Understanding Web Analytics and Google Analytics
- 1. Understanding Web Analytics BY THEODORE HAYES AND PRATHAMESH KULKARNI
- 2. What is Web Analytics?
- 3. Formal and wordy definition (1) the analysis of qualitative and quantitative data from your website and the competition, (2) to drive a continual improvement of the online experience that your customers, and potential customers have, (3) which translates into your desired outcomes (online and offline).
- 4. OLD Ways of Analyzing data Insights Clickstream
- 5. WEB ANALYTICS PARADIGM Insights Competitive Intelligence Voice of Customer Experimentation and Testing Multiple Outcomes Analysis Clickstream The What The How Much The Why The What Else The Gold!
- 7. Digital Marketing and Measurement Model
- 8. Steps… Step 1: Identify the Business Objectives. ◦ Why does your website/campaign exist? Step 2: Identify Goals for each Objective. ◦ Goals are specific strategies you'll leverage to accomplish the business objectives Step 3: Identify the Key Performance Indicators. ◦ A key performance indicator (KPI) is a metric that helps you understand how you are doing against your objectives.
- 9. Cont… Step 4: Identify the Targets. ◦ numerical values you’ve pre-determined as indicators of success or failure. Step 5: Identify valuable Segments for analysis. ◦ A group of people, their sources, onsite behavior, and outcomes.
- 10. Focus on 3 key areas of marketing Acquisition. ◦ How are you anticipating acquiring traffic for your website ◦ Did you cover all three components of successful acquisition: Earned, Owned, Paid media? ◦ Where are you spending most of your efforts? Behavior. ◦ What is the behavior you are expecting when people arrive? ◦ Should they visit repeatedly? ◦ Are there certain actions they should take? Outcomes. ◦ What outcomes signify value delivered to the business bottom-line?
- 11. Exercise
- 12. Take any site for your liking Why does the site exist? ◦ Macro And Micro conversions What parts of the website should you focus on first? ◦ What content on the website is directly tied to driving Macro and Micro Conversions? ◦ What sections of the website might be most valuable to the visitors? ◦ What does the top nav and left/right nav groupings tell you about priorities?
- 13. How smart is their digital marketing strategy? ◦ SEO rankings, are they doing any paid search, do they have a facebook page? How well are they doing in context of their competition? ◦ Use google trends, similarweb Is it Mobile Optimized? What is the fastest way to have the impact on the business?
- 15. The secret to being an AGILE analytical ninja Ask a lot of Questions Make VOC analysis your daily habit Solve for customer problems not just KPIs Maintain a database of all significant changes that affect data. Be in the Loop Avoid random optimization at all costs.
- 16. Google Analytics COLLECTION, CONFIGURATION, PROCESSING, AND REPORTING
- 17. How it works
- 19. Contents 1. Quick Glossary 2. Understanding Google Analytics Account Structure 3. Universal – What is it? 4. Understanding and using custom Dimensions and Metrics 5. Understanding Filters and how to use them effectively 6. Regular Expression 7. Understanding and setting up goals and conversion funnels 8. User Interface – Live Example 9. Campaign tracking 10. Alerts 11. Creating Segments 12. Creating Custom Reports 13. Data Driven Attribution
- 20. Quick glossary Bounce Rate Session Exit Rate Dimensions is the percentage of single page visits is a group of interactions that take place on your website within a given time frame. (default 30mins) is the percentage of pageviews for a page that were the last in the session. is a descriptive attribute or characteristic of an object or user
- 21. Quick glossary Users New Sessions Hit Entrances Users whom have had at least one session in the selected date range. An estimate of the percentage of first time visits An interaction that results in data being sent to Google Analytics. is incremented on the first pageview or screenview hit of a session
- 22. Google analytics data model
- 24. Basic FrameWork Account :- your access point for Analytics, and the topmost level of organisation. Property :- Website, mobile application, blog, etc. An account can contain one or more properties. View :- Your access point for reports; a defined view of data from a property. You give users access to a view so they can see the reports based on that view's data. A property can contain one or more views. • Users:- You add users to an account. You can assign four different permissions to a user (Manage Users, Edit, collaborate, or Read & Analyse), and you can assign different permissions at the account, property, and view levels.
- 26. Cookies explained – ga.js and dc.js _utma cookie “Persistent” cookie, 2 years, distinguishes users and sessions. ◦ _utmb cookie (and _utmc urchin.js) “Session” cookie, 30 mins, used to determine new visitors/sessions • _utmz cookie “Source” cookie, 6 months, record how users reach there site. • _utmv cookie “Custom Variable” cookie, 2 years, used to store visitor level custom variable data.
- 27. Google Analytics Tracking Code (GATC) 100% Client side Session-based 4 Cookies:_utma _utmb _utmc, _utmz Universal Analytics Measurement Protocol (UAMP) Client side via analytics.js Server side via measurement protocol Visitor-based 1 cookie with a unique visitor id:_ga Ecommerce tracking requires separate JavaScript GATC VS UAMP TECHNICAL DIFFERENCES
- 28. AND SO? – Universal upgrade
- 30. Online and Offline purchases using loyalty cards can be correlated A voucher printed from your website is redeemed in a physical store. A scanable bar code and store purchases can be connected to website behaviour CONNECTING ONLINE AND OFFLINE • Passing the encrypted FedID will link offline and Online
- 31. Connecting Devices During a telephone order or store purchase user ID is captured. This can be correlated to website data with the same user. E-shop returns CONNECTING SYSTEMS AND DEVICES
- 33. Custom dimensions and metrics In UA, custom variables are called custom metrics and dimensions 20 custom dimensions and 20 custom metrics per account (200 in premium accounts) Custom dimensions and metrics have to be recorded with another data type in universal analytics (like a pageview, event or ecommerce transaction), they cannot be recorded on there own.
- 35. Macro Conversion on an ecommerce site: add a product to the cart and complete a purchase. Micro Conversions are other desirable interactions / communications, including: • Download a catalog .pdf • Request a catalog via mail • Get Special Offers & Savings • Review a product • Add to wishlist • Call to order RECOMMENDED: Setup Conversion Goals for each of the Micro Conversions. Tweak design, content, Calls to Action, then follow with multiple A/B and Multi-Variate (MV) Testing experiments to achieve the best Conversion Rate for each task. This testing process is the key to continuous improvement via Conversion Optimization. Ultimately, assign revenues to each conversion, as each will provide a measure of value and that will help focus your efforts. MACRO AND MICRO CONVERSIONS
- 36. Goal settings Only 20 goals per view. Try and create a view for every type of Macro Conversion. Then you have 19 goals for all the Micro Conversions. This way your account set-up will start to replicate your business structure. Don’t try and analyse too much at once segmentation is key!
- 37. Goal Setup Assign a monetary value to goals. This extra measurement allows economic value reporting. First step to predictive analysis.
- 38. Regular Expression SOME BASICS TO GET YOU STARTED
- 39. Some Basics – MetaCharactersStart of String End of String Escape following character Match zero or many Match one or many Match either one (or) Group Range Match any single character ^ $ * + | () [a-b] [1-5] .
- 41. All Ips IN ONE LINE 1(92.(0.2.[1|2]2|168.1.[2|4])|08.12.(13.[1|5]| 20))
- 42. Filters and effective use
- 43. Filters Filters:- allow you to limit and modify the traffic data that is included in a view. For example, you can use filters to exclude traffic from particular IP addresses, focus on specific subdomain or directory, or convert dynamic page URLs into readable text strings. 1. Predefined Filters 2. Custom Filters Exclude Include Lowercase Uppercase Search and Replace Advanced
- 44. Adding Filters
- 45. Campaign tracking URCHIN TRAFFIC MODULE
- 46. Campaign tracking Adwords will be tracked automatically provided you have specified within Adwords (auto tagging). All other referring campaign traffic sources need parameters adding where practical. Example:- http://iprospectc2.com/reports/search-engine- optimisation?utm_medium=affiliate&utm_source=generic&utm_campaign=bestever Variable Meaning utm_campaign The name of the marketing campaign, e.g. World Traveler Credit Card utm_medium Media channel (E.g. cpc, organic), What is the ‘distribution method’ that is used to get our message out to our customers? utm_source Who are you partnering with to push your message. A publisher, Google, Yahoo etc. utm_content The version of the ad (Used for A/B Testing). You can identify two version of the same ad using this variable. utm_term The search term purchased (if your not buying keywords).
- 47. Campaign tracking – Cont.. Default Channel Groupings Definition Direct Source exactly matches Direct and Medium exactly matches (not set) Or Medium exactly matches (none) Organic Medium exactly matches organic Referral Medium exactly matches referral Email Medium exactly matches email Paid Search Medium matches regex ^(cpc|ppc|paidsearch)$ And Ad Distribution Network does not exactly match content Other Advertising Medium matches regex ^(cpv|cpa|cpp|content-text)$ Social Social Source Referral exactly matches Yes Or Medium matches regex ^(social|social-network|social-media|sm|social network|social media)$ Display Medium matches regex ^(display|cpm|banner)$ Or Ad Distribution Network exactly matches Content Note:- Channel groupings are case sensitive, Google / CPC will fall into (other) NOT Paid Search.
- 48. Check (other) for incorrect UTM_Parameters Tag is incorrect This happens when text ads are used on the display network, consider creating another default channel utm_parameters are case sensitive Adding utm_parameters to banners overwrites original source/medium Over 13.36% of traffic has been incorrectly defined. Check spelling Display Channel has been user defined to ‘DisplayCPC’ auto- tagging will be affected.
- 49. DO’s and Don’ts Do ◦ Always use small case characters when creating UTMs ◦ Always escape special characters ◦ Use URL builder ◦ Create additional default channel for sms campaigns ◦ Adhere to set naming conventions Don’t ◦ Tag internal banners or elements on the website. ◦ Create your own utm_mediums. (Exception sms campaigns) ◦ Create utm_source which is difficult to identify.
- 50. Tracking methods in ga
- 51. Tracking methodolgies GATC and UATC or ‘Tag/code snippet’ ◦ Urchin.js ◦ Ga.js ◦ Dc.js ◦ Analytics.js Event Tracking – Action, Category, Label, Value Virtual Pageview Ecommerce
- 54. 1. Segmentation by Referrer/Traffic Source. Paid, Natural, Paid brand and non-brand By number of keywords -2,3,4 Social Media 2. Segmentation by Visitor Type New. Returning & Registered Visitor 3. Segmentation by Engagement 5 Pages <3 Pages 4. Segmentation by Content Viewed Key Landing Page Product Page Checkout Complete Folders for Large organisations 5. Segmentation by Landing Page Type 6. Segmentation by Event Conversion goal types and E-commerce 7. Segmentation by Platform (less important) Browser Screen Resolution Mobile 8. Segmentation by Location Main Markets UK US FIGS ROW CREATING SEGMENTS
- 55. Conversion funnels Every online business has steps that people must take to become a customer. Use Conversion funnels to:- Determine what steps are causing customer confusion or trouble. Figure out what language or copy might be altering our customer’s emotional behaviour during checkout or sign up. To be aware of bugs, browser issues and other technical nuisances.
- 56. Conversion funnels Funnels Provide Greater Insights than Goals Alone As essential as goals are in Google Analytics, they serve by themselves more as basic KPIs (key performance indicators) than as actionable starting points for conversion optimisation.
- 58. Essential Alerts Significant Traffic Drop Traffic Spikes • Significant Drop in Goal Completions • Analytics has flatlined
- 59. Essential alerts Spike in goal Completions Significant drop in Google Referrals Direct Traffic bounce rate over 70% ◦ When you create a custom alert, that alert is visible only to you in your current view and in any other views to which you apply the alert. ◦ Period: Select the frequency at which the alert can be generated (Day, Week, Month).
- 60. Alerts
- 61. alerts When you create a custom alert, that alert is visible only to you in your current view and in any other views to which you apply the alert. Apply to: If you want to apply the alert to additional views, open the other views menu, then select the check box for each view to which you want to apply the alert. The alert is then available to you in any of the other views you select. Period: Select the frequency at which the alert can be generated (Day, Week, Month).
- 63. Enhanced ecommerce – web tracking (analytics.js) The enhanced ecommerce plugin for analytics.js enables the measurement of user interactions with products including product impressions, clicks, viewing details and adding a product to shopping cart, initiating checkout process, transactions and refunds. Can send multiple types of ecommerce data using analytics.js. Impression, Product, Promotion and Action data.
- 66. Last interaction/last click interaction model Standard attribution model in all web analytics tools (except Google Analytics) •Very unfair attribution model.
- 67. Last non-direct click attribution All standard reports in GA give 100% of conversion credit to the last “campaign” prior to conversion. • This model is also the irritating reason why none of your standard analytics reports match your multi- channel funnel reports. • This model under values the direct channel.
- 68. Last adwords click attribution This model is profoundly value-deficient.
- 69. First interaction/click attribution Reverse of last click – gives all credit to first click. •This model makes even less sense, than the last click model. • First click attribution is akin to giving your first girlfriend 100% of the credit for you marrying your wife
- 70. Time decay attribution A Much Better Attribution Model • The media touch point closest to the conversion gets most credit. • This passes a common sense test and is a simple and more accurate model.
- 71. Position based attribution This attribution model is a customizable model. • Don’t forget the default model is not to be used. You need to have a good understanding of your business to allocate credit.
- 73. Understanding the business. Key questions you need to ask. ◦ What type of user behaviour do you value? ◦ Is there an optimal conversion window you are solving for? ◦ What does the repeat purchase behaviour look like historically? ◦ Are there any micro-conversions defined with engagement type goals, tied to the economic value? ◦ Are offline conversions being sent back into GA using Universal Analytics?
- 74. Step 1 Select the Baseline Model. ◦ Good to start with position based model, ◦ Here is a good model setup which applies to a lot of industries.
- 75. Step 2 Select the look back window. ◦ Good rule is to pick close to the upper limit of the number of days to conversion, excluding outliers + a bit more. ◦ B2B have longer conversion look back windows, B2C such as ecommerce have much shorter.
- 76. Step 3 Select the Engagement based credit option. Page depth is a better metric for engagement as time on site is not calculated for bounce visits etc. This means campaigns with higher engagement will get more credit.
- 77. Step 4 Apply Custom Credit Rules. (Tricky Bit) ◦ Here you can apply any custom rule you want, so this is to tweak your attribution model to more accurately fit the business. • This custom rule is ensuring ads that get clicks to be extra rewarded and, in this case, get 1.4 times the credit of other campaigns in the conversion paths (in comparison to ads that just get impressions)..
- 78. Summary Experiment with the custom rules to see what fits. Use common sense when creating your attribution model. Ask the important business questions. If you use common sense, your attribution model even if not perfect will be a lot more accurate than a last click model.
- 79. Thank You
Editor's Notes
- WHAT is collecting, storing, processing, analyzing your click level data. Using tools such as Webtrends, Google Analytics, Ominiture etc. THE HOW MUCH: websites attempts to deliver three types of outcomes. Increase revenue, Reduce Cost, Improve Customer Satisfaction/ Loyality. Example missing reports were not missed due to not measuring outcomes. THE WHY – Finding out the why the customer did something. THE WHAT ELSE: The power to know if your performance is good. Are you winning the race! Competitive intelligence helps you identify new opportunities and keeps you relevant. THE GOLD: Insights to Action.
- Your objectives should be DUMB: Doable. Understandable. Manageable. Beneficial.
- Ask all the business questions, even if you can see the answers in Google Analytics. For example this can give valuable insight into whether the data/client is correct, save valuable time, give a better understanding into consumer/behavior and business, allowing you to make the connections in the data. Always listen to the customer – this is the source of the revenue. The happier the customer, the more likely they are to purchase or be retained.