The Problem with Marketing Analytics (and Why Traditional Tools Aren’t the Answer)

If you’ve been in marketing as long as I have (going on 15 years now), it probably doesn’t seem like long ago that marketing was primarily viewed as a creative discipline. In those days, marketing departments were filled with writers, designers, and market researchers — and the function’s primary responsibility, with the exception of sales enablement, centered on building the brand and owning top-of-funnel activity.

Pretty amazing how things have changed.

Today, marketing’s ownership of the buyer’s journey is increasing rapidly. We no longer create content and campaigns only for the sake of generating awareness. We do those things to contribute to organization-wide objectives and to help prospects self-educate until the very end of their buyers’ journeys.

For CMOs, this bump in responsibility and accountability has proven to be a double-edged sword. On the plus side, they’re being given larger budgets than ever before. On the downside, they’re also now being asked to track and attribute exactly how those increased investments are contributing to revenue growth — much like a VP Sales is held accountable for an incremental rep.

Trying to Answer Important Questions with Imprecise Marketing Analytics

As a former CMO, I feel that pain. In fact, in 2011, I remember standing in front of the Board of Directors of the late-stage SaaS company that had recently hired me, preparing to address questions I wasn’t really confident I knew the answers to.

Why is the company’s revenue fluctuating? What’s causing pipeline surges and dips, and what can be done to make it more consistent? How effective are the campaigns we’re executing, and which ones should we be investing more into? How will what you’re doing impact revenue growth next quarter?

I knew the only thing worse than giving the wrong answer was not having one at all. So, I answered them as best I could — with presumptive responses that were largely based on hunches and loose interpretations of available data. Thankfully, it worked. The data told enough of a story that the board let me live. Deep down, however, I wasn’t sure I’d really said much of anything.

Sound familiar?

The truth is, data can be incredibly powerful and it can help paint a picture of how marketing campaigns are influencing revenue results. In fact, according to a recent Forbes article, data-driven marketing leaders are almost three times more likely to have increased revenues.

The issue, however, is that even for the savviest marketer, acquiring the right data and analyzing it in the right context can be incredibly difficult — particularly as marketing becomes more and more involved in the middle and later stages of the buyer’s journey.

The Challenge: Full Sales Cycle Visibility

Many marketing technologies (e.g. CRM, marketing automation systems, content curation tools, etc.) offer plenty of top of funnel reporting, but that insight typically only provides very basic campaign influence metrics — and it does nothing to explain what’s happening in the middle 80% of the customer lifecycle.

To combat this, marketers typically lean on one (or several) other attribution models to supplement first and last touch, and create multi-touch attribution:

• Evenly-weighted (linear) attribution: An improvement over single-touch, linear attribution makes it easy to apply credit to the middle stages of the funnel. The problem: While this approach does apply credit to all touches along the buyer’s journey, it runs the risk of overvaluing lower impact touches, which can lead to faulty assumptions about the effectiveness of a campaign.

• Time decay attribution: The core premise of the time decay model is that the closer a touch point is to conversion, the more credit it should receive. While that’s fine in theory, this model on its own sometimes fails to capture critical contextual information that might allow for better campaign assessments.

• Position-based attribution: In this model, greater revenue credit is given to specific touches in the cycle (typically the first and last touches), while the remainder of the credit is divided among mid-funnel activities. This model effectively gives credit where it’s due (first and last touch) and factors in the influence of mid-funnel campaigns.

• Interaction-based attribution: A custom interaction-based attribution model relies on historical analysis to apply different weights to varying interactions. The challenge with relying only on interaction-based models is that they’re often subjective, and marketers must put considerable thought into which types of user behaviors are most valuable.

While each of those models can provide CMOs better visibility into campaign performance on their own, they’re most powerful when combined into an all-encompassing attribution model. This aggregate approach allows CMOs to gather deeper insight into their buyer’s journey, attribute revenue impact with greater precision, and shed light on the ideal sequence of campaigns to most efficiently — and cost-effectively — convert a lead into a closed sale.

Welcome to the Era of Multi-Touch Attribution

What I’m describing above is multi-touch revenue attribution — a model that incorporates both campaign and revenue data to help marketers evaluate performance across every touch point in the buying process, calculate true ROI, identify what’s working, and immediately make quantifiable decisions for their organization. While that’s not necessarily a new concept, for most senior marketers it’s an approach that’s always proven far too complex or costly to tackle.

The good news? This is the year that changes. Seriously.

Click here for my next post, where I describe four key conditions that will make 2015 the year that multi-touch attribution goes mainstream, and I’ll explain how your organization can take easily take the steps to embrace it.

Photo by Tim Green