MarkoHurst Site Search Analytics Emetrics Madrid

Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 46

Search Analytics For Your Site:

Una Conversación Con Tus Consumidores

eMetrics Madrid, Espana 2009


Keynote: Marko Hurst

Marko Hurst
Me

Book: Search Analytics - Conversations With Your Customers


• Anticipated release: December 2009
• Book website: RosenfeldMedia.com/books/SearchAnalytics
• Co-Author: Lou Rosenfeld

Consultant, Author, & Speaker


• User Experience
• Web Analytics
• Behavioral Targeting: (Machine Learning & Neuroscience)

Blog: MarkoHurst.com “Insightful Analytics”


Twitter: MarkoHurst
Contact: [email protected]

Marko Hurst
Before We Begin

Audience Survey & Search Info

Marko Hurst
Audience Survey

Question 1
Who currently has ‘search’ on your company website or
on your intranet/extranet?

Marko Hurst
Audience Survey

Question 2
How would you rate your experience with your
onsite search engine?
Criteria of success for a well performing search engine means that it
delivers relevant results you expected on the first page without
refining

Defies statistical probability


You are pleasantly delighted and use it often
Less dangerous, but about the same odds as Russian Roulette
There is a greater chance of finding a new continent than content
You’d rather contract an incurable disease than use it again

Marko Hurst
Finding Information Online

Three (3) ways to find information


• Browse (navigation, links)
• Ask (help - automated, live person)
• Search (query input field/box)
SEARCH due to the popularity of search engines (Google,
Yahoo!) is the #1 most used feature to find information
online
• BUT, once a visitor gets to your site, ‘browse’ is the #1 method
(with some exceptions)
• About 1/3 of all visitors use search, but sites range from 5 - 70%*
• Why does the information seeking behavior change onsite?

* Observations from client engagements


Marko Hurst
Over Promising and Under Delivering On Search Results

Onsite search ‘mildly’ sucks. Why?


I.T. only project
Implemented without consideration of the user
No optimization program in place
Lack of experience
Search is believed to be a ‘turnkey’ solution that is
simply ‘plugged in’
• <evil grin> admit it, you thought it was that easy <evil grin/>
• Getting search right is hard work and ongoing
• Be weary of anyone who says anything to the contrary, they are
probably trying to sell you something

Marko Hurst
Unfortunately For You As A…

VISITOR that means you are….


• left unsatisfied
• annoyed
• unfulfilled
• sometimes feeling stupid
BUSINESS that means you…
• didn’t think about your entire visitor/customer experience
• lost a potential customer, perhaps forever
• have reduced your customer satisfaction rate
• your site isn’t a usable as you ‘think it is’
• you are not listening to your what your customers are telling you

Marko Hurst
A Shining Light of Hope

Book Objective

Provide simple and effective Site Search


analysis techniques for user experience &
web analytic professionals of all levels to
better understand their visitors in order to
deliver a superior experience

Marko Hurst
Benefits & Expectations

SSA* produces actionable insights


• Techniques used are about analysis, NOT reporting (and there was
much rejoicing!)
• For some, this like reaching Nirvana
• For others, this is like opening Pandora’s Box
To achieve the benefits of SSA expect to:
• Change site design/layout
• Change keywords, copy, metadata, labels, navigation,
taxonomy, etc.
• Add &/or remove pages
• Change your SEO & SEM
• And much more!

* SSA = Site Search Analytics


Marko Hurst
Agenda

What Is Site Search Analytics?

How We Find Information

Anatomy of Search

SSA Techniques

Q&A

Marko Hurst
What Is Site Search Analytics?

And What It’s Not

Marko Hurst
SSA Is Not About Improving…

Marko Hurst
SSA Is About Improving…

Public facing websites, intranets/extranets, mobile, etc.


Marko Hurst
A Few Onsite Search Engines You May Encounter

Marko Hurst
How We Find Information

Information Foraging

Marko Hurst
Information Trail

Humans forge for information similar to how animals


forage for food
• Move outwards in a direction
we think (predict) will
provide the expected results
• Continue on a path as long as
we ‘smell’ signs that we are
still on the correct path
(information scent)
• When we no longer smell
those signs we retrace our
path or find a new path
entirely where the ‘smell’ is Flickr Photogrpher : a walk on the wild side

stronger, which we remember


for next time (recursive
learning) to better predict
where/where not to go

Marko Hurst
Information Scent

Information scent is how people evaluate options they


encounter looking for information on a site
• Strong information scents are
good at guiding users to the
content they want/need
• Weak information scents
cause visitors to spend more
time evaluating options and
increase the chance that they
will select the wrong option and
be forced to backtrack or leave
entirely Flickr Photogrpher : RaffertyEvans

Marko Hurst
How Humans Find Information Online

Three ways of finding information


– Browse
– Ask
– Search

Marko Hurst
Browse (Navigate)

Marko Hurst
Ask

Marko Hurst
Search

Marko Hurst
Where To Begin

Getting Started

Marko Hurst
The Basics

Analyzing data
• Zipf Distribution
• Excel (spreadsheet) skills
• Low / no budget software
• No need for code or higher mathematics
Where to get data
• Log files
• Search / Web Analytics
A mindset / desire to improve your site, increase ROI, and deliver a
better user experience
• Outside the scope of book :)

NOTE: everything I show you is 100% technology agnostic

Marko Hurst
Zipf Distribution (The Long Tail)

3 characteristics: head, torso, & tail

Thick
head

Flickr Photogrpher : hjallig


Middle
Torso
Long Tail

Marko Hurst
Data: Search Logs (Google Search Appliance)

Critical elements in red


IP address, time/date stamp, query, and # of results
XXX.XXX.XX.130 - - [10/Jul/2006:10:24:38 -0800] "GET /search?
access=p&entqr=0&output=xml_no_dtd&sort=date%3AD%3AL
%3Ad1&ud=1&site=AllSites&ie=UTF-8&client=www&oe=UTF-8&proxystyleshe
et=www&q=regional+transportation+governance
+commission&ip=XXX.XXX.X.130 HTTP/1.1" 200 9718 62 0.17

XXX.XXX.X.104 - - [10/Jul/2006:10:25:46 -0800] "GET /search?


access=p&entqr=0&output=xml_no_dtd&sort=date%3AD%3AL
%3Ad1&ud=1&site=AllSites&ie=UTF-8&client=www&oe=UTF-8&proxystyleshe
et=www&q=lincense+plate&ip=XXX.XXX.X.104 HTTP/1.1" 200 971 0 0.02

XXX.XXX.X.104 - - [10/Jul/2006:10:25:48 -0800] "GET /search?


access=p&entqr=0&output=xml_no_dtd&sort=date%3AD%3AL
%3Ad1&ie=UTF-8&client=www&q=license+plate
&ud=1&site=AllSites&spell=1&oe=UTF-8&proxystylesheet=www&ip=XXX.XXX.
X.104 HTTP/1.1" 200 8283 146 0.16

Marko Hurst
Data: Search Engine / Analytics

Data collection & options


vary by vendor
Data collection is typically a
separate step if you want to
combine it with web
analytics
• Google & Omniture are the only
major analytic vendors to have
built-in search data capabilities
(NOTE: Omniture’s Search is an add-on
product)

Marko Hurst
Improving Your Site Search

SSA Techniques & Search Behavior

Marko Hurst
The Anatomy of Search: Search Components

Six (6) components comprise a single experience

1 2 3 4 5 6
Flickr Image : Peter Morville
Based on original image from “In Defense of Search” by Peter Morville

Marko Hurst
Component 1: Visitor (User)

When / Where Do Visitors Search?


Most often when a visitor becomes frustrated with
browsing (i.e. your design, navigation, information architecture, whatever)
eCommerce and Government websites search is often
the first choice
• This behavior also occurs when a visitor ‘knows’ what they are
looking for

Some visitors use search as their first / main method

For more on this, see Peter Morville’s “Search Patterns” IA Summit 2008
Marko Hurst
Component 1: Visitor (User) Search Analysis

When /where did your visitors initiate search from?

Marko Hurst
Component 2: Query (Keywords)

When a visitor users onsite search they are speaking to


you their Natural Language
• They are confessing their needs & desires to you hoping you can
help them

This is your chance to have “Una Conversación Con


Tus Consumidores”
• REMINDER: Conversation = Good. Monolog = BAD!

Marko Hurst
Component 2: Keyword Analysis

What are your visitors looking for?

Marko Hurst
Component 2: Keyword Analysis

Trends

Marko Hurst
Component 3: Search Interface

Minimum: search query box & search button

Sometimes a filter will also be used

Marko Hurst
Component 3: Search Interface Analysis

How many characters should your query box display?

Marko Hurst
Component 4: Search Engine

While the Search Engine is an essential component…


• Options and details vary by vendor
• Common features: Ranking, best bets, stemming, facets, weighting,
most frequent, etc.
• Opening the ‘black box’ is beyond scope of book -- and I
promised you didn’t need to code & be a mathematician

Marko Hurst
Component 5: Content

Search is about getting visitors to CONTENT


What type of content can you improve? (Hint: all of it)
• Metadata
• Navigation
• Labeling
• Ontology
• Taxonomy
SEO / SEM people LISTEN UP
• SEO is NO LONGER about being #1. It’s about getting visitors to your
content!
• This is great place for UX & SEO to work together NOT against each
other
• Perceived differences should be fixed
• Usability testing, copywriting, & SEO are about content bring them together

Marko Hurst
Component 5: Content Analysis

Power of CONVERSATION: Are you listening or ignoring your visitors?


What content / products your visitors are looking for?
• Do you not have it? Or can’t they find it? Maybe you should add /
remove products?
• Your visitors may be speaking a language you don’t understand
• Worse you may be trying to speak to them in a language they don’t
understand
• Labels, navigation, taxonomy, metadata, SEO, SEM

Surveys: tie attitudinal & behavioral data together


Look for relationships between content http://4q.iperceptions.com

• Informs you ontology & metadata, as well as SEO / SEM


You can look at the data, but this is a GREAT place to explore and
play with your data

Marko Hurst
Component 6: Results (SERP)

SERP (Search Engine Result Page)


The (inferred) quality of your results can be determined
by:
• Refinement
• Null results
• Bounce Rate
• Where did they go?

Marko Hurst
Component 6: Results (SERP) Analysis

There are lots of great reports out there, here are a few
I find critical for successful analysis…

Marko Hurst
Single Greatest Piece of Advice I Can Give Today…

Reports & data are fantastic and essential


for analysis. But if you REALLY REALLY
want to find out how well or poor your
search engine is working all you have to do
is… “walk a mile in your visitors shoes”.
MEANING: Take your visitors’ keywords and manually
input them YOURSELF and experience what they did

Marko Hurst
Internal Site Search KPIs

Top searches
Zero / null results
Search conversion
Search usage
Total Unique Searches
Results Pageviews / Search
# of Visits with Search
Search Exits (Bounce Rate)
Search Refinements
Time After Search
Search Depth

Marko Hurst
Summary

Remember this…

Marko Hurst
Summary

Search is not a ‘turnkey’ solution, it takes ongoing effort to get it


right
Improving search improves customer satisfaction, site usability,
SEO, SEM, ROI, & overall user experience
Combine & use both qualitative & quantitative data for your analysis
6 components to search
• Visitor
• Keywords
• Search interface
• Search engine
• Content
• SERP (results)
Site Search book will be out in December

Marko Hurst
Muchas Gracias!
Book: RosenfeldMedia.com/books/SearchAnalytics

Blog: MarkoHurst.com

Contact: [email protected]

Twitter: MarkoHurst

Marko Hurst

You might also like