Internet+Online Advertising

About Internet Advertising


Internet advertising (also referred to as online advertising or interactive advertising) is a marketing channel, through which advertisements are placed on the pages of a web portal. With access to the Internet becoming simpler, creative agencies are developing content that can deliver better value to advertisers and consumers at much lesser costs.

New tools are being developed and perfected to a) enable the analysis of non-personally-identifiable data to detect patterns in users’ interests and consumption habits and b) to allow for predictive modeling based on user responses, thus enabling the development of better-targeted ads.
As the acceptability of this channel increases, so do the total revenues. As per the latest reports compiled by the Interactive Advertising Bureau and PricewaterhouseCoopers LLP, Internet advertising revenues exceeded $5.2 billion for the third quarter of 2007 – a $1.1 billion increase (25.3%) over Q3 2006.


Figure 1. Internet Advertising Revenues, U.S.

Online advertising has become a de-facto revenue source for most start-up portals and many established portals, as well. In fact, Internet powerhouse Google earns most of its revenue through online advertising, with an annual turnover of approximately USD $14 Billion.1

Concepts
Internet advertising is no different than any other form of advertisement or promotion in terms of its objectives, which is to advertise products and services to potential customers.

Key Players
There are four main categories of actors in the business: publishers, advertisers, advertising networks and advertising agencies. Generally, big publishers are dominant players in this market, while advertising agencies and network strive to meet the publishers’ needs.

• Advertisers are corporations, manufactures, retailers, service providers, individuals and any other entity who has a product or service to sell. Typical objectives of an advertiser include building and reinforcing a brand, informing the audience of a new product, offering a service, capturing or defining newer customer segments or markets, etc. Examples of advertisers include Coca Cola, Ford Motor Company and Aviva Life Insurance.
  
  
• Publishers are all the web portals and websites that have a user base accessing the portal for various reasons. Publishers define various categories of content and information. A publisher is sometimes also known as a “channel,” examples of which include sports, auto, kids, fashion, etc. Internet companies such as Yahoo.com, Rediff.com and Indiatimes.com are all publishers.
  
  Portals that intend to be in the publisher category must lease or sell “real estate,” “zones,” or “property” on web pages across various portal channels. It has become an inherent part of web page design to create or provision for zones as part of a monetizing strategy. Publishers are ranked on various parameters, including:
• Page views
• Quality and type of content
• Demographics
• Reputation
• Unique users and user base
• Advertising Agencies are brokers that form a bridge between advertisers and publishers. Working on a commission-based model, advertising agencies depend on advertising networks and publishers to run campaigns for their clients.
  
  
• Advertising Networks are entities that have the technology and infrastructure to develop online advertisement campaigns, deliver advertisements to publishers’ pages, record information, and develop metrics and reports on the budget and health of the campaign, as well as implement behavioral and geographical targeting. An advertising network has a substantial number of advertisers, smaller networks and agencies as its clients. It deals with or negotiates with publishers on terms and conditions, revenue models, cost of property, etc.
  
  
  

Internet Advertisement Formats

Advertisements vary based on technology and format. Some standard types of Internet advertisements include:

• Banner advertisements (also called “display ads”) are one of the more popular formats of Internet advertising. They are often made of text and graphics, either static or animated.
  
  
• Interstitials include pop-ups and pop-unders and are the most controversial format of Internet advertising. Interstitials are perceived to be a nuisance, as they put users in a forced exposure mode.
  
  
• Keyword search has become the most popular format of Internet advertising in terms of revenues. Keyword search, also called “paid listings,” are major revenue sources for Internet search sites such as Google.
  
  
• Rich media is a generic term for a variety of highly interactive, visually influential Internet advertising formats. With Flash, advanced JavaScript, streaming audio and video, and JavaScript powered interactivity, rich media advertisements deliver enhanced impact and result in improved user response.
  

Campaign Parameters
Internet advertising campaign parameters are closely related to one’s marketing objectives for communication. These parameters take into consideration marketing requirements like target audience age groups, ethnicity, taste, interests and passions, etc., while also leveraging other factors like brand recall, short term memory, internet usage patterns, etc. Below is an indicative list of parameters that define a campaign’s structure and tag an advertisement server’s logic for campaign execution:

• Geo targeting delivers different content to users from different locations, as determined by the users’ IP addresses (e.g., advertising Sambhar mix to Indian users and advertising popcorn to American users).
  
  
• Channel targeting runs campaigns only on relevant websites (e.g., placing perfume advertisements on fashion and personal accessory pages).
  
  
• Frequency capping restricts the amount of times a specific visitor to a website is shown a particular advertisement. Advertisers must determine the amount of exposure that will ensure memory retention without wasting inventory.
  
  
• Click capping prevents the same advertisement from repeating once a user has clicked on it, as the user will most likely not click on it again.
  
  
• Duration capping determines the length a campaign should run (e.g., no more than two weeks).
  
  
• Day capping determines the specific days on which a campaign should run (e.g., only on weekdays).
  
  
• Hour capping determines the specific hours during which a campaign should run (e.g., between the prime surfing hours of 11:00 am and 4:00 pm).

Revenue Models
Advertisers utilize many models to generate revenue through Internet advertisement campaigns, including some of the following:

• Cost-per-thousand-impressions (CPM): In this approach, the advertiser agrees to pay to the agency based upon the number of “impressions” achieved for a given campaign (e.g., 200,000 impressions in two months with/without cappings). An impression is any instance in which an advertisement is viewed. It is important to note that because of user advertisement blockers that prevent the delivery of images, a page view does not necessarily guarantee an impression.
  
  CPM agreements are the most commonly deployed revenue model in Internet advertising. As such, agencies are required to provide much reporting and information to advertisers to develop a trusting relationship. In fact, the ability to self-configure campaigns and deliver transparent information are becoming key differentiators among agencies in retaining and attracting clients.
  
  
• Cost-per-lead (CPL),also called Cost-per-acquisition (CPA): This model works mostly for rich media advertisements where users are prompted to provide their contact information. These are workflow-based advertisements, such as for home loans. After a user has filled in the required details, the last step Is to produce a “Thank-You” page or redirect the user to the advertiser’s landing page. Both these steps are recorded by the advertisement server and enable the network to charge for the “leads” generated.
  
  A CPL/CPA agreement is an advanced revenue model through which a network receives a predefined payment amount or percentage based on the actual sale of goods or services. Since these transactions may occur many days after the time the advertisement was generated, tools are being developed to better track this relationship.
  
  
• Cost-per-click (CPC): In this model, revenue is generated every time a user clicks on a given advertisement. Click-through-rate (CTR) is the most important metric in this situation.
  
  Although the above models have been explained from an advertiser/network interaction standpoint, publishers are a dominant player in the field and generally define the terms and conditions of the models. Publishers can take some or all of the following approaches:
  
  
• Fixed price sale: Internet advertising inventory such as zones and number of page views are sold for a fixed price to the advertiser/network. In this scenario, it is the network’s responsibility to optimize ROI.
  
  
• Bidding approach: A publisher may choose to sell a space to multiple networks and advertisers at the same time. In this case, multiple advertisements compete for a page view. Networks are currently developing complex algorithms that optimize ROI for the publishers, thereby earning the page view.
  A variant on the bidding approach can be demonstrated by Google AdWords. In this scenario, advertisers bid on particular search words. When a user Googles one of these reserved words, that advertiser’s advertisement on the search results page.
  
  
• Pay-per-click: In this approach, contextual advertisements are placed on a publisher’s website. If a user viewing the page clicks on the advertisement, revenue is generated for the publisher. This approach has also been popularized by Google.

When to Use Which Model

• When an advertiser picks a publisher, the publisher’s risk needs to be mitigated. In this scenario, the CPM model is most appropriate.
  
  
• When a publisher picks an advertiser, the advertiser’s risk needs to be mitigated. In this scenario, the CPA model is most appropriate.
  
  
• When advertisers and publishers are matched by a third party, both sides assume risks. In this scenario, a CPC model is most appropriate. It is also important to note that the size of the network matters just as much as the matching algorithm in this situation.

Technical Implementation
The foundation of online advertising is the adserver, a web server that hosts advertisements. This is the tool that updates ad content each time a user visits or refreshes a website. An Adserver can also count the number of impressions/clicks for an ad campaign and generate reports, which helps advertisers with ROI analysis. Depending on its size, an advertising network may deploy many ad servers in a clustered and load balanced environment to assure ad delivery and provide alternate paths in case of technical snags. The following work flow describes the lifecycle of serving an ad. An accompanying figure can be found below:

• The advertiser provides campaign objectives to the network or agency. The network then converts the objectives to a campaign that maps the requirements to various parameters.
  
  
• These parameters are fed into the adserver to create a “tag” that is based on the type of ad desired by the advertiser. A tag is a JavaScript code that is embedded within the desired property or real estate space on a publisher’s page.
  
  
• When a user views a page, the JavaScript alerts the adserver. It informs the adserver about the user’s ID and other parameters (if built into the tag) such as the user’s IP address, operating system, browser, connection type, screen resolution, etc.
  
  
• The adserver records this page view and identifies the type of ad and campaign to be displayed.
  
  
• The above information is relayed back to the page, and the required ad or banner/creative is loaded by the content server.
  
  
• The adserver is informed after an ad has been displayed, and an “impression” is recorded. It is important to differentiate between a “request” and an “impression” in this scenario. A request is recorded when a page with the desired tag is loaded, whereas an impression is recorded only if the ad creative has been displayed.

Figure 2. Online Advertising Infrastructure and Workflow


EMERGING TRENDS
With the greater availability of inexpensive memory and broadband access, the Internet is becoming a domain in which many people spend a substantial amount of time – especially with the steady rise in social computing. As a result, the concepts of viral marketing and word-of-mouth advertising are gaining prominence. In order to better target user bases and create highly productive advertisements, refined techniques are being developed and implemented.

Behavioral Targeting
With the advancement of technology, it has become possible to capture user behavior while surfing the web. Advanced tools are increasingly being deployed to capture user clicks, mouse movements, mouse-overs, time spent on various pages, etc. This data is collectively termed as “click stream,” and it is fed into complex algorithms to develop highly targeted user segments and clusters. These algorithms also take into consideration user purchasing decisions, age and interest information. This sort of behavioral targeting enables advertisers to reach a receptive, qualified audience with a highly relevant marketing message.

Advertisement serving engines are being developed to identify user clusters and deliver highly targeted advertisements. At a glance, behavioral targeting works by:

• Intelligently segmenting and qualifying users based on interests, behaviors, geography, searches made, keyword phrases, workplace attributes, etc.
  
  
• Reaching qualified audiences at the scale necessary to meet marketing goals.
  
  
• Providing relevant advertising in alignment with purchasing lifecycles.
  
  
• Eliminating waste and maximizing campaign efficiency.
  
  
• Producing better results at generally less cost than premium inventory
  

Social and Virtual World Advertising
According to one survey, nearly 70% of “tweens” (children between 8 and 12 years old) in the U.S. are registered with a social networking site (Facebook, MySpace, Hi5, etc.), and at least 50% of these users login twice per week. Similarly, the user virtual world known as Second Life is growing by hundreds of new registrants each day. Based on these statistics, social networking and virtual world websites are very potent vehicles for advertising. In fact, Facebook has created a new platform, Facebook Ads, to create and disseminate socially-aware advertisements on the website. Furthermore, many well-known names have bought real estate in Second Life to prominently display their logos throughout this virtual world.

Intelligent Advertisement Engines
Until recently, advertisement engines were pure play delivery tools that applied constraints as part of a campaign’s requirements (i.e., capping) and delivered the advertisement with no added intelligence. Today, elaborate algorithms are being incorporated into advertisement engines to maximize revenue. The newer advertisement engines can also capture data to provide advanced reports on the viability of a creative in terms of attracting clicks or users. Campaign managers can then analyze the graphical reports to check when a user response gets petrified and when to introduce a new banner.



SUMMARY
Using predictive modeling and non-personally identifiable user information, Internet advertising enables companies to greatly refine their marketing campaigns. As a result of the sophisticated new tools being developed for Internet ad targeting, companies can now deliver ad content that provides greater value at much lesser costs. As the benefits and revenues of Internet advertising continue to grow, it is only a matter of time before this form of marketing is regarded as mainstream as other commercial channels.