Source data in an ACDL implementation

[aablank1]

We'd like to keep the custom rules in Data Collection to a minimum.  A clean data layer can solve that.  What is the best practice for source data in an ACDL implementation?  

There are two ways I've seen to set it up:

1. Use a reference file in a cms or data store, that doesn't require a code change to update names.
2. Use the cms to generate the page names automatically based on the navigation hierarchy.

Automatically generating names sounds the best.  However there are some issues in our environment which complicate things:

• To save server calls, we specify which links are tracked.  It really adds up.  (All pages and almost all modals are tracked.)
• We adhere to a breadcrumb page naming convention (specified by Adobe consulting many years ago).
  • Example - 'mens: shirts: polo'
• The page titles on our site tend to be inconsistent and vague in areas (updating them is moving heaven and earth - and a potential seo struggle).  That's why we specify the names.
• Not all parts of our site have a cms or the file structure to adapt for naming rules (again inconsistent).

So this is a bit of a puzzle.  Any thoughts would be welcome.

[bjoern__koth]

Hi @aablank1 

First of all, the type of website should dictate the structure of your data layer. Meaning, if you have products, make sure you have a "products" array, etc.

In terms of page information, I would always try to capture 

• the document title, even this may change over time (document.title)
• the page path (document.location.pathname)
• if possible a language-independent path that reflects the content structure of that one specific page. In other words, a /path/to/page that is identical for the same page in English, French, Italian page. Big benefit here is that this will give you an aggregated value over all variations of the same page (that also should be a independent from any future changes to the page title
• document language and country information 

but yeah, it really depends on what type of website you have

[aablank1]

Really appreciate the advice.  Capturing the path and title sounds good.

With the "products" array, are there standard array structures documented so we could pick what we need?  It seems like creating new ones would be a bit like reinventing the wheel.