As presented at #BrightonSEO and #SearchElite Manchester September 2017
Uploaded version is the longer deck as used for #SASCon 2017 - examples and use cases for Google Analytics features and tools to make more from the data you can collect.
21. Google Analytics API to Excel, csv or database
Also API for Facebook, Twitter, AdWords, Search Console & more
Allows scheduled and batch data pulls
22. Google Analytics API to Excel, csv or database
Also API for Facebook, Twitter, AdWords, Search
Console & more
Allows scheduled and batch data pulls
Also has save to Drive or Cloud database options
23. Combine Business & Web Data
Deeper Analysis and visualisation
PowerBI also has native API link to Google Analytics API etc.
PowerQuery and PowerBI have REST JSON API
(make your own query from any source API)
Also have csv folder source extract for batch processing
24. Requires a little more work than just using the Analytics UI, more powerful than VLOOKUP
This is an example Data Model to split detail out
This lets you combine, calculate and filter as needed
25. Bigger datasets, more powerful statistical analysis
PowerBI cloud has real language querying
“What is the best performing product for revenue this Quarter?”
26. Geo lookup or CRM for customer addresses (or combine both)
Visualise on drill down map or overlay ArcGIS data
e.g. Conversions > Avg from Areas with Household income > average
27. Average alone can mask opportunity & risks
Tableau, R or PowerBI to analyse distribution patterns
Best for your lowest value, least loyal customers
May be Worst for your big spenders
Often not a business win
even if it reports as a Test or Campaign win
33. Success Events
If [Good Thing] happens
Track [Good Thing]
Ideally [Type of Thing] | [Which Specific Thing] | [What Details]
34. Success Events - Detail
Why this much Detail?
• Custom Segments
• Remarketing
• Event Goals
• Filters and Offline Analysis (Dealer ID & stockID)
(need to be a bit careful of high cardinality on unfiltered or unsegmented data)
45. Clicks By Race Day
With 3 Event Goals – break down ways Total is reached
Compare Click Volumes per Race Day AND Total
46. Buy to Click Ratios
Calculate Buy to Click Ratios by Day AND Totals
Sunday has higher drop-off between Click and Buy
1560 ÷ 3164 = 49.30% 466 ÷ 1041 = 44.76% 2015 ÷ 4067 = 49.54%
48. Custom Dimensions
• Product SKU
• Site ID
• Visitor ID
• User Lifetime Value
• Product Availability
• Previous Submit
• Checkout Account Type
• Guest
• New Account
• Logged In
• Experiment ID and Experiment Variant
Practically almost any Dimension you might want to track (non PII)
Limited to 20 on a standard account. 200 on GA Premium
49. Custom Dimensions
Dealer ID to link back to Network CRM
Dealer Name varies (multi-franchise)
Report whole network performance but link back to Dealer Sales
50. Custom Metrics
Dimensions Describe Data
Metrics Measure Data
• User/Session of Described Type – Count
• Lead Value
• Original Price
• Displayed Price
• Product Level Discount Value
• Volume Level Discount Applied
• Finance Deposit
51. Custom Metrics
Frequently eCommerce will pass either
• Original Price (even if discounted Price is applicable)
• Discount Price (but no idea how much discount vs. Original Price)
Custom Metrics allows you to record & calculate the differences
Reports can then show Value Lost in discount/promotions
This allows you to report real Revenue performance
52. Calculated Metrics
Calculate Business Metrics not Web Metrics
Original Price – Discount Price Applied = Value Lost
Buy Exeter Tickets ÷ Click Exeter = Buy to Click Rate
53. Calculated Metrics
Buy Exeter Tickets ÷ Click Exeter = Clicks to Buy Rate
But we can now use this in Custom Reports
By Channel | By Campaign | By Device | By Gender | By Age etc.
56. Custom Reports
Drill down to Individual Tickets – Exeter Saturday has most Clicks
Compare performance of Exeter Saturday 2016 vs Exeter Saturday 2017
Are sales behind schedule? Traffic growth but lower Clicks to Buy Ratio?
58. Lots of Data
Pull data via API
Match against Business Data
Automate
Crunch Segments, Dimensions, Metrics
Take Action
59.
60. PowerQuery > Data Model > PowerPivot Measures > Excel visualisation
Same Data Model, Measures and Query could also be used in PowerBI
61. Manual: Pull data via API to Sheets and Download. Download csv from GA UI
Automate: Next Analytics, Analytics Edge, Supermetrics, Power Query, PowerBI
63. • Automatic Refresh
• Custom Filters
• Date Range
• Dimension Filters
• Custom Segments
Up to 20 Pages per Report
Unlimited Reports
Free!
64.
65.
66.
67. Distribution matters
• 2 users off the scale of this chart
• 130+ items per purchase
“Average” is misleading if you don’t know how it is distributed
74. Segment Share of Contribution
Share of Sessions or Event Clicks easily compared
Or Targeted. Or Reported against # Stock of each Year being advertised etc.
78. Regular Expressions
“Price To” from ?= Search string from Referral Path BUT
Different format for “Price To” Search on EVERY Referrer site
Expresso http://www.ultrapico.com/expresso.htm (Windows Only)
http://regexr.com/ (Browser)
80. Custom Channel Groups
Default Channels are ok if you tag inbound links and adverts accurately
But Custom Channel Groupings are better
81. Custom Channel Groups
Start with Default as Template
Divide Channels more precisely
Rules apply top to bottom
Different Groups
Different Levels of Detail
Paid Search is a good example
By Type:
• Shopping
• Display
• Text
• Video
By Spend:
• High Cost
• Medium Cost
• Low Cost
Buying Cycle, Competition, Volume etc.
82. This is a Top line Channel splitting – slightly more detail than Default Channels:
• Paid: Main Ad types, Top Funnel, Middle Funnel, Bottom Funnel
• Organic Search: Google or Other – Keyword Known/Not Provided
• Email: Which Mailing list (Mailchimp or EventBrite)
Custom Channel Groups
83. Custom Channel Groups
More detail on Paid Campaigns
• Paid Search
• Paid Social
2nd Dimension of Keyword
Exeter Race promotion
• Keywords mostly Exeter
• Some Top Funnel generic
92. Content Grouping
Tag Manager makes it easier
dataLayer makes it simple
Page Template: Used Bikes > All Used Stock Page Titles
Makes: Ducati > Ducati Models
Models: Multistrada > Multistrada Variants (S, Enduro etc.)
Model Year: 2009, 2010, 2011 etc.
Price Bands: £5,000 - £7,499 | £7,500 - £8,999 | £9,000 - £10,499 etc
Up to 5 Groupings
99. Resources
Get Started With Tag Manager & Google Analytics
https://analytics.google.com/analytics/academy/
Learning Resources
https://www.simoahava.com/
http://www.lunametrics.com/blog/
http://www.analytics-ninja.com/blog.html
Demographics and Advertising Tools Policy Requirements
https://support.google.com/analytics/answer/2700409?hl=en-GB&utm_id=ad
Regular Expression Tools
http://regexr.com/ (Online)
http://www.ultrapico.com/Expresso.htm (Windows)
Data Studio
https://www.google.com/analytics/data-studio/
API Tools
http://www.analyticsedge.com/
https://supermetrics.com/
http://www.nextanalytics.com/
https://powerbi.microsoft.com/en-us/
https://www.scitylana.com/
Connect on Linked In https://www.linkedin.com/in/timstewart/
Or ask me a question on Twitter @pjeedai