that make search results bad as well, including useless and duplicate content, bad document authoring, bad pdf titles because the titles are generated from the word document’s file name and simple misunderstanding of search users needs. Analytics can help improve these or bypass the need to make major content improvements with the use of keywords and best bets. Other ways to fix these problems can be seen on our recorded webinars at www.surfray.com
  • Fixing these problems that we can find in the analytics in SharePoint is really quite easy. First of all, you need to access the site collection web reports. Then check them regularly for about 3 months. Make best bets for the most failed queries with the content you think people should get. Poll your users if you don’t know what to put there. Check the improvement on a regular basic and follow the cycle here: Search, Analyze the search statistics, and optimize them with best bets and period content improvement.
  • Let’s look at the analytics in SharePoint 2010. Go to site settings in the site actions menu. Then choose site collection web analytics reports. Then look at the different reports. Failed queries is the most important and the best starting point.
  • The failed queries report is you friend! It is the best place to start to get an overview of the failure points of search. Check the queries that returned no results and the queries that had a high percentage of abandonment. These are the places to start adding best bets. Click on the search keywords menu on the left to get to the search keywords page and make a best bet. Take the top terms and create a keyword and best bet for each one. Other reports of interest are the top queries page which you can compare with the page views report to see how much search is actually used. After several best bets are made, you can monitor their use and look for new suggestions. This will allow you to improve the best bets over time.
  • Thanks for attending my webinar. Please feel free to send me any questions or add me to twitter to get updates on blog posts and future webinars.

Transcript

  • 1. Search Analytics in SharePoint 2010Presented by: Robert Piddocke Author: Pro SharePoint 2010 Search Working with Microsoft FAST Search Server for SharePoint rcp@surfray.com robertpiddocke @rpiddocke www.surfray.com
  • 2. What is Good Search?Basic challenge: Finding information• Browsing vs. Searching• Challenge: to find information as quickly and easily as possible• To respond to direct questionsLocal/Site Search is a different beast than Global Search• Global search users often look to see IF content exists, and WHAT exists. They also look to locate sources of information instead of detailedinformation.• Local users expect the information to exist and are frustrated when theydon’t find it. They use terms that are specific to the site’s business. Theysearch for content they are working with or have created themselves. www.surfray.com
  • 3. Search Engine Theory• Documents (usually) have meaning and purpose – One or a number of themes• Authors describe these themes in their content with terms• Users search for ‘themes’ by using terms• There is often a disconnect between the terms used by authors and users for the same themes• Your content may have these themes but not the language or viceversa – spelling errors, corporate language, noise etc.• Too much time is spent anticipating instead of learning and reacting www.surfray.com
  • 4. 4 Behaviors of Search UsersSuccess – The user searches, gets a result page and chooses the page with theinformation she wants.Partial Success - The user searches, gets a result page and chooses a page, but notthe best one. She clicks back and tries again.Unsuccessful with Results - The user searches, gets a result page but sees nothingshe likes. She searches again or leaves.No Results - The user searches, gets no results, searches again or leaves. www.surfray.com
  • 5. How long is your tail? The majority of search terms represent unique searches but only a few searches represent a third of all users. 5722 27000 12% 55% 100+ 99 - 2 <2UsersSearching ...49201 Unique Terms
  • 6. Typical Search ProblemsFailure points• Garbage in, Garbage out• Poor meta data• Too much noise – Irrelevant content on top – Useless content – Duplicate content• Bad document authoring (pdf, MSOffice)• Misalignment with/misunderstanding users’ conceptions• No improvement over time www.surfray.com
  • 7. What to do?ACTIONS• Use Site Collection Web Reports• Check statistics weekly in the first 3 months•Make 20-50 Best Bets• Check monthly and adjust• Improve titles descriptions• Add metadata• Map exisiting metadata
  • 8. Analytics in SP2010!1 2 3
  • 9. Failed Queries are your friend!
  • 10. Please feel free to ask questions in the chat!Presented by: Robert Piddocke Author: Pro SharePoint 2010 Search Working with Microsoft FAST Search Server for SharePoint rcp@surfray.com robertpiddocke @rpiddocke www.surfray.com