Blog Post:This week marks 31 years since the pilot release of . . . dare I say . . . the most epic and profoundly influential television series to ever grace us (or at least me): He-Man and the Masters of the Universe.  The only semblance of criticism I can make about the television series is that it released only 10 days before my second Halloween and the He-Man costumes were not available for another year. Each week during these formidable years of my childhood, Prince Adam from the fictional planet of Etheria would raise his Sword of Power and I would recite along with him as he proclaimed, “By the Power of Grayskull! I HAVE THE POWER!” transforming Prince Adam into He-Man, the most powerful man in the universe.  With this endowment of power, He-Man battles the evil forces of Skeletor to save planet Etheria and to protect the secrets of Castle Grayskull. As digital analysts and marketers, you repeatedly face the challenge of accurately measuring the impact of your marketing investments—a process known as attribution. Marketers are continually asked to demonstrate the effectiveness of their marketing decisions across paid, earned, and owned channels. Increased spend of digital marketing dollars has led to the critical need to measure and manipulate the channel controls and customer interactions that impact business goals. Attribution is the process of measuring and assigning credit to touchpoints and customer interactions that influence a customer along the entire path to conversion. As part of our bold reinvention of the Adobe Analytics offering, our vision is to help make the enterprise smarter. To this end, Adobe has invested and continues to invest heavily in providing world-class advanced analytics capabilities designed for the non-quant—empowering both marketers and analysts with powerful data mining and machine learning capabilities.  Among the many capabilities we have recently released, expanding our attribution offering has been a major strategic theme. The Value and Benefits of Attribution Much like the endowed power from the Sword of Power that transformed Prince Adam into He-Man (or the Sword of Protection for She-Ra: Princess of Power for those of you who relate more to Princess Adora), the comprehensive attribution capabilities recently released in Adobe Analytics juxtapose traditional cross-channel marketing attribution with comprehensive customer analytics to provide a holistic understanding of each interaction on driving business success. You do not have to settle with only understanding how top of the funnel marketing decisions impact bottom of the funnels sales without any understanding of the interactions in between. Adobe Analytics applies attribution to each interaction along the customer journey leading toward conversion. Attribution provides valuable insights into the impact of your marketing investments and the value of each customer interaction along the customers’ journey toward conversion. Marketers use attribution for a range of objectives including (but not limited to): Ok, let’s move past the fun anecdotes (put your swords down for a second) and drill down into the details of all of the attribution models you have at your disposal and best practices for each approach. Attribution Models within Adobe Analytics My colleague, Jeremy King, recently published a blog post outlining the capabilities released in our most recent release of Adobe Analytics, including an overview of Best Fit (or Algorithmic) Attribution. There are two core approaches to cross-channel attribution: 1) rules-based (subjective) attribution models, and 2) data-driven/algorithmic (objective) attribution models. The following illustration depicts each of the standardized attribution models currently available in Adobe Analytics.

JBates-Table 1

Algorithmic Attribution Explained Although each of the rules-based approaches are fairly straightforward, the algorithmic attribution model used within Adobe Analytics applies advanced statistics and machine learning to objectively determine the fractional impact of each marketing touch along a customers’ journey toward conversion. In an effort to restrain my nerdy side from diving deep into the algorithms, I will simply say that the algorithms we use fall into a class known as multivariate regression and classification models. The real value of using the algorithmic attribution approach instead of the rules-based methods is that you do not have to guess how to allocate the impact of each marketing interaction. You just saved countless hours of trial-and-error guessing. Additionally, only with Adobe Analytics can you combine marketing channel touches with owned touches (e.g., website interactions). In this way, each and every customer interaction is being analyzed for its impact on conversion apples-for-apples so you can optimize the entire customer experience. Which Attribution Model Is Right for Me? You may be asking yourself, “self, which attribution model should I choose?” Because of the inherit biases each of the single and multitouch rules-based models introduces, I almost always encourage our clients to adopt the objective, data-driven algorithmic attribution model. In reality, there are many valid reasons for both exploring and adopting the rules-based models instead. For example, if the CMO asks for a consistent measurement strategy across all digital marketing efforts, the algorithmic attribution model is a great option. However, if the email marketing manager needs to know how many days before an event should an email blast be sent to a set of customers, the latency attribution will help provide the needed insight for the decision while the algorithmic attribution model continues to serve as the “single source” of truth for measuring marketing effectiveness across the business. Generally, I have found that the decision of which attribution model to use is largely determined by a combination of the organization's analytical maturity, campaign strategy, and the general length of the customer consideration period. Although the end game should be to move your analytical maturity along as quickly as possible toward taking an algorithmic approach to attribution, the following represents a mapping for helping you identify which attribution models might make the most sense today:

JBates-Table 2

The following are four examples that illustrate how you might want to evaluate different attribution models within the context of your business:

JBates-Figure 1

Once you have identified a couple of candidate attribution models, it is always a good idea to review the outputs/implications of each model with stakeholders within your organization. Executive and stakeholder buy-in is critical in adopting an attribution solution; otherwise, you end up with marketers using their own "favorite" model that shows the absolute best performance for their channel even when it may not reflect reality.  Remember that this attribution model evaluation map is only a general way to evaluate some of the more common factors that influence which attribution model to adopt. There are obviously many potential edge cases and circumstances that could impact your own evaluation. It is a healthy exercise to compare/contrast the outputs from many attribution models during your initial exploration phase before you settle in on a single approach. It is a best practice to eventually decide upon one attribution approach so as to create consistency in measurement over time across the entire organization. With the Sword of Attribution Power, you are now Masters of the "Marketing" Universe—the most powerful analysts/marketers in the world. Say it with me . . . “By the Power of Attribution! I HAVE THE POWER!”

Disclaimer: I am not liable for any coworker hospitalizations resulting from uncontrollable outbursts as you begin using Adobe Analytics like “By the Power of Attribution! I HAVE THE POWER!” at your place of employment. You will learn to harness the power overtime. Adobe partially shares with you in the liability of any and all amazing analyses you complete with this newly found attribution power and the pure AWESOMENESS of data-driven insights you now provide to your business.

Adobe’s commitment to innovation in reinventing the marketer shows in our aggressive approach to making the enterprise smarter through advanced analytics. Combined with our other investments, Adobe Analytics provides the most comprehensive analytics and attribution solution in the market.  
Author: Date Created:October 27, 2014 Date Published: Headline:Become a Master of the “Marketing” Universe with Cross-Channel Attribution Social Counts: Keywords: Publisher:Adobe Image:https://blogs.adobe.com/digitalmarketing/wp-content/uploads/2014/10/sb10062905w-001-e1414096611453.jpg

This week marks 31 years since the pilot release of . . . dare I say . . . the most epic and profoundly influential television series to ever grace us (or at least me): He-Man and the Masters of the Universe.  The only semblance of criticism I can make about the television series is that it released only 10 days before my second Halloween and the He-Man costumes were not available for another year.

Each week during these formidable years of my childhood, Prince Adam from the fictional planet of Etheria would raise his Sword of Power and I would recite along with him as he proclaimed, “By the Power of Grayskull! I HAVE THE POWER!” transforming Prince Adam into He-Man, the most powerful man in the universe.  With this endowment of power, He-Man battles the evil forces of Skeletor to save planet Etheria and to protect the secrets of Castle Grayskull.

As digital analysts and marketers, you repeatedly face the challenge of accurately measuring the impact of your marketing investments—a process known as attribution. Marketers are continually asked to demonstrate the effectiveness of their marketing decisions across paid, earned, and owned channels. Increased spend of digital marketing dollars has led to the critical need to measure and manipulate the channel controls and customer interactions that impact business goals. Attribution is the process of measuring and assigning credit to touchpoints and customer interactions that influence a customer along the entire path to conversion.

As part of our bold reinvention of the Adobe Analytics offering, our vision is to help make the enterprise smarter. To this end, Adobe has invested and continues to invest heavily in providing world-class advanced analytics capabilities designed for the non-quant—empowering both marketers and analysts with powerful data mining and machine learning capabilities.  Among the many capabilities we have recently released, expanding our attribution offering has been a major strategic theme.

The Value and Benefits of Attribution

Much like the endowed power from the Sword of Power that transformed Prince Adam into He-Man (or the Sword of Protection for She-Ra: Princess of Power for those of you who relate more to Princess Adora), the comprehensive attribution capabilities recently released in Adobe Analytics juxtapose traditional cross-channel marketing attribution with comprehensive customer analytics to provide a holistic understanding of each interaction on driving business success. You do not have to settle with only understanding how top of the funnel marketing decisions impact bottom of the funnels sales without any understanding of the interactions in between. Adobe Analytics applies attribution to each interaction along the customer journey leading toward conversion.

Attribution provides valuable insights into the impact of your marketing investments and the value of each customer interaction along the customers’ journey toward conversion. Marketers use attribution for a range of objectives including (but not limited to):

  • Understand marketing’s value to the business
  • Quantify the impact of marketing spend for each channel
  • Understand marketing effectiveness at all stages of the customer sales/life cycle
  • Understand the interactive synergies across marketing touchpoints
  • Know the performance of each tactic, device, and audience across channels and understand the interplay
  • Develop the right interactions to increase engagement
  • Know the best possible customer purchase path and experience
  • Develop more informed strategies to get the most out of marketing dollars across media and channels
  • Ensure impartiality through a data-driven, algorithmic approach to measuring attribution

Ok, let’s move past the fun anecdotes (put your swords down for a second) and drill down into the details of all of the attribution models you have at your disposal and best practices for each approach.

Attribution Models within Adobe Analytics

My colleague, Jeremy King, recently published a blog post outlining the capabilities released in our most recent release of Adobe Analytics, including an overview of Best Fit (or Algorithmic) Attribution.

There are two core approaches to cross-channel attribution: 1) rules-based (subjective) attribution models, and 2) data-driven/algorithmic (objective) attribution models. The following illustration depicts each of the standardized attribution models currently available in Adobe Analytics.

JBates-Table 1

Algorithmic Attribution Explained

Although each of the rules-based approaches are fairly straightforward, the algorithmic attribution model used within Adobe Analytics applies advanced statistics and machine learning to objectively determine the fractional impact of each marketing touch along a customers’ journey toward conversion. In an effort to restrain my nerdy side from diving deep into the algorithms, I will simply say that the algorithms we use fall into a class known as multivariate regression and classification models. The real value of using the algorithmic attribution approach instead of the rules-based methods is that you do not have to guess how to allocate the impact of each marketing interaction. You just saved countless hours of trial-and-error guessing.

Additionally, only with Adobe Analytics can you combine marketing channel touches with owned touches (e.g., website interactions). In this way, each and every customer interaction is being analyzed for its impact on conversion apples-for-apples so you can optimize the entire customer experience.

Which Attribution Model Is Right for Me?

You may be asking yourself, “self, which attribution model should I choose?” Because of the inherit biases each of the single and multitouch rules-based models introduces, I almost always encourage our clients to adopt the objective, data-driven algorithmic attribution model. In reality, there are many valid reasons for both exploring and adopting the rules-based models instead. For example, if the CMO asks for a consistent measurement strategy across all digital marketing efforts, the algorithmic attribution model is a great option. However, if the email marketing manager needs to know how many days before an event should an email blast be sent to a set of customers, the latency attribution will help provide the needed insight for the decision while the algorithmic attribution model continues to serve as the “single source” of truth for measuring marketing effectiveness across the business.

Generally, I have found that the decision of which attribution model to use is largely determined by a combination of the organization’s analytical maturity, campaign strategy, and the general length of the customer consideration period. Although the end game should be to move your analytical maturity along as quickly as possible toward taking an algorithmic approach to attribution, the following represents a mapping for helping you identify which attribution models might make the most sense today:

JBates-Table 2

The following are four examples that illustrate how you might want to evaluate different attribution models within the context of your business:

JBates-Figure 1

Once you have identified a couple of candidate attribution models, it is always a good idea to review the outputs/implications of each model with stakeholders within your organization. Executive and stakeholder buy-in is critical in adopting an attribution solution; otherwise, you end up with marketers using their own “favorite” model that shows the absolute best performance for their channel even when it may not reflect reality.  Remember that this attribution model evaluation map is only a general way to evaluate some of the more common factors that influence which attribution model to adopt. There are obviously many potential edge cases and circumstances that could impact your own evaluation.

It is a healthy exercise to compare/contrast the outputs from many attribution models during your initial exploration phase before you settle in on a single approach. It is a best practice to eventually decide upon one attribution approach so as to create consistency in measurement over time across the entire organization.

With the Sword of Attribution Power, you are now Masters of the “Marketing” Universe—the most powerful analysts/marketers in the world. Say it with me . . . “By the Power of Attribution! I HAVE THE POWER!”

Disclaimer: I am not liable for any coworker hospitalizations resulting from uncontrollable outbursts as you begin using Adobe Analytics like “By the Power of Attribution! I HAVE THE POWER!” at your place of employment. You will learn to harness the power overtime. Adobe partially shares with you in the liability of any and all amazing analyses you complete with this newly found attribution power and the pure AWESOMENESS of data-driven insights you now provide to your business.

Adobe’s commitment to innovation in reinventing the marketer shows in our aggressive approach to making the enterprise smarter through advanced analytics. Combined with our other investments, Adobe Analytics provides the most comprehensive analytics and attribution solution in the market.