Downloaded 35 times
Uploaded byCharles Meaden
31 Ways To Destroy Your Google Analytics Implementation
The document outlines 31 common mistakes that can undermine Google Analytics implementations, highlighting issues caused by marketing teams, developers, and agencies. It covers problems like missing or duplicate tags, incorrect settings, and improper tracking, along with their consequences and methods for detection and resolution. By identifying and fixing these issues, organizations can improve their data accuracy and analytics effectiveness.
More Related Content
Similar to 31 Ways To Destroy Your Google Analytics Implementation
31 Ways To Destroy Your Google Analytics Implementation
- 1. 31 Ways todestroy your Google Analytics implementation Charles Meaden Digital Nation
- 2. 30 seconds aboutme • Founder of Digital Nation • Been using web analytics in one form or another for 20 years • Using Google Analytics since it was launched
- 3. The Spirit ofMeasurecamp http://www.wired.com/2015/01/rodney-mullen/
- 4. A Lesson FromMy First Boss •Shit in •Shit out
- 5. It Used ToBe So Easy • 20 years ago, all we had to was process log files • It’s quite difficult to corrupt a log file
- 6. Aggregation of MarginalGains http://jamesclear.com/marginal-gains
- 7. We Shouldn’t BeHaving This Talk • If everyone associated with web analytics on a project did their job properly • So who’s to blame..
- 8. 1. Marketing Teams •‘Don’t have the time’ – to write a proper brief – tag campaigns – keep people informed of new campaigns • ‘Its not my job’ attitudes • Won’t take the time to learn
- 9. 2. Developers • Insistthey know the best place to add tracking codes • Won’t read technical briefs or just cherry pick from them • Don’t see the need to comment or document their code • And my all time favourite • Not testing their code in a browser to make sure it deploys correctly
- 10. 3. Digital Agencies •“Of course we know analytics” • Just adding the code onto the page and nothing else • Not taking the time to understand what the client wants • Don’t see how they can make money from this
- 11. Why Do MistakesHappen Lack of knowledge prevents sites from being tagged correctly Organisations don’t truly value the data from their tools and won’t commit sufficient budgets to getting them setup correctly Agencies who claim ‘expert’ knowledge of analytics yet don’t even exclude the IP address of the client Not enough attention being paid to analytics, so mistakes are not spotted for weeks or months
- 12. You can help •If there is anything I’ve missed let me know and I’ll add it to this deck • Get hold of me via – [email protected] – Twitter: @charlesmeaden – Linkedin: uk.linkedin.com/in/charlesmeaden
- 13. 4. Tags MissingFrom Pages Consequence • Pages not being recorded at all • Self referrals from your own domain name • Incorrectly recorded landing pages How to spot • Use a crawling tool such as WASP or Screaming Frog to search the UA code on the page Fix • Maintain a list of all page types and monitor on a regular basis – Custom Alerts will give you an early warning
- 14. 5. Too ManyGoogle Analytics Tags Consequence • Your bounce rate drops to zero • Multiple sets of data running around your organisation How to spot • Use either HTTPFox , Google Analytics Debug or WASP to see which tags are being fired • If using Wordpress or CMS check which plugins are being used Fix • Hunt through the source code till you’ve find them and work out which one can go
- 15. 6. Mixing ‘Classic’and Universal Code on the same site Consequence • Multiple sessions being recorded • No guarantee that your data will be recorded correctly How to spot • Use a crawling tool such as Screaming Frog or Wasp Crawler to crawl your site and look for tags on the site Fix • Upgrade the site to Universal Analytics You can run classic and Universal side on the same page without any issues. This refers to using classic in one section and universal in another
- 16. 7. Set DomainName Not Set Or Misconfigured Consequence • Traffic from other sites can appear in your domain • Visits from blog.yourdomainname.com come in as separate visits How to spot • Pages start appearing that don’t belong to your site • Your subdomains appear as referrers Fix • Universal - Add the domain name to the referrer exclusion list • Add the setdomainname into the your GA code (classic)
- 17. 8. Your CodeDeployed On Other Sites Consequence • Inflated visits • Affects all of your metrics How to spot • Look at Year on Year patterns for anything that looks runs unusal • Use Custom Alerts to alert you to any big increases in traffic Fix • Remove the incorrect filter • Set a filter to only include traffic from your hostname
- 18. 8. Your CodeDeployed On Other Sites Consequence • Main site and shop on a subdomain • Intention was to setup two Google Analytics accounts – Cross domain tracking not possible for “techical reasons” • The shop fired both accounts • Vists from the shop ended up in the main account • Over the year, figures overinflated by 52%
- 19. 9. Forgetting ToExclude Yourself / Client / Employees Consequence • Inflated visitor figures • Inflated conversion rates from all those tests you run How to spot • Look and see in Audience > Technology > Networks for large amounts of traffic from one network • Large visitor numbers from countries outside your target market Fix • Use a filter to exclude all known static IP addresses and known ISP names • If you have a lot of visits coming from dynamic IP’s, consider using a custom variables / custom dimensions
- 20. 10. Not EnablingCross Domain Tracking Consequence • New sessions created as you move between domains • No true picture of the user journey How to spot • Traffic coming in from other subdomains or domains within your organisation Fix • Deploy cross domain tracking • Google Tag Manager and Universal Analytics make this a far simpler process
- 21. 11. Excluding KnownBrowser Bots Consequence • Inflated visitor figures • Distorted metrics How to spot • High levels of direct traffic with 100% bounce rate • Look and see in Audience > Technology > Networks for large amounts of traffic from one network • Visits coming from Mozilla User Agent Fix • Use a filter to exclude IP address and networks
- 22. 12. Excluding MaliciousBrowser Bots Consequence • Inflated referring sites visits • Distorted metrics How to spot • Look for ‘Semalt’ or ‘Buttons for Websites’ • Other referring web sites that have 100% bounce rates Fix • Enable the Bot Filtering in View > View Settings • Create a filter to exclude traffic from Semalt and Buttons
- 23. 13. Not QA’ingCode in the Browser Consequence • Data either not recorded or wrong • If not spotted in time, long term data quality issues • People stop trusting the data How to spot • Immediate drops in data • Look for code – wrapped in comment tags – no code at all Fix • Use HTTPFox , Google Analytics Debug or WASP to check whether the GIF is being fired • Have a test GA view that simulates the standard account
- 24. 13. Not QA’ingCode in the Browser
- 25. 13. Not QA’ingCode in the Browser
- 26. 14: Losing UTM& Gclid Codes Due To Redirects Consequence • Marketing information not correctly recorded • Increase in direct traffic How to spot • Direct traffic is far higher than you’d expect • Missing medium and sources • Use HTTPFox to analyse the redirects and see what gets stripped out Fix • Discuss with your developers the most efficient way of developing the redirects that handle the codes
- 27. 15. Not LinkingAdwords to Analytics Consequence • All Adwords traffic is reported as organic • Lose your Matched Search Query report How to spot • Nothing reported in the Adwords section Fix • Link Adwords to Analytics
- 28. 16. Not TaggingCampaigns Consequence • Visits appear as referrals or direct – Email will appear as mail.google.com or direct – Display campaigns just appear in the referring sites How to spot • Peaks in direct traffic or referring traffic when email is sent out • No campaign traffic when you’re running them Fix • Email - talk to your ESP about the most effective way to tag your pages • Teach those who run campaigns how to tag them
- 29. 17. Using CampaignTracking for Internal Promotions Consequence • Existing medium / source / campaign information overwritten • New sessions are started How to spot • People ‘landing’ on pages that don’t make sense such as • Medium, Source and Campaign tags that refer to internal Fix • Remove the tags • Replace them with events • Look at Universal Enhanced Ecommerce for banner tracking
- 30. 18. Events BeingFired on Page Load Consequence • Drops the bounce rate to zero for that page • Affects the overall bounce rate How to spot • Use HTTPFox, Google Analytics Debug or WASP to see what is being sent back to Google and what is firing the event Fix • Set the event to be a non-interaction event
- 31. 19. None orOne Thank You Page for All Forms Consequence • Incorrect conversion rates How to spot • Goals have almost identical conversion rates Fix • Create separate thank you pages • Set the pageview element to record a distinct name for each form • Fire an event when the form is successfully completed – Downside is you can’t create a funnel
- 32. 20. Goals IncorrectlySetup Consequence • Incorrect conversion rates • Zero conversions How to spot • Funnels that don’t look right – bulge in the middle – same numbers for each step Fix • Use regex to tightly define goals • Test, Test and Test again
- 33. 21. Session ID’s Consequence •Increase the number of unique page names and running into Google limits of 50,000 per table How to spot • 1000’s of combinations of the same name Fix • Exclude specific variables in the admin section
- 34. 22. Other TrackingID’s Consequence • Increase the number of unique page names and running into Google limits of 50,000 per table How to spot 1000’s of combinations of the same name Fix • Exclude specific variables in the admin section
- 35. 23. Site SearchNot Implemented Consequence • Searches recorded as page views • Can’t use the Site Search tools to segment traffic How to spot • The Content > Site Section is empty Fix • Look at URL’s to determine where the query string is – http://www.example.com/search?controller=search& search_query=zoo • Set up site search in the Google Analytics Admin section
- 36. 24. Zero SearchResults Not Tracked Consequence • You don’t know which search terms failed to produce zero results How to spot • You can’t as you’re not tracking them Fix • Modify your search results URL to include a phrase such as noresults – http://www.example.com/search?controller=search&search_query=noresults:zoo • Fire an event that contains missing search term • Use Google Tag Manager to scrape the DOM element
- 37. 25. Incorrect TimeZones Set Consequence • It’s not clear when the peak times are on your site How to spot • Traffic appearing at strange times, normally early in the morning Fix • Change the timezones within the admin section – The change isn’t applied retrospectively
- 38. 26. Incorrect CurrencySet Consequence • Transactions are recorded in the wrong currency – Actual number will be correct if not specifying currency in ecommerce data How to spot • The wrong currency symbol is shown Fix • Change the currency within the View Settings
- 39. 27. Default PageSet Incorrectly Consequence • Pages created that don’t actually exist • Can break goals How to spot • All directories contain the default page Fix • Before removing it, see if has been used in goals and segments
- 40. 28. Not SettingAll Pages To Lower Case Consequence • End up with duplicate pages – Happens on Windows IIS servers which are not case sensitive How to spot • Pages will appear as two separate rows Fix • Eyeball pages names to check for any issues • Setup a lower case custom filter
- 41. 29. Agencies creatingProperties Not Accounts Consequence • Hassles when trying to get admin access to accounts How to spot • Agencies refusing access to accounts Fix • Shoot a short video explaining how to change the user accounts
- 42. 30. Deleting YourGoogle Analytics Account Consequence • Your data has gone! How to spot • The account doesn’t appear to be there Fix • Select the Trash Can icon to recover the account – You have 35 days to recover account
- 43. 31. Not CorrectlyTracking 404 Pages Consequence • If you have no clear 404 page or way of identifying it, then data from the error pages gets mixed up with live pages How to spot • Enter a URL that you know doesn’t exist • Typically, sites will do one of the following – Show a dedicated 404 error page on it’s own unique URL – Show the a 404 error message on the URL you entered Fix • Fire an event or virtual page view to identify the pages • Lunametrics has a good article on tracking 404’s using Google Tag Manager