Media Ad Concepts

Personalization concepts:

There are many categories of web personalization including

  1. Behavioral
  2. Contextual
  3. Technical
  4. Historic data
  5. Collaboratively filtered
There are several camps in defining and executing web personalization. A few broad methods for web personalization may include:
  1. Implicit
  2. Explicit
  3. Hybrid

---------------------------------------------------------------------------------------------------------------------------

Media Ads Concepts

  1. Ad inventory: The number of ad spaces available with the publisher. If a publisher has 100 spots, where s/he can display the ads then 100 is ad inventory.
    1. It is valued in terms of the site traffic or ad views that can be delivered to the advertiser.
    2. It is calculated in number of page impressions. If site A has 100 impressions/day and site B has 1000 impressions/day then site B can ask for more money from advertiser because site B can reach to many websites. 
  2. An ad value can be calculated as page impressions X average number of ads on your website pages Although above formula is simplest of all. 
    1. Moreover, it can be calculated by the location of the ad. Ads displayed at the top of the webpage demands for more money since it is the prime stop as compared to ads displayed at the bottom. 
  3. Fill Rate: It shows how much of your ad space has been rented out and how much of it is empty. Higher the fill rate better are the chances since it means your website is in high demand.
  4. Revenue: A website’s revenue can be calculated by number of leads or actions are performed after clicking on the ad or after cpm(cost per 1000 impressions)

---------------------------------------------------------------------------------------------------------------------------


Metrics Concepts:

Here is the summary of articles on Product Metrics. Article: Only Metric that counts
Summary: 
  1. Are people using the product? Once you find that out, find out:
    1. Why do they use your product? 
    2. How frequently? 
    3. What key actions they perform while using it? 
  2. Answers to above questions will tell you who are core users, who:
    1. Come directly
    2. Return weekly/monthly
    3. Share it with others
  3. Reference:
    1. https://news.greylock.com/the-only-metric-that-matters-now-with-fancy-slides-232474cf414c

---------------------------------------------------------------------------------------------------------------------------

Ad Ecosystem:

This section will explain you the basics architecture of ad ecosystem. Definitions of different components and how those components work together to render ads:


Publisher:
  1. Runs online business or websites that has user traffic.

Ad Network:
  1. Aggregator of publisher ad inventory that is packaged by contextual channels, audience types or buying options(CPC, CPA, CPM). This inventory is sold. 
  2. It is an entity/company/organization that connects advertisers to websites.
2.     It enables to deliver ads to consumers, perform targeting, tracking, and reporting of impressions.
3.     Inventory of online ads include:
a.     Blogs
b.     Rss feed
c.      Web pages
d.     Mobile pages
e.     Instant messages
f.      Mobile apps
g.     Adware Emails
h.     Other media
4.     A publisher(new york times) can integrate with an ad network. Ad network can then display certain ads
5.     Large publisher sells remnant(not sold = 10-60%) inventory through ad networks. Smaller ones sell everything through the ad networks.
6.     Large ad networks include:
a.     Search engines
b.     Media companies
c.      Tech vendors


Seller Side Platform:
  1. Platform for publishers to manage ad inventory and enables routing to different ad exchanges


Ad Exchange:
  1. Enables buying and selling of Media ads from multiple ad networks.A marketplace for buying and selling of ads. Ad inventory is auctioned to advertisers/agencies.
  2. Handles Real time Bidding for publisher that includes:
  3. -       Purchases inventory at individual impression level
  4. -       Bidding in real time(milliseconds)
  5. Ensures that publishers can maximize value of their inventory.
  6. Ensure that advertisers purchase individual impressions at fair price.            

DSP:
  1. Platform that enables media and audience buying for advertisers in real time on ad exchanges    
  2. Evaluates and bids on each individual impression through RTB based on available user data.
  3. Takes care of RTB for advertisers.
  4. Centralized campaign management, reporting, optimization to meet advertiser’s objective.

RTB process:
  1. Publisher puts the available inventory on the ad exchange
  2. Ad exchange holds auction for each impression
  3. DSP receives the bid from advertisers.
  4. DSP evaluates the impression and calculate bid amount based on its value to advertiser. 
  5. Ads from winning bid are shown on the publisher’s site.

Advertiser:
  1. Spends on marketing to grow their business.

---------------------------------------------------------------------------------------------------------------------------


More resources:

  1. 5 things ad tech product managers should consider when planning their company's future
    1. NOTES:
      1. While contacting SSPs lot of information like user browsers, location, etc can be collected from requests & processed auctions.
      2. User information can be collected via the DMP. They will have services that will provide you user information via cookie syncing services.
      3. Consider inventory types like TVs or VRs where ad can be served in future. Product requirements should take care of such things. 
      4. Consider use of machine learning algorithms to generate metrics that will help describe if correct data is collected and if right metrics are generate. 
      5. TODO: Look at new video ads protocol
      6. Make the product flexible to allow for change in trends
      7. Cost of scale: Don't let technology costs kill the when scaling comes into picture.  
  2. Confessions of an adtech product manager

No comments:

Post a Comment

NoSQL

This one is reviewed but I need to delete its copy from hubpages or somewhere NoSQL Data models: key-value  Aggregate model.  key or i...