GTM listener for non-GTM-based GA tracking (hardcoded on-page, plugins, etc) [a.k.a. spy/hijack GA]

- README - spyGoogleAnalytics.md

This has been moved to a repo: smhmic/ga-spy.

• Example Usage
• Code for Async
• Code for Universal

EXAMPLES - spyGoogleAnalytics.js

//
// These examples are for UA, but would be very similar for GA Async.
//
(function spyGoogleAnalytics( callback ){
  /* GA library -specific code goes here. (See other gist files for full code.) */
})(function( a ){ 
     
    //   v  v  v  CUSTOM CODE GOES HERE  v  v  v  
     
    // RETURN FALSE to prevent original hit from firing.
    // By default, the original hit fires just as it normally would.

    // ARGUMENTS passed to the GA object are available in the array `a`.
    // See GA documentation for parameter formats: https://goo.gl/muCY7Q
     
    // FOR DEBUGGING: console.debug.apply( console, a );
     
    
    var dataLayer = (window.dataLayer = window.dataLayer || []);

     
     
    //
    //  EXAMPLE 1: Expose everything (pageviews, events, tracker config, etc) to GTM.
    //
    var frame = {}, 
        namespace = 'spy.ga.' + a[0]
                        + (typeof a[1] == 'string' ? '.'+a[1]
                            + (typeof a[2] == 'string' ? '.'+a[2] : '') : '');
    frame.event      = namespace;
    frame[namespace] = a;
    dataLayer.push( frame );
    // return true;  // Allow non-GTM hits to fire unmodified.
    // return false; // Prevent all non-GTM hits from firing.

    
    
    //
    //  EXAMPLE 2: Expose every event to GTM (simple).
    //
    if( a[0] == "send" &&  a[1] == "event" ){
       dataLayer.push({
          'event'    : 'spy.ga.event.' + a[2],
          'spy.ga.event' : {
             'category'       : a[2],
             'action'         : a[3],
             'label'          : a[4],
             'value'          : a[5],
             'nonInteraction' : a[6] && a[6].nonInteraction
          }
       });
    }
    
    
    
    //
    //  EXAMPLE 3: Expose every event to GTM. Unlike Example 2 above, this
    //  handles alternate parameter formats (see https://goo.gl/muCY7Q).
    //
    var fields;
    if( a[0] == "send" ){
          
       // If last argument is object, use as fields object.
       fields = typeof a[a.length-1] == 'object' ? a.pop() : {};

       if( ( a[1] || fields.hitType ) == "event" ){
          fields.category       = fields.eventCategory  || a[2];
          fields.action         = fields.eventAction    || a[3];
          fields.label          = fields.eventLabel     || a[4];
          fields.value          = fields.eventValue     || a[5];
          fields.nonInteraction = fields.nonInteraction || undefined;
          //TODO: If using fields other than the standard event fields above,
          // either: a) use same code as `nonInteraction` above to explicitly
          // set new value, OR b) clear all old values before pushing new ones
          // with `google_tag_manager[{{Container
          // ID}}].dataLayer.set('ga.event',undefined);`
          dataLayer.push({
             'event'        : 'spy.ga.event.' + fields.eventCategory,
             'spy.ga.event' : fields
          });
       }
    }


    
    //
    //  EXAMPLE 4: Intercept built-in tracking from ShareThis and AddThis
    //  so custom naming conventions can be applied via GTM.
    //
    if( a[0] == "send" && a[1] == "event" && ( a[2] == "ShareThis" || a[2] == "addthis" ) ){
       dataLayer.push({
          'event'       : 'socialShare via' + a[2],
          'socialShare' : {
             'network'  : a[3],
             'url'      : a[4]
          }
       });
       return false; // Return false to stop original hit from attempting to fire. 
                     // Returning true allows analytics.js to try to fire ShareThis/Addthis 
                     //   built-in tracking.  Only return true if either on-page tracking 
                     //   OR AddThis/Sharethis is configured to track to a separate 
                     //   GA Property than the Property tracked by GTM.
    }
    

    //   ^  ^  ^  CUSTOM CODE END  ^  ^  ^   
   
 });

spyGoogleAnalytics-async.js

/**
 * Spy on asynchronous calls to Google's Traditional Analytics (ga.js) library.
 * 
 * Built with GTM in mind, but can be used for other TMSes or custom JS.
 * 
 * Intended to piggyback and/or block external (non-GTM-based) tracking (i.e.
 * on-page code, plugins/platforms with built-in tracking, etc), in order to:
 *  - leverage benefits of GTM for non-GTM-based tracking
 *  - override naming system of external tracking with custom conventions
 *  - integrate external tracking with your measurement implementation
 *  
 * NOTE: NOT THOROUGHLY TESTED.
 * 
 * NOTE: This is as reliable as any custom code, including dataLayer snippets.
 * Performance impact is minimal. But it's preferable to migrate/modify/remove 
 * the external tracking, for a cleaner implementation. This code is provided 
 * for cases where editing the external code is not an option.
 *
 * NOTE: In order to spy on commands that fire immediately / page load/ready,
 * this must run before the code that loads ga.js.
 * 
 * @see https://gist.github.com/smhmic/492213ea74bb47c5878b
 * @author Stephen M Harris <[email protected]>
 * @version 0.0.1
 */
(function spyGoogleAnalytics( callback ){

   var
    gaqObjName = '_gaq',
    _gaq       = window[gaqObjName],
    console    = window.console || {error:function(){}},
    k, i,
    
    // Processes each set of arguments sent to _gaq.push.
    // If returns false, arguments will not be passed through to _gaq.push.
    handler = function( a ){ try{
      // If command is not via GTM Tag, return result of custom code.
      return ( a[0] && a[0][0] && a[0][0].substr && a[0][0].substr( 0, 4 ) == 'gtm.' )
          || ( false !== callback( a ) );
    }catch(ex){ console.error(ex) } },
    
    // Process array of push'ed args, and return filtered array.
    processArgSet = function( arr ){
       var aFiltered = [];
       for( i=0; i< arr.length; i++ )
          if( handler( arr[i] ) )
            aFiltered.push( arr[i] );
       return aFiltered;
    },
    
    // Spy on the _gaq.push function.
    spy = function(){
       var gaqPushOrig = window[gaqObjName].push;
       // Replace _gaq.push with a proxy.
       window[gaqObjName].push = function(){
          var aFiltered = processArgSet( [].slice.call( arguments ) );
          // If some args passing through, or no args were passed,
          //  pass through to _gaq.push.
          if( !arguments.length || aFiltered.length )
             return gaqPushOrig.apply( window[gaqObjName], aFiltered );
       };
       // Ensure methods/members of _gaq.push remain accessible on the proxy. 
       for( k in gaqPushOrig )
          if( gaqPushOrig.hasOwnProperty( k ) )
             window[gaqObjName].push[k] = gaqPushOrig[k];
    };
   
   if( ! _gaq ){
      // Instantiate GA command queue a la GA Async snippet.
      _gaq = window[gaqObjName] = [];
   }
      
   if( window._gat ){ // GA already loaded; cannot see previous commands.
      spy();
   } else if( _gaq.slice ){ // GA snippet ran, but GA not loaded.
      // Filter all existing command queue items through custom code.
      _gaq = window[gaqObjName] = processArgSet( _gaq );
      _gaq.push( spy );
      spy();
   } else { 
      throw new Error('spyGoogleAnalytics aborting; '
                      +'`'+gaqObjName+'` not the GA object.' );
   }

})(function( a ){ 
    /** @var [Array] a - The arguments pushed onto `_gaq` **/
     
    //   v  v  v  CUSTOM CODE GOES HERE  v  v  v  
     
    // RETURN FALSE to prevent original hit from firing.
    // By default, the original hit fires just as it normally would.

    // ARGUMENTS passed to the GA object are available in the array `a`.
    // See GA documentation for parameter formats: https://goo.gl/muCY7Q
     
    // FOR DEBUGGING: console.debug.apply( console, a );
    
    // EXAMPLES: https://git.io/vK4VJ
    


    //   ^  ^  ^  CUSTOM CODE END  ^  ^  ^   
   
 });

spyGoogleAnalytics-universal.js

/**
 * Spy on calls to Google's Universal Analytics (analytics.js) library.
 * 
 * Built with GTM in mind, but can be used for other TMSes or custom JS.
 * 
 * Intended to piggyback and/or block external (non-GTM-based) tracking (i.e.
 * on-page code, plugins/platforms with built-in tracking, etc), in order to:
 *  - leverage benefits of GTM for non-GTM-based tracking
 *  - override naming system of external tracking with custom conventions
 *  - integrate external tracking with your measurement implementation
 * 
 * NOTE: This is as reliable as any custom code, including dataLayer snippets.
 * Performance impact is minimal. But it's preferable to migrate/modify/remove 
 * the external tracking, for a cleaner implementation. This code is provided 
 * for cases where editing the external code is not an option.
 *
 * NOTE: In order to spy on commands that fire immediately / page load/ready,
 * this must run before the code that loads analytics.js.
 * 
 * @see https://gist.github.com/smhmic/492213ea74bb47c5878b
 * @author Stephen M Harris <[email protected]>
 * @version 0.4
 */
(function spyGoogleAnalytics( callback ){

   var
    gaObjName = window.GoogleAnalyticsObject || 'ga',
    ga        = window[gaObjName],
    console   = window.console || {error:function(){}},
    k, q, i,
    
    // Processes each set of arguments sent to GA object.
    // If returns false, arguments will not be passed through to GA object.
    handler = function( args ){ try{
      // If command is not via GTM Tag, return result of custom code.
      return ( args[0] && args[0].substr && args[0].substr( 0, 4 ) == 'gtm.' )
          || ( false !== callback( args ) );
    }catch(ex){ console.error(ex) } },
    
    // Spy on the GA object.
    spy = function(){
       var a, gaOrig = window[gaObjName];
       // Replace GA object with a proxy.
       window[gaObjName] = function(){
          if( handler( a = [].slice.call( arguments ) ) )
          // If command is via GTM Tag or custom code does not return false, 
          // pass through to GA command queue.
             return gaOrig.apply( gaOrig, a );
       };
       // Ensure methods/members of GA object remain accessible on the proxy. 
       for( k in gaOrig )
          if( gaOrig.hasOwnProperty( k ) )
             window[gaObjName][k] = gaOrig[k];
    };
   
   if( ! ga ){
      // Instantiate GA command queue a la UA snippet.
      ga = window[gaObjName] = function(){
         (window[gaObjName].q=window[gaObjName].q||[]).push( arguments ); };
      ga.l = 1 * new Date();
   }
      
   if( ga.getAll ){ // GA already loaded; cannot see previous commands.
      spy();
   } else if( ga.l ){ // UA snippet ran, but GA not loaded.
      if( ga.q ){
         // Run all existing command queue items through custom code.
         for( q=[], i = 0; i < ga.q.length; i++ )
            if( handler( [].slice.call( ga.q[i] ) ) )
               q.push( ga.q[i] );
         ga.q = q;
      } else {
         // No commands queued yet, instantiate queue.
         ga.q = [];
      }
      ga( spy );
      spy();
   } else { 
      throw new Error('spyGoogleAnalytics aborting; '
                      +'`'+gaObjName+'` not the GA object.' );
   }

})(function( a ){ 
    /** @var [Array] a - arguments passed to `ga()` **/
     
    //   v  v  v  CUSTOM CODE GOES HERE  v  v  v  
     
    // RETURN FALSE to prevent original hit from firing.
    // By default, the original hit fires just as it normally would.

    // ARGUMENTS passed to the GA object are available in the array `a`.
    // See GA documentation for parameter formats: https://goo.gl/muCY7Q
     
    // FOR DEBUGGING: console.debug.apply( console, a );
    
    // EXAMPLES: https://git.io/vK4VJ
    


    //   ^  ^  ^  CUSTOM CODE END  ^  ^  ^   
   
 });