Web Analytics Dashboard and Analysis System Daniela Fernandez Student Id Supervisor: Michael Johnson ‘On the Web, absolute numbers rarely matter, trends do, and segmented trends really do’. Kaushik A. ITEC811 1

 Basic Concepts  Aim of Project  Problem vs. Solution  Overall Process  Related Work  Requirements Specification  Design (WADAS System, Dashboard, Suggestion Module)  Implementation  Tests and Results  Conclusions 2

Basic Concepts Metrics Dimension Web Analytics Dashboard Measures Give context Key data in 1 page 3

 Basic Concepts  Aim of Project  Problem vs. Solution  Overall Process  Related Work  Software Requirements Specification  Design ( WADAS System, Dashboard, Suggestion Module )  Implementation  Tests and Results  Conclusions 4

Aim of Project  PROBLEM  Too much data  Reports difficult to understand  No time for strategic decisions  SOLUTION  Web Analytics Dashboard  Analysis System with suggestions 5

6 Web application Provides flexibility to create our own dashboard in iGoogle Could fail for users who are not experts Does not store GA authorization credentials Desktop application Needs the Polaris software to be downloaded Maintaining the information available Several reports on the screen

 Basic Concepts  Aim of Project  Problem vs. Solution  Overall Process  Related Work  Software Requirements Specification  Design ( WADAS System, Dashboard, Suggestion Module )  Implementation  Tests and Results  Conclusions 7

 Functional Requirements  User Management  Security  Retrieving and Displaying Data  Design Requirements  The system shall be developed in a web environment, not desktop.  Each dashboard must fit a page not bigger than A4  Non – Functional Requirements 8 User Requirements Provided by

 Basic Concepts  Aim of Project  Problem vs. Solution  Overall Process  Related Work  Software Requirements Specification  Design ( WADAS System, Dashboard, Suggestion Module )  Implementation  Tests and Results  Conclusions 9

10 Designed following the UML approach Based on the model- view-control (MVC) design pattern

11 Four Dashboards -Sales - IT/Design - Marketing - Managment

12 Users can: Review behaviour of visitors over the last 12 months Identify traffic sources with better revenue. Measure loyalty See changes between last and current periods Check percentage of goals reached Identify daily products performance View better campaigns and products.

13 Users can: Analyse % of visits using different browsers and platforms. Compare visits and revenue per traffic source Analyse performance of pages in the site. Measure value of top 5 pages viewed. Recognize use of internal search Identify which network vendor used by visitors produces more ROI

14 Users can: Analyse the use of the internal search Identify most valuable pages Observe performance of referring sources Consider the advertisement of better keywords Compare the performance of better products between last and current period

15 Users can: Determine which traffic source is not performing well. Compare AdWords performance between last and current periods Analyse behaviour of visits from different sources. Identify opportunities and possible problems that affect ROI Measure goals performance Observe better sold products

 Users will see a list of suggestions to:  Improve the accuracy of the data  Understand advanced options enabled in GA  Take action about less obvious elements in the dashboards You have not set goals for this profile. Goals can help you to know how many visitors are completing your goal. (e.g. average of visitors that have purchased x product). 2. You should set your default page in your profile settings to avoid distinct entries in the report referencing the same page (e.g and can be the same but are shown twice because the default page ‘index.php’ is not set). 3. The campaign ‘Camp 1’ is not generating any revenue since it was created, you should consider making changes for that.. 1. You have not set goals for this profile. Goals can help you to know how many visitors are completing your goal. (e.g. average of visitors that have purchased x product). 2. You should set your default page in your profile settings to avoid distinct entries in the report referencing the same page (e.g and can be the same but are shown twice because the default page ‘index.php’ is not set). 3. The campaign ‘Camp 1’ is not generating any revenue since it was created, you should consider making changes for that..

 Basic Concepts  Aim of Project  Problem vs. Solution  Overall Process  Related Work  Software Requirements Specification  Design ( WADAS System, Dashboard, Suggestion Module )  Implementation  Tests and Results  Conclusions 17

 The system was implemented using:  PHP (Zend Framework) + MySQL database  JavaScript  HTML  Security issues were considered for  Authentication to WADAS ▪ Login and Password ▪ Sessions  Authentication and Authorization to GA API ▪ Authsub Protocol  Access to source code from interface 18

19 Retrieve account feed from GA Account Retrieve data feed from GA Account >&dimensions= &metrics=<ga:namemetric >&start-date= &end-date=<ga:enddate Call to Google Visualization API library Load graphic package (piechart) //creating an DataTable object dataPieVisits ← DataTable(); number_rows ← size of feed add number_rows to dataPieVisits Object for each (element in feed) set value of feed in dataPieVisits Object //Creating a Pie Chart chartPieVistis ← new PieChart( ) //Drawing a Pie Chart chartPievisits ← Draw( )

 Basic Concepts  Aim of Project  Problem vs. Solution  Overall Process  Related Work  Software Requirements Specification  Design ( WADAS System, Dashboard, Suggestion Module )  Implementation  Tests and Results  Conclusions 20

 Evaluation of dashboard mock-ups using questionaries  10 people answered the questionaries  Time of responding was measured (Avg Time Spent – 29 minutes for the total dashboards)  Feedback meeting performed after evaluation. 21 Dashboard/ Questions Q1Q2Q3Q4Q5Q6Total Sales IT-Design Marketing Manageme nt

 Security Issues  Creation and Deletion of sessions for WADAS  Login and Logout process to WADAS  Requesting of the single-use token to Google  Upgrading and Revoking the token for long-lived session using the AuthSubSessionToken and AuthRevokeToken methods.  Verification of denial of access to other profiles when the user does not have a role of Administrator.  Requesting and Retrieving data from GA 22

 Basic Concepts  Aim of Project  Problem vs. Solution  Overall Process  Related Work  Software Requirements Specification  Design ( WADAS System, Dashboard, Suggestion Module )  Implementation  Tests and Results  Conclusions 23

 Was found that:  WADAS brings benefits to basic and expert users.  Tools help, but are not the key factor in Web Analytics. People are the key.  It’s possible to make decisions with less than 100% accurate data.  Future Work  Complete implementation of dashboards designed.  Collect additional information from AdWords to provide better suggestions.  Make the Suggestion Module robust adding neural networks processing 24

25