The Dark Side of Marketing Attribution

Throughout my career at ANNUITAS, I have probably spent more time thinking about and enabling Marketing Attribution models than any other activity. Over the years, I have seen Marketing Attribution models transform from a nice-to-have tool to help demonstrate marketing’s performance into a must-have core business planning component. To understand the outcome that more and more business are trying to move towards, it makes sense to first examine the fundamentals of Marketing Attribution models.

The Interactive Advertising Bureau, a conglomerate of 650 leading media and technology companies, defines attribution as:

The process of identifying a set of user actions (“events”) across screens and touchpoints that contribute in some manner to a desired outcome and then assigning value to each of these events.

When it comes to Demand Marketing, “Marketing Attribution” is the categorization process that aids in building rules to determine which leads and opportunities came from marketing.

While it seems simple enough, Marketing Attribution falls into what I would describe as “marketing philosophy.” As much as I would love for this topic to be completely objective, unfortunately Marketing Attribution models are quite the opposite. Many sales and marketing leaders will outright reject this hard truth at first, as they want to believe that there is a “perfect” model out there that is completely beyond questioning. Rather than unrealistically overdesigning their attribution model to be completely impervious to all attacks, sales and marketing leaders should focus on coming together and finding common ground. Marketing Attribution models don’t draw their power from capital-T “Truth” – instead they become powerful tools through the collective trust of all business stakeholders impacted by the model.

To earn this trust, sales and marketing leaders need to work as a team to determine what they all agree is driving impact and build their model out with those value judgments as a starting point. This magically changes the corporate posture surrounding attribution models from one of attack-and-defense to that of an alliance focused on deliberate optimization.

I have seen these key mistakes made in the construction of Marketing Attribution models that poison the model and make it misleading. These mistakes include:

  • Stretching the same model to serve too many purposes
  • Lax rules that allow too many false positives to come through
  • Rigid rules stuck in outdated thinking
  • Losing sight of the outcome being measured by the KPI

When mistakes are made and business leaders do not present a unified front to Demand Marketing, this can create a Marketing Attribution model that actually encourages people to optimize towards meeting the wrong metrics and not focus on driving real growth.

One Model to Rule Them All?

The most common mistake I see made with Marketing Attribution model design is stretching the scope of the model so far that it fails to do any one thing particularly well. Marketing Attribution models are often burdened with serving dual functions:

  • Creating trust in data and alignment across multiple business stakeholders [Executive Model]
  • Optimizing investment to capitalize on predictive indicators of future buying intent [Operational Model]

To help support each Demand Marketing leader’s day-to-day, attribution models need to exist that help drive insight into the effect of individual marketing tactics on the big picture in order to ensure every dollar is wisely invested. To do this, Demand Marketing leaders will often need to lean on their data science and analytics teams to build models using advanced methods to help make better decisions about where to invest.

Under no circumstances should these advanced models make their way into the board room. While it may be tempting to build one model to serve both masters, combining models focused on intense technical analysis with models focused on a 10,000-foot executive overview will only serve to undermine trust in these models.

As somebody that has made this mistake myself, attribution models designed to help create trust in data and alignment on performance across different business stakeholders need to be accessible enough that a person from any walk of life or field of study can not only understand it, but wholeheartedly believe in its ability to measure performance.  Without this level of accessibility, all an attribution model will do is foster mistrust of the data and send each business leader back to Salesforce to create a myriad of conflicting reports supporting a variety of different worldviews.

If you build your Operational Model correctly, its efficacy should speak for itself by translating into direct impact on the Executive Model.

If Everything Is From Marketing, Nothing Is From Marketing

In our CEO, Adam Needles’s, blog post on this topic, one important factor he outlines as necessary to measuring marketing’s impact on the bottom line is causation.

The question is whether these activities were ‘causal.’  I.e., did demand marketing activities lead to a positive sales outcome, versus occurring in tandem with (or after) the positive sales outcome.  If there is net sales lift, then demand marketing is causal.

Sadly, the majority of Marketing Attribution models that I’ve seen are built favoring correlation as opposed to causation.

The most common reason Marketing Attribution models are built this way is because companies frequently don’t collect the right data, or have it structured correctly in order to drive deeper insight. At ANNUITAS, our most fundamental axiom is that strategy should not be dictated by technology.

Frequently, organizations will approach their Marketing Technology stack using Salesforce as its central database. If you’ve ever heard somebody say “If it’s not in Salesforce, it’s not real,” they work at a company with this kind of structure. Here, a SFDC campaign related to the Opportunity is “good enough” to qualify as attribution because Salesforce alone cannot track and drive more precise marketing attribution models. This philosophy fails to consider:

  • The substantivity of the interaction
  • If the interaction was what created the Qualified Lead or Opportunity
  • If the Person on the campaign is even on the buying committee

The outcome is a world where anything with a pulse is considered “Marketing Sourced.” Demand marketers aren’t incentivized to think about how to drive value – instead they are encouraged to find ways to add campaign members to opportunities. The solution to this is orchestrating people, process, and technology to create a more granular data model that enables precise attribution. Once “false positives” are cast away, it is much easier to both build a foundation of trust in Marketing Attribution across stakeholders and optimize towards the right goals.

2005 Called. They Want First-Touch Attribution Back.

Buyer behavior has changed dramatically in the last two decades. As BrightEdge has noted, the balance of power has shifted from seller to buyer:

Before most people started using the Internet, customers interacted with brands largely on the company’s terms. Consumers had to wait for businesses to reach out to them in their buyer journey with information about the latest products… Now 67% of the journey takes place online, where the customer is in control. They expect to receive personalized care and they want to receive the information desired on their schedule.

Back in the day, first touch attribution models made sense because marketers served as rainmakers – they simply needed to reach out to leads that had few options to address their pain points and even fewer options with which to do research.

Today, the opposite is true. Buyers are paralyzed by the tyranny of choice when it comes to competing products trying to solve the same pain point and are expected to defend their purchase decisions with mountains of research. There are fewer excuses for making a poor purchase on the corporate card.

Marketing needs to have a heavily personalized answer to all points of the buying process for all members of the buying committee – because if they don’t somebody else will. This speaks to the need to account for Demand Marketing’s efforts at all phases of the funnel, not just acquisition. Yet disturbingly, a recent study by Tune found that 43% of marketers are still using First Touch as their primary method of attribution.

If Demand Marketers are forced to use old models, they’ll be stuck using old strategies to compete, leaving their companies as the biggest loser.

Hate the Game. Not the Player.

A successful Demand Marketing leader is uniquely positioned such that they would both directly benefit from the degree to which they grow the business or share in the company’s pain should they fail. Because of this, they have many kinds of incentives and disincentives forcing them to think about how their actions affect the final outcome and not just the number they need to hit. This is a concept that Nassim Nicholas Taleb, author of Antifragile and The Black Swan, refers to as “having skin in the game”:

…the balancing of incentives and disincentives, people should also be penalized if something for which they are responsible goes wrong and hurts others: he or she who wants a share of the benefits needs to also share some of the risks.

Unfortunately, not every company can organize themselves such that everybody has equal skin in the game. This makes it essential to ensure that Marketing Attribution models are as reflective of the real buyers journey as possible, as it can allow those that support leaders with skin in the game to drive outcomes.

If Marketing Attribution models radically diverge from “reality,” business leaders leave themselves open to their employees gaming the system – either consciously or unconsciously. For example, if criteria is too loose, it is possible for people to “buy” attribution to meet their quarterly goal through high-volume, low-quality channels—such as Content Syndication—to create “technically marketing sourced” opportunities. Suddenly marketing budgets stop becoming investments into revenue growth and instead become another brick in the wall protecting certain peoples’ bonuses.

Unfortunately, people realize that the Marketing Attribution rules have a direct impact on how easy or difficult it will be for them to meet their target which can also lead to heavy politicization of Marketing Attribution at organizations. To keep this from happening, focus on keeping Attribution models as true to “reality” as possible and giving different leaders skin in the game to incentivize working towards the outcome as opposed to the number.


At the end of the day, Demand Marketing leaders with skin in the game recognize what’s important isn’t how much marketing contributes versus sales or partners, but instead that everybody is driving towards the same business growth. As our Chief Revenue Officer, Lauren Goldstein, notes:

I think there’s a real benefit when operations functions are aligned to the revenue goal…when everyone is marching in the same direction there’s no question about what success looks like.

Marketing Attribution models are essential for helping define at all levels of the business what marketing investment is translating into growth. This can be achieved by properly defining the scope, building trust across all parties, ensuring the business rules are consistent with reality, and incentivizing outcome-oriented actions.

More Great Content