Monthly Archives for September 2010


Track HBX Style Links in SiteCatalyst

I hear from a lot of people migrating from HBX to SiteCatalyst who are looking for ways to make that process a little easier. One of the hurdles I see is trying to migrate all of their HBX link tracking that is already in place to a format that SiteCatalyst can easily understand. That is an easy one to tackle. It’s called the setupLinkTrack plug-in.

Other than tracking all of you current HBX coded links, any SiteCatalyst user can get some benefit from it. Here is another great use for this plug-in. Recently I was reading a great article from @adamgreco about learning to track website navigation. It is a really great article and if you have not had a chance to read it you should go check it out. In one part of the article Adam writes “you should have your developer write code that will pass the name of the link to a Traffic Variable (sProp) when a visitor clicks on each link in your navigation”. Well I’m the guy responsible for the tracking code for my website, so how can I track those links without adding a whole ton of JavaScript onclicks or some other server side hacks? This plug-in will easily handle that, without adding the extra server calls that come with the standard SiteCatalyst link tracking.

First take the plug-in code and add it to the plug-in section of your s_code.js file. Make sure you have the utility functions s.split and s.apl.

 * Plugin: setupLinkTrack 2.0 - return links for HBX-based link 
 *         tracking in SiteCatalyst (requires s.split and s.apl)
s.setupLinkTrack=new Function("vl","c",""
+"var s=this;var l=s.d.links,cv,cva,vla,h,i,l,t,b,o,y,n,oc,d='';cv=s."
+"x in vla)s._hbxm(vla[x])?s[vla[x]]=cva[x]:'';}s.c_w(c,'',0);if(!s.e"
+"o&&!s.lnk)return '';o=s.eo?s.eo:s.lnk;y=s.ot(o);n=s.oid(o);if(s.eo&"
+":o.parentNode;if(!o)return '';y=s.ot(o);n=s.oid(o);}for(i=0;i<4;i++"
+"&&i>-1){eval(\"__f=/ src\s*=\s*[\'\\\"]?([^\'\\\" ]+)[\'\\\"]?/i\")"
+"+'^^'+s.pageName+' | '+(o.lid=o.lid?o.lid:'no &lid')+'^^'+o.lpos;if"
+"(t&&(h||l)){cva=s.split(cv,'^^');vla=s.split(vl,',');for(x in vla)s"
+"._hbxm(vla[x])?s[vla[x]]=cva[x]:'';}else if(!t&&oc.indexOf('.tl(')<"
+"0){s.c_w(c,cv,0);}else return ''");
s._IL=new Function("a","var s=this;return a!='undefined'?a.length:0");
s._II=new Function("a","b","c","var s=this;return a.indexOf(b,c?c:0)");
s._IS=new Function("a","b","c",""
+"var s=this;return b>s._IL(a)?'':a.substring(b,c!=null?c:s._IL(a))");
s._LN=new Function("a","b","c","d",""
+"var s=this;b=a.href;'';c=s._LVP(b,'lid');d=s._LVP("
s._LVP=new Function("a","b","c","d","e",""
+"var s=this;c=s._II(a,'&'+b+'=');c=c<0?s._II(a,'?'+b+'='):c;if(c>-1)"
+");return e}return ''");
s._LS=new Function("a",""
+"var s=this,b,c=100,d,e,f,g;b=(s._IL(a)>c)?escape(s._IS(a,0,c)):esca"
+"rn unescape(b)");
s._LSP=new Function("a","b","c","d","var s=this;d=s._IP(a,b);return d"
s._IP=new Function("a","b","var s=this;return a.split(b)");
s._RP=new Function("a","b","c","d",""
+"var s=this;d=s._II(a,b);if(d>-1){a=s._RP(s._IS(a,0,d)+','+s._IS(a,d"
+"+s._IL(b),s._IL(a)),b,c)}return a");
s._TL=new Function("a","var s=this;return a.toLowerCase()");
s._NA=new Function("a","var s=this;return new Array(a?a:0)");
s._hbxm=new Function("m","var s=this;return (''+m).indexOf('{')<0");
s._hbxln=new Function("h","var s=this,n=s.linkNames;if(n)return"
+"n,',','lnf',h);return ''");

Next in the s_doPlugins section of the s_code file you need a couple lines of setup.

s.hbx_lt = "auto" // manual, auto

The plug-in will use 4 variables and a name for a cookie it will set. For this example I am going to use props 1 thru 4.

Next we have the actual code that will appear in the anchor tag that the plug-in will look for. It is a simple name attribute tag of lid and lpos. This is what a link would look like with the code properly formatted.

<a href="my-page.php" name="&lid=Featured Articles&lpos=Left Nav">Cool Article Name</a>

Here’s what it will do. After clicking on a link that contains a lid and lpos (the exact link shown above), this is what you will find in the debugger on the page you land on:
Debugger 1
The plug-in entered the s.pageName value of the page that the click occurred into prop1, the value of lid into prop2, a combination of the prop1 and the value of lid into prop3 and the value of lpos in prop4.

Lets change up what we include in the actual link code. Lets remove the lid from the tag and just use this link:

<a href="my-page.php" name="&lpos=Left Nav">Cool Article Name</a>

This is what we will get from the debugger:
Debugger 2
You can see by not using the lid tag then the plug-in will use the actual anchor text in those positions.

Now lets say you are not interested in capturing the previous pageName value, or the combined previous pageName/lid value. You can just omit those variables in the plug-in setup. Change the code to:

s.hbx_lt = "auto" // manual, auto

You can see I left the commas in there as empty place holders. Clicking the link with that setup and a lid and lpos value will return this in the debugger:
Debugger 3

What about the other links on the page? What if they do not include a lid or lpos value? The plug-in will track those also. Clicking one of those links will return this in the debugger:
Debugger 4
The plug-in will still capture the previous pageName value, the anchor text and the combined values of the two.

I’m sure by now you noticed that auto/manual setting. Up to now its been set on auto. Let’s flip it to manual and click a link on the page that does not include any code. Here is what you will get:
Debugger 5
You will only get the two previous pageName values. If I would have omitted those two variables like I did earlier then it would have not returned anything. If you click on a link with a lid and lpos variables then it will perform the same as if it were set to auto.

If you want to not add the code to the links so you can track all of the links on a page, but you only want to do it on a specific page and not the whole site you can then wrap the code in an if statement like this:

s.hbx_lt = "auto" // manual, auto

You could also do that using document.location (or a million other ways to identify a specific page).

Since it also grabs the previous page value I no longer need the previous pageName plug-in (useful when using the getPercentPageViewed plug-in).