Six Omniture SiteCatalyst Plugins for better data

We've been doing lots of work recently with clients recently where we have installed plugins. So I thought "Hey, why don't I write a blog post about plugins." So I am. This blog post is going to tell you what they are, what you need to do to get them, some popular ones and the sort of data that they'll give you at the end. The most important thing, of course is always what you do with that data in the end, so I'll give a couple of examples of those as well.

What is a plugin?

There are two parts of any SiteCatalyst implementation. One part is putting a bit of HTML on your page that describes a series of variables that you wish to set: your custom traffic variables (props); your custom conversion variables (eVars); Your custom Products; and your custom events. Campaigns, PageNames, etc are just standard versions of these four things. You can do the entire lot by setting them in the HTML on the page.

The second part is that you need a bit of JavaScript. When the JavaScript runs it picks up all those variables that you set in the page code and sends them off to Omniture's servers. Inevitably you might decide that you want to use the JavaScript file to set some of those variables so that you don't have to code them on to every page. Some of these things will happen automatically (like your browser version, for example), others you will need to set in the code.

The plugins for SiteCatalyst get put into this JavaScript file. It's a way of setting variables without having to code them on the page. The advantage of putting them in the JavaScript file is that you can do some clever stuff by setting values in the cookie and then recalling it or grouping variables together.

If you read through Adam Greco's post on the Omniture website from a couple of years ago (on my birthday no less!), then you'll be able to see all the different types.

I suggest, as does Adam, that you do two things when you think about putting one of these plugins into your JavaScript:

1. Get someone who knows what they are doing to help you (talk to Omniture, hire a consultant, etc)
2. Test thoroughly. I've also discovered recently that even this doesn't always work so make sure that your release schedule will allow you to make fixes easily and quickly if something goes wrong

There is plenty of help in the knowledge base, but I would still suggest the two steps above as there are probably pitfalls that you will come across that a consultant or Omniture would know about that will save you time and money.

Previous Page Plugin

One of the best plugins that you can install is the previous page plugin. It does what it says on the tin really - it sets the value of the current page on the cookie and then when the user loads a new page it loads the previous page in the variable of your choice.

What do you do with this information? Well if you have your previous page in prop1 and your internal search variable in prop2 you can set up a correlation between the two of them and get some very valuable information. You can find out what people searched for on a particular page. When you've found out what people are searching for on a particular page you can make it so that it is easier for them to find that thing by creating links to those things. Better user journeys should mean more sales. But of course you can measure that using an eVar and your events.

Previous Search Term Plugin

Of course the next step in this situation is to create a previous search term report. This report showing the search terms that people used on the previous page, if correlated with the current page, will tell you the sort of thing that people are clicking on when doing a particular search.

With this information you can start altering your search results to give your users better information when they do their searches. This in parallel with an eVar report telling you the conversions that were made for searches is particularly useful to improve your internal search.

Campaign Analysis Plugin

Talking of events and eVars that measure your conversions, there are a few plugins that they've created over at head office can create reports that will give you a bit more data than before. What they've done is created a couple more variables by tracking your campaigns over a longer time period. I talked a couple of months back about tracking your campaigns using a first, last and linear conversion method. Measuring like this will help you work out which campaigns are helping with your conversions.

Of course using a plugin you can write all of those campaigns into your cookie so that when your user comes to convert you have a chronological list of the order they went through them. If you were really clever you could put the dates in as well, but I think that you'd end up with a bit too much information. What you'll end up with in this situation is a nice way of working out how your users interact with campaigns. Knowing this information will help your Marketing guys build their campaign messages based on the stuff that users have responded to in the past. This will help you improve your conversion rates and make more money.

Of course you can do a number of other things with this plugin, including counting the number of campaigns that the person responds to. Remember that if your campaigns report is set to Full Subrelations (as all will be in version 15) then you'll be able to create a report showing for a particular campaign that converted how many other campaigns the users responded to.

Conversion Analysis Plugin

Another part of the conversion analysis plugin actually makes use of two different parts of the plugins. Firstly you use the ability to pick up existing variables to create your own 'marketing channels' report. It sounds a bit odd doing this in the code, but the next step makes it all worthwhile. Every time a user comes to the site you write to the cookie what channel they responded to, so that when the user converts you can write events to each of the other channels showing that they 'assisted' the sale.

This report is invaluable in showing all channels that helped with a sale, not just the first and last as your Marketing Channels does. It may enable you to spend your money in the areas that help with the conversions and not just those that initiate or end it.

Form Analysis plugin

A popular plugin in helping to build a better user journey to get users to convert is the form analysis plugin. This is a helpful plugin that says for a particular page did the user manage to complete it correctly, complete it incorrectly or abandon halfway through. These values are set by using the javascript to 'monitor' what the user is doing. Of course, for the privacy people out there, it will only send signals back to the server in the case of those three events.

So you end up with a report that has the name of your form followed by the field that it was abandoned or incorrectly filled out at along with the number of correct completions. Using this information you can help build better forms for improved conversion. An improved conversion rate for a form will improve conversion for all your campaigns.

Percentage Page Viewed Plugin

As a nice one to finish on, this is something I learned about at 2010's London Omniture summit. This plugin shows the amount of a page that a user has viewed. It allows you to work out if people are getting all the way down to your call to action or reading all of your content. Of course it can only be set on the page afterwards, so you'll need to run it in conjunction with the previous page report. There is a nice post from Adam on this from last year.

Your designers will be the most grateful people in the world if you can give them this report.

Have I missed any that you particularly like out?

5 Tricks for better SiteCatalyst Dashboards

About a year ago I wrote a post on the new Omniture SiteCatalyst Dashboards. This time around I'm going to give you a few hints and tricks of things that you can do to make it all work a bit better. I have been expecting to write about SiteCatalyst version 15 at some point, but there haven't been enough people who have moved to it yet, so it is going to have to wait. You don't want to read a post with examples from a dummy account, do you? In case you were wondering, the dashboards in SiteCatalyst 15 are the same as in version 14. Obviously with the added improvements of things you can do in the normal reports. So lets get going.

1. Put your company logo on the reports

Ok this isn't going to make your dashboards more useful in any way possible, but it does make them look more professional. 

You don't have to have an Adversitement logo

The more professional your reports the more likely the management are to listen to the recommendations that you make because of them! 

Want to go a step further? Why not put a front page on your reports with your name on it so that people know to contact you in case they get stuck. That is easy too because you can add text and images to the front of reports!

You can add in images, text, html pages, RSS feeds and XML feeds so you can customise your pages how you want. If you're going to have a front page to describe your dashboard, you might want to include stuff that will work when it gets published out (which the HTML won't). You can also upload images for free to Flickr to link to for these reports.

2. Add Gauges

Ok this isn't really adding some specific data to your report either, but it does allow you to present it in a way that will mean something to your management. It is adding that vital ingredient of context.

So what do you put in as your maximum and minimum values? Well you have a couple of options:

• Put in your average monthly targets as a top target and 90% as your bottom target (or put in your target as your bottom one and 110% as your top one)
• If you don't have targets, why not put in your highest and lowest values from the last 12 months - this gives nice context as to whether it is within a range or doing better than expected

These targets are for your standard metrics and your custom events, but they are also for your calculated metrics as well. If you want to show how your conversion rate is doing then this is a very good way of showing it.

The downside of these reports are that they are only for metrics - you can't have a split down for the values of one of your variables (eg you can't have conversion from a particular referring domain).

3. Have two graphs

When you set up your reports you quite frequently want to have your graph of two different metrics, but they aren't in the same scale. For example how often have you wanted to have your entries for a page and bounce rate on the same report, but because bounce rates are so much lower than entries the thing looks ridiculous and you can't see the difference of your bounce rates per page.

The annoying thing is, if you create two graphs on the same report then you end up with an annoying situation in your dashboard of only one graph (unless they are pie charts - but that isn't appropriate in all situations).

So your solution is a bit of an annoying way around of creating the same report twice in your dashboard. Do it how you would do normally for just one metric. Then when you come to add your second metric, you need to leave the first metric in your report, but put it at the bottom. Then move your first metric to your right hand graph, but make sure that you are still sorted by the first metric (like so):

You might have to alter your report again to be sorted by entries once it has loaded (why does it do that? Why bother asking if it then doesn't pay attention to it?).

Now when you add to your report they'll be in the same order as before. This is important because you don't want your top pages and your top bounce rates - they won't match up.

4. Copy your dashboard to another report suite

So you've got two websites being tracked through Omniture. The websites come out of the same template of your CMS, so you've done identical SiteCatlayst implementations on them. That's great, it makes your life a lot easier. It makes it a lot easier if you've created one report that you want to replicate in another report suite.

So it is as simple as clicking on 'More Actions' and then 'Copy Dashboard'. When in this menu you should choose to change the report suite and you'll be given a list of them all!

As a little aside to this one - you can quite easily change your reporting cycle at this point as well. If your reports are set up to be monthly reports then you can copy your dashboard to be identical and then change your date rang. Remember when choosing the date range that if you choose 'last week' or 'last month' from the drop down then the reports will roll over at the end of each month/week. If you choose a particular date range using the free form boxes or by clicking on a date then your date ranges will roll over by a day every day. You probably don't want this.

5. Use Classifications to get better graphs

Classifications are useful ways of transferring one variable that has no meaning to the average person to another with a sensible name. I've written about how to do SAINT before - it is really, really simple. 

But the other way you can use SAINT is to help you create graphs for your dashboards that would normally have been unavailable. For example one client had split out the different types of a conversion in an eVar so that they could easily report on each type. It turns out that the ten different values related to two or three different types. We classified these into the three different types so that we could turn them into a graph.

The other option you can use for SAINT is the rather useful ability in Excel of turning text into columns. This is especially useful if you have set up, for example, your page names to be 'Site Section : Page Title'. You want to have entries and bounce rate of your site sections, so you can export the list into excel, use text to columns with your colon as your delimeter.

Another clever way of using the classification is to create a line in a graph that is 'everything else'. An annoying state of SiteCatalyst is that it doesn't have this option. But you can create it using SAINT. Remember that you can only have five values in your graph, so if you are going to have an 'everything else' you need to classify into just four others.

So there you have it. Have I missed out the golden egg that you've been using all this time?