Jul 012011

Do you know how data is collected when Advertisements are placed on various Websites? Excerpt from previous article Vendors and Tools in Web Analytics Data 

For Example: If Intel asks Google DoubleClick DART to place Intel ad on CNN.com, then CNN is the publisher, Intel is the advertiser or Brand Website, and Google DoubleClick technology helps Intel to place their ad on CNN.com.

What happens next? How is this data collected?

The data will be collected depending on various tags on CNN.com website. If CNN.com has third party Double Click DART tags the data will be collected as follows:

  1. When users sees Intel Ad on CNN.com an Impression File is created (CNN.com has DART Impression TAG).
  2. When a user clicks on the ad, a click file is created (CNN.com has DART Click Tag).
  3. After clicking on Intel AD, user is taken to Intel Website, an Action File is created (Intel Website has DART Spotlight/Floodlight TAG/Action TAG to collect actions/visits on intel.com).

The data is collected at Double Click DART Adserver and the three files are created. See sample files below:

This data is then collected into databases. Great Web Data Analyst can analyze user level data using data mining tools such as SQL, SAS etc. for advanced analytics and help in optimizing media and better site experience. Notice all this information is non PII. The data has user_id 1826881611 (a system generated number) and so on.

After this data is collected there are several ways an analyst can segment this data. Depending on Advertiser, Impression file is huge and often contains billions of impressions/rows of data. Similarly, size of click and action file will depend on brand size (the more the activity on website, the larger will be the files).

Using above data, Analytics Team can help answer many questions such as below:

  • Which site, creative concept, creative size id is doing better?
  • What attributes bring quality traffic?
  • What is the Optimal Frequency (Optimal Impressions served before converting)?
  • Dashboards integrating benchmarks, goals and Key performance indicators (KPI) for all Display Advertisements?
  • Day Parting (Which parts of day and what hour of day bring maximum traffic)?
  • Geo Segmentation (Which states, DMAs, countries generates maximum traffic)?
  • Reach (Unique users per day, month, or campaign lifetime by Site, Creative, Placement etc)?
  • Total Online Sales, # of Units, Cost Per Sale?
  • Cross Site Duplication etc?

Similarly, various Search Engines collect data for Search. If both Paid Search Campaign (PPC) and Display Campaign are trafficked (Third Part DART Tags) in DART then Search and Display data can be combined for advanced Attribution Analysis. For Example: Impact of Display ads can also be calculated on Search Traffic because same user id (aka cookie id) will be seen in both Display and Search Data File.

Does above help you in understanding Ad Serving and data collection methodology? Do you have any questions or want to share a similar methodology?

Please participate via comments. Thanks.

  4 Responses to “Digital Ad Serving Data Collection Methodology and Applications”

  1. This is very helpful!

  2. Hi Shilpa…it is is quite useful information…I might require more details…I am working on Multivarite Data Analysis..

  3. Ying & Sujeet: Thanks. I am glad you found it useful.
    Sujeet: Sure. Please send me more details. I look forward to help further and more discussions.

  4. I have been searching the web for a site like this. Using analytics is a HUUUUUUUGE profitability point for progressive marketing. Thanks very much I will be a regular visitor!

 Leave a Reply



You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

CommentLuv badge