Implementation of Java Script for Analytics

WebSideStory vs. CoreMetrics vs. Omniture

[ThePhoenix]

We are looking at the vendors above for analytic solutions. We are concerned about a couple things:

1. How much code needs to be implemented on each page?
2. How much manipulation needs to be done to the code for each page?

Does anyone have experience with these vendors? Any help would be appreciated.

-ThePhoenix

[Receptional]

I don't use any of the above, but the principal is the same. You will need to add a snippet of code to every page of your website. It is the same snippet on every page EXCEPT, you may need to differ this code or add to it on pages where significant actions (like a sale) have occurred.

If you have a templated page, the code usually fits neatly into the template somewhere. Otherwise, you'll need something to do a global search and replace, adding multiple line of text. ( they may have a wizard to do this for you).

So - easy job if you have a web developer. Nightmare if your web development experience is limited to wysiwyg or using a content mamangement system as a user only.

[cgrantski]

Just implementing the snippet of code is the plainest most vanilla approach that will get you little more than log files with cached pages included. It's hardly worth paying the extra $$ if you don't add extra, page-specific code to get additional information. I'm talking about many pages besides the receipt page. In fact some of those solutions you're looking at require the extra code to be added to pages in order to get things that are able to be set up in the UI of other products. Uh, that wasn't real clear. I mean, what you may be using as Content Groups or something like that in WebTrends, set up by specifications in the WebTrends GUI, can only be collected for other analytics programs if specifically preconfigured into the tags . (WebTrends isn't the only one that allows a lot of setup to be done through the GUI rather than its tags; I'm just using it as an example because you're already familiar with it.)

Sometimes the cost of implementing those tags is massively undestimated or conveniently overlooked by vendors during the selling process. Implementation is a major profit-maker for their professional services groups. You may find out too late that some of the coolest features are add-ons. Figure out what you'll need first and get it all spec'd out by the various vendors.

--- voice of experience

[ThePhoenix]

Thanks for the input. I am not concerned about having experienced developers. We are a very large organization with pleanty of resources.

I just have the head developer in my office asking me about how time intensive the implementation process is going to be. All of vendors talk about how great their products are, but none of them have told me how much time it is going to take to implement.

-ThePhoenix

[cgrantski]

Well, this afternoon we implemented plain vanilla tags on a site of about 1,000 pages and it took 10 minutes not including QA. The site has includes in the templated footers. It really would have taken 3 or 4 minutes but there were a few popup pages that did not use the footer and needed to be modified individually. It would have taken about the same amount of time if the site had 10,000 pages or more.

For each page that needs extra coding, such as a receipt page, count on anywhere between 10 minutes to 4 hours for the special code, depending on the number of extra pieces of info you want to be collected. Once the engineer gets the basic idea, that is.

My time estimates are for WebTrends tags. They have a nice feature where the tag itself needs no modified coding; you just have to figure out how to get the extra information you want into <meta> tags which automatically get scooped up by the WebTrends javascript. I've been told this makes it a lot easier than some other products, but I'm not an expert.

On the other hand, I've watched a team of two people from *********** (one of the vendors you listed) struggle on-site for four weeks to tag a commerce site. I don't want to slander them by describing a bad situation that may not be typical for that vendor, so I won't name them. Phoenix, a huge organization like yourselves will probably be able to get the vendors to get really specific about implementation, considering what's at stake for them.

[ThePhoenix]

Thanks for the input. I am just very concerned that the implementation will be very intense. I am working with some very agressive timelines. I cannot afford to miss my deadlines, nor can I hold up this project.