Sunday, 11 June 2017

What are different types of Attribution Models

In Digital Marketing, you might have often heard about Attribution. What is an Attribution? Attribution is giving the credit for purchase to the marketing channel. For example: if the customer has purchased a product by clicking on the Adwords(SEM) ad then the advertiser gives credit to the Adwords. You might be wondering why should we give Attribution Credit? Attribution Credit is required for measuring of Return on Investment(ROI) on the marketing campaigns or success of a marketing channel.



But in real-world, during the customer's journey of purchasing a product, there may be multiple touchpoints. A touchpoint can be a viewing/clicking on the ad. Since Adwords has promoted/influenced the customer to buy the product. Therefore, advertiser gives credit for the sale to Adwords. The advertiser uses multiple channels to reach their audience across multiple inventories available.

For example, Amazon may use various marketing channels such as Facebook campaigns, Adwords, Retargeting Campaigns, Email Campaigns, Instagram, LinkedIn, Twitter to reach their audience, etc,. It is difficult for the advertiser to credit for a sale when the users interact with multiple channels before purchasing the products. To credit for a purchase accurately, different attribution models have been introduced, so that advertiser can go after any attribution model. Let's look into the different kinds of Attribution models based on their touchpoints:
  • First Touch Attribution Model
  • Last Touch Attribution Model
  • Last Click Attribution Model
  • Linear Attribution Model
  • Position-Based Attribution Model
  • Time Decay Attribution Model
Before digging into the attribution models, let's take an example of an advertiser who is using multiple channels such as Adwords, Facebook, Twitter, LinkedIn, Email Retargeting(AdRoll) and customer's journey of purchasing the product of worth $1000. The customer searches for the product on google search, then clicks on the paid search ad and leaves the website without purchasing the product. After 1 day, he saw the ad on Facebook, Twitter, LinkedIn and clicked on the Facebook ad but didn't purchase the product at the time. Then, after 5 days, he viewed the AdRoll email ad and did not convert. Later, he purchased the product directly visiting the website. 

First Touch Attribution Model
In the First Touch Attribution Model, the advertiser gives the credit to the channel who initiated the sale. In the above example, since Adwords initiated the sale, the advertiser gives the credit of $1000 to the Adwords.

Last Touch Attribution Model
In the Last Touch Attribution Model, the advertiser gives the credit to the channel where the customer has last touched meaning that the customer has viewed the ad before purchasing the product. In the above the example, the advertiser has last touched AdRoll Email. Therefore the advertiser gives the credit of $1000 to the AdRoll Email.

Last Click Attribution Model
In the last click Attribution Model, the advertiser gives the credit to the channel where the customer has last clicked. In the above example: the advertiser gives the credit to Facebook as the customer clicked on the Facebook ad. Priority will be given to the channel where the customer has last clicked channel. Therefore, the advertiser gives the credit of $1000 to the Facebook channel though the customer has first clicked on Adwords and also viewed the AdRoll Email.

Linear Attribution Model
In Linear Attribution Model, the advertiser gives the credit equally to all channels that contributed for a purchase. In the above example, the advertiser gives the credit of 25%($200/channel) for each channel, AdRoll Email, Adwords, Facebook, Twitter, LinkedIn

Time Decay Attribution Model
In the Time Decay Attribution Model, the advertiser gives the highest percentage of credit to the channel that is closest to purchase. Here in the above example, the last touchpoint AdRoll Email, therefore the advertiser gives 50% credit to AdRoll Email channel, 20% to LinkedIn, 15% to Twitter, 10% to Facebook, 5% Adwords. Note: Percentages may vary depending on the advertisers choice.

Position-Based Attribution Model
In Position-Based Attribution Model, the advertiser gives the credit of 40% for purchase to the First and Last Touchpoints each and remaining 20% of the credit is distributed among rest of middle channels which are in between First Touchpoint and Last Touchpoint. In the above example, the advertiser gives credit of 40% to the Adwords, AdRoll Email each and distributes 20% of the credit to Facebook.

Mostly widely used Attribution model includes Last Click, Last and First touch attribution. Again, it is completely Advertisers choice to which attribution model, they have to go after to analyze the success of a marketing channel.

Friday, 19 May 2017

What is CRM data?

Generally, brands manage their customer data using CRM software to better control over customer data which helps sales team, marketing team, support team to take the faster decisions. Marketing team can leverage their customer data from CRM data to target their audience. but what is a CRM data? CRM stands for Customer Relationship Management and CRM data is nothing but information about the customers such as email address, phone number, postal address etc,.

CRM data
CRM data


CRM software is used in enterprise business to store the Customer data, Customer interaction, Tracking leads, Contracts, Employees etc,.. The CRM data can be used by various teams in an enterprise business such as sales team, marketing team and support team.  The CRM data is considered to be potential audience for marketing campaigns because they are already interested in your service. Examples of CRM software includes, Zoho CRM software, Salesforce CRM etc,.

Let's understand by considering an example, an advertiser has multiple electronic stores across the country and he collects their customers data such as phone number, email or postal address while purchasing the products for shipping the product to customer postal address. All the customers data is stored in the CRM software. He can leverage that data for marketing campaigns.  Now, let's say the advertiser is running a seasonal offer on his website and he wants target his display ads to past converted customers across the web and Facebook.

Many ad tech companies such as AdRoll, Facebook, Google Remarketing helps the advertiser to target the display ads. The advertisers have to upload the email address to ad tech companies. Under the hood, when they uploaded email addresses, the data is converted into human unreadable hash format. To target the visitors in a effective way through the marketing campaigns, ad tech companies appoint data on-boarding platforms such as LiveRamp, Datalogix, Conversantmedia, Neustar, Experian etc,. because ad tech campaigns has no idea about CRM data and their customer online behavior.

Data on-boarding platform analyze the CRM data and converts offline data into online data so that ad tech companies can use it for marketing campaigns.

Ad tech companies uses the information(hashed format) provided by the data on-boarding companies and starts cookie matching process across the Web and Facebook. The ad will be served to the customer who logged into Facebook or Google.

What you can do additionally with CRM data?

  • CRM Retargeting
  • CRM Prospecting

CRM Retargeting

CRM Retargeting is a kind of retargeting which allows the advertisers to retarget their audience based on the email address or postal address. Since CRM data is high quality traffic, visitors to tend engage with website audience. CRM Retargeting is an effective way targeting the audience using the marketing ad campaigns. With the help CRM retargeting, the marketers can target the tailored and personalized messages to the visitors which means the right customers at the right time.


Prospect the new customers 

Based on the existing data, ad tech companies use machine learning algorithms to create a lookalike or prospect new users. This helps the advertisers to increase the brand reach and attract new customers to their business.

if you got CRM data then you can directly contact the liveRamp or you can work with any DSP, Retargeting companies to target your audience.


Monday, 15 May 2017

What is Cross Device Advertising?

Cross device targeting enables the advertisers to target their audience across multiple devices such as Mobile, Tablet, Desktop and Laptop. For example: If a user has visited the Amazon website to purchase a Smartphone via desktop, then with the help of cross device targeting, it is possible to retarget that user on multiple devices such as Tablet, Mobile, Laptop etc,.

Cross Device Advertising
Cross Device Advertising

You might be thinking that when we are targeting the ads based on cookies on the desktop, we can also do that same thing on mobile based on the cookie then you're wrong, it is difficult to track the user identity using cookies because of the following challenges:

What are the challenges of Cross device targeting?
  • Whenever the mobile browser is closed, the browser will reset the cookies
  • Few mobile browsers doesn't allow the 3rd party cookies by default such as safari, however they do allow the 1st party cookie. Let's understand the what is 1st party cookie and 3rd party cookie. 
    • 1st party cookies are used by the website owners(the website which you have visited for example, you have visited Amazon website and Amazon.com has placed cookies and the cookie information is used by the Amazon.com for enhanced user experience). 
    • 3rd Party cookies are used by the website vendors(the visited website may also website vendors cookies, for example you have visited foxsports.com and advertisting.com website might use cookies of foxsports.com for targeting the relevant ads).
  • Mobile Apps doesn't share information with Mobile Apps/browsers as they mobile uses Appview concepts formally called sandbox where the information is shared within the mobile  app, Please refer to the below screenshot by IAB
availability of cookies on apps and browser, for android and ios
Availability of cookies on apps and browser, for android and ios
Keeping the goal of identifying or matching the right users with existing available data across multiple devices is broadly divided into two types:
  • Deterministic data in Cross Device Advertising
  • Probabilistic data in Cross Device Advertising

Deterministic Data in Cross Device Advertising

Deterministic Data is based on the user login(Personal Identifiable information) across the multiples devices and it is considered as the accurate data because when the user logged into mobile app or browser(we can track device ID, user login), and the same user is logged into other devices such as desktop, tablet browser then we can determine that the user is a unique using multiple devices. 

Therefore we can accurately target the user who is connected multiple devices based on the user-login. However, this is possible when you have large database such as Facebook, Google, Twitter, LinkedIn etc,. 
Deterministic Data
Deterministic Data

Probabilistic Data in Cross Device Advertising

On the other side of coin, Probabilistic Data is based on non user login(non Personal Identifiable information) which means we infer the data using different attributes such as IP address, browsing data, device information and prospect the user based on the existing data by creating data points using machine learning algorithms. Since, we are inferring the data based on various attributes may change for example, Operating System updates, it is not accurate as Deterministic Data. 

Probabilistic Data
Probabilistic Data

With the help deterministic data and probabilistic data, the ad tech companies target the right users across the multiple devices effectively with cross device advertising or cross device targeting

Advantages of Cross Device Advertising

  • It solves cross channel tracking issues meaning, before the invent of cross device tracking, ad tech companies used track as single conversion as two different conversions, one for mobile and one for desktop. It eliminates the de-duped conversions
  • Cross device helps understand and analyze customer behaviors 
  • It helps to serve impressions effectively and utilize the multi-channel inventory
  • Marketers can spend effectively for each impression served to the visitors
  • Marketers can able to set up the sequential messages

Disadvantages of Cross Device Advertising

  • Sometimes, in a family, people tend to share the devices and ads may be served to other members of family. For example, in a family, if X is logged into the tablet browser and Y is browsing the web using X logins on tablet, then the ads will be served to the X across multiple devices. This is a rare scenario in cross device advertising
  • To identify the user accurately, you need to huge amount of data.

It is important to refrain the number of times an ad is shown to the user because the user fed up viewing the ads across all devices and also save the impressions from being wasted when Cross device targeting is implemented. As of now, there are many ad tech companies which offers cross device targeting such as AdRoll, Drawbridge Criteo, Facebook, etc,.

Tuesday, 25 April 2017

What is an Ad Exchange?

Ad Exchange allows the advertisers to programmatically bid on each impression in just fraction of seconds through Real-time bidding. Ad Exchange is a nest of Ad Networks, Ad Exchanges, DSP's.

Ad exchanges eliminates manual negotiation process and advertiser don't need to directly contact the publisher for buying the space

What is an Ad Exchange?
What is an Ad Exchange?


There are two types of Ad Exchanges:
  1. Public Ad Exchange
  2. Private Ad Exchange

Public Ad Exchange

Public Ad Exchange is also know as Open Ad Exchange.Through Public Ad Exchange, billions of impressions are served every day. Publishers sold their inventory to the advertisers whoever wins the bid in Real Time Bidding. Publisher never know which ads are served on their website and at the same time, advertiser doesn't on which websites their ads are served. It is not possible to track each all the information because we don't know where the impressions are served.

Examples of Ad Exchanges are AppNexus, AOL, DoubleClick for Ad Exchange, OpenX etc,.

Private Ad Exchange

Publishers couldn't able make more profits as premium inventory was sold at a very low cost to the advertisers. Through Private Ad Exchange, Publishers can carefully control who can buy their inventory and at what price. On the other side of coin, it allows the advertisers to carefully choose their publishers high-quality inventory t a premium price and protect themselves from brand safety. Private Ad Exchange is also know as Private Ad Exchange or Private MarketPlace(PMP). 

Private Ad Exchanges are formed by hand picked premium ad networks, DSP,  Ad Networks, Ad exchanges and by small other Private ad exchanges to serve guaranteed impressions. Private Ad Exchange is considered as an safer environment for 1st party data.

Sunday, 13 November 2016

Difference between Ad Network and Ad Exchange

Though the main objective of the Ad Network and Ad Exchange are the aggregation of the ad space from the publisher website and selling them to the advertiser and sometimes they sounds similar but there are few differences between the Ad Network and Ad Exchange:






Ad NetworkAd Exchange
Ad Network is an aggregator of unsold ad inventory from the publisher websites and sell them to the advertiser at a predefined priceAn Ad Exchange is an aggregator of ad space from the publisher website and sell them to the advertiser based on the Real Time Bidding
Ad inventory is sold in packages to the advertisersAd Inventory is sold in Real Time Bidding to the advertiser
The advertisers cannot bid high for the potential customer instead, they can only bid ad space from the publisher at a predefined price set by the Ad networksAdvertiser can bid high amount for a potential customer
Each impression on the publisher website is bought for a fixed priceEach impression on the publisher website is bought for a different price
Advertisers don't have control over ads.Advertisers have control and can target a specific audience by bidding.
Ads are targeted based on the contextual targetingAds are targeted based on the Behavioral targeting meaning ads are targeted based on browser cookie history
Ex: Tribal Fusion(exponential), Adsense, Yahoo Ad Network etc,.Ex: AppNexus, BidSwitch, OpenX, DoubleClick for Ad Exchange etc,.
Best suited for small publishersBest suited for large publishers
A publisher can earn less when compared to the Ad ExchangeThrough Ad Exchange, Publishers can earn more money using Open Exchange and Private Marketplace

Top 10 videos to learn about Online advertising

In the recent years, the internet is evolved in such a way that we can't image our world without it. It is an inseparable part of our life. Every device either it may be Mobile or an ATM's are interconnected to the internet.

In this world, billions of the people use internet services for various reasons such as seeking information or for watching movies, listening songs, checking an email etc,.This has opened the doors for a huge marketplace in the advertising industry. The ad industry landscape is very large and it is very difficult understanding the whole workflow.

If you are new to online advertising or if you have been associated with any one of the teams in the online advertising then it might be difficult to understand. You might have googled and ended up with nothing.

Below are top 10 YouTube videos to learn about online advertising quickly:




The Evolution of Online Display Advertising

This video has great content about Evolution of online advertising by Iabuk which lasts for less 5 mins. This video answer many questions such as How Ad Networks are emerged and what are the disadvantages of the Ad Networks. How Ad Exchanges overcome the limitation of the Ad Networks and how each ad tech companies(DSP's, SSP, Ad networks, ATD) are interconnected with each other.

Online Ads 101: The history of online ads

This video has useful content about the history of online advertising. The video begins with history about advertising, how the advertising is in today's world. This video is made by Online Ads Edu. This channel helps the enthusiasts to learn about online advertising.

200ms: The Life of a Programmatic RTB Ad Impression

This is great video by MediaCrossing which gives you the complete understanding of RTB and how it works within 200ms. The video in the background is awesome

Real Time Bidding 101

This video is made by Acuity presented by Marco Muzzi, he has great understanding of Ad Industry and you find answers to the questions like what is real-time bidding? How does real-time bidding work? What are the benefits for digital marketers?


How an Ad is Served with Real Time Bidding (RTB) - IAB Digital Simplified

This video is made by IAB - Interactive Advertising Bureau. This video gives you an overview of how ads served to the users and how exactly each party involved are clearly mentioned in the video.


Intro to Programmatic Advertising - Part 1

This video is made by Meteora. Here in this video, the presenter gives an overview of ad technology by taking a live example of Nike brand and he also relates each parties responsibilities. The presenter has presented with video by using Photoshop tool.

Intro to Programmatic Advertising - Part 2

This video is made by Meteora. In this video, the presenter takes you to in-depth of how exactly DSP's bids on the potential customer. The presenter has more emphasized on DSP functionality.


Real-Time Bidding (RTB) Basics

This video is made by Donart Nummer. In this video, Donart explains how RTB works with a simple explanation. This video great content easy to understand and video length is 9 mins.


Display Advertising Basics (DSPs, RTB, Ad Exchanges, DMPs)

This video is presented by Pete Kluge. This is part 1 of the video series on Display Advertising Basics (DSPs, RTB, Ad Exchanges, DMP's). Pete Kluge clearly explains about basics of the online advertising or display advertising and how each party involved in the bidding. The length of the video is 16 mins.

Display Advertising Basics P2 (RTB, DSP, Programmatic, Private Exchanges, PMP)

This video is presented by Pete Kluge. This is part 2 of the video series on Display Advertising Basics P2 (RTB, DSP, Programmatic, Private Exchanges, PMP). Pete Kluge explains about scenario's by considering premium inventory sold to Private Marketing Platform and Private exchanges via Programmatic guaranteed direct buys.

Bonus: You can also watch some more at MediaMath Training videos. Don't miss to watch the video on RTB.

Happy Learning :)

Saturday, 12 November 2016

What is Demand side platform?


DSP is short form for Demand Side Platform. Demand Side Platforms are web-based tool which helps the advertisers or agency's to buy the ad space from the publisher website. Ex: MediaMath, Turn, DoubleClick Bid Manager etc,. 

The end user of the DSP is an Advertiser or Ad Agency that allows the access to the real-time bidding through ad exchange, where the each ad will be served through bidding and the winning ad will serve an impression on the publisher website. 

DSP's are enhanced version of the Ad Network because they offer additional features
  • Campaign Management
  • RTB capability
  • Tracking the performance
  • Optimizing the campaign performance
  • Transparency and control
  • Better Reporting
For Example:

AdRoll is an aggregator of Facebook, Web, Yahoo Right Media, AppNexus etc,. An advertiser can target the audience on Facebook, Web, AppNexus by launching a single campaign through AdRoll instead of launching 3 different individual campaigns in Facebook ad manager, Adwords and in AppNexus.

DSP's usually bid on CPM basis and will take a certain of part of the commission from the advertiser spend. In this way DSP's makes money.


What is Demand side platform?


To better understand let's take an example of DSP:
  1. Users or visitors request for a publisher website in the browser and the browser will send a request to web server then web server will send ad tags along with HTML Code. Ad tags will send ad request to the publisher ad server. Along with ad request some additional user information will be sent to Supply Side Platform(SSP). SSP's will send a request to the ad Exchange for further bidding(To understand clearly or to keep diagram simple, the above diagram doesn't contain information about SSP's)
  2. Along with bid request Ad Exchange will passes user information to the DSP's to bid for the ad slot on the publisher website
  3. DSP's takes information provided by the Ad Exchange and request DMP's to provide more information about the user, so that DSP can high amount of website visitor
  4. DMP's process the DSP's request and will send the information User such as Age, Gender and more information about the users
  5. DSP's will estimate the bid amount based on the relevancy of the advertiser product. If the likelihood of clicking an ad or buying a product is high then DSP's will higher amount to win the bid. On the contrary, DSP's will bid low amount if the likelihood of clicking an ad is low.
  6. Similar to the DSP's(this DSP) many other DSP's also participates in the Real-Time Bidding
  7. Winning ad via Real-Time Bidding will be passed as an ad response to the publisher website and ad will be displayed on the publisher website.
Each DSP's have different functionality and works in the different way, but ultimate goal of the DSP's to bid high amount for potential customer using their logic.