How Auto Content Discovery Works in Predictive Content

[#1]

Image: Step Flow for Predictive Content for Web and Email

What is Auto Content Discovery?

This post explains the first part of the flow image above: discovering, mapping, and tracking your content assets. Content Discovery is used in the Predictive Content app to auto-discover all the content assets (videos, pdfs, blog posts, press releases, white papers, etc.) on a website/domain. Once discovered, you can see how your content is performing and decide which pieces of content should be prepared, approved, and enabled for the Predictive Content outcomes on web (either in the recommendation bar or rich media) and email.

Setting up Predictive Content and Content Discovery

See Set Up Docs: http://docs.marketo.com/display/public/DOCS/Getting+Started+with+Predictive+Content

Steps

1. RTP Javascript tag is installed on all your web pages (Note: RTP is now called Web Personalization)
2. Set Asset Discovery to On
   1. Asset Discovery auto-discovers extensions (PDFs, PPT, MP4, OFF, WEBM) and a click/view on embedded videos (Youtube, Vimeo, Wistia)
3. Create URL Patterns
   1. Setting URL patterns auto-discovers content when a web visitor clicks on the HTML web page relevant to the content pattern.
   2. The URL syntax pattern determines your ability to use this feature optimally (e.g., to create a pattern for press releases, your press releases should all be stored on a page identified as www.yoursite.com/press-releases/*).

How is content auto-discovered?

The content discovery technology uses an event listener that runs on every web page (where the RTP Javascript tag is installed) and waits for a web visitor to click on a URL link or arrive directly on that web page in a browser. If that link includes an extension (PDF, ppt, embedded video) or matches the URL pattern defined, then it will be discovered and added to the All Content page in the Predictive Content app.

Only content pieces that a web visitor interacted with (clicked on or viewed) once the RTP Javascript tag has been installed are discovered via Content Discovery. If the content is already discovered, it will add to the tracking and views of those discovered pieces. You can also manually add new content to be listed and tracked in the All Content page.

Image: All Content Page in the Predictive Content app displaying and tracking all discovered content

What data is discovered?

Visitor Data (Used for Analytics + Predictive Algorithm)

• Pages Viewed  - number of page views by the user in the session
• Visit Count for Web Visitor
• Last Content URL Seen
• Last 5 Content Assets clicked via Predictive Content that this visitor has seen in the last 90 days
• Last Web Campaigns Seen - 10 last campaigns per session (within the last 5 sessions)
• Inferred Organization
• Inferred, Industry, Size, Revenue
• Inferred Country, State, City
• Search Term
   

Content Data  Extensions

• Video (Youtube, Vimeo, Wistia)
  • Video Name
  • Video URL
  • Video Image URL
• PDF
  • PDF Name
  • PDF URL

URL Patterns (HTML pages)

• Found via metadata of the HTML page
  • Content Name
  • Content URL
  • Content Image URL
  • Content Description

What data is auto-populated during the auto-discovery phase?

Based on the content data we discover, the aim is to populate as much of the Predictive Content as possible, making it quicker, easier and involves less prep work for you. However, you still need to review the discovered content and then approve and enable it for the Predictive Source (Email, Rich Media, Recommendation Bar). Assuming we discover an HTML page defined in URL patterns, this HTML content piece will be populated in the following fields in Predictive Content:

Predictive Content Fields	Auto-Populated Value	Notes
Content Name	(Content Name)	Unique Value
Content URL	(Content URL)	The URL is consistent for all sources (email, bar, rich media)
Categories	Video OR Category Name from URL Pattern	If Video is discovered, the category is populated as Video. Category populated from defined URL Pattern. e.g., Marketo.com/blog = Blog all discovered content based on this URL would receive Blog as a category Category is consistent for all sources (email, bar, rich media)
Content Title	(Content Name)
Email Title	(Content Name)
Email URL	(Content URL)
Email Image URL	(Content Image URL)
Email Button Label	Read More	Not Auto-populated. Default is "Read More."
Rich Media Title	(Content Name)
Rich Media URL	(Content URL)
Rich Media Image URL	(Content Image URL)
Rich Media Description	(Content Description)
Bar Title	(Content Name)
Bar URL	(Content URL)

Image: Example of Populating Metadata for HTML content into Predictive Content