Marketing Correlations: A Better Way to Measure Marketing Operations


Accountability and financial performance has shifted away from accountants and financial advisors. It’s now up to marketers to measure the effectiveness of their activities and justify the money they spend.

This short article highlights the traditional way marketing performance was measured, how marketing performance has been affected by CRMs, and a way that marketers can extract greater insight towards their customers’ experience with their brand.

By focusing on conversion correlations, marketers will be able to identify which edits to make to their marketing strategy, promotions, and budget.

The Problem with Current Marketing Analytics

Marketers spend money on ideas.

They try new ideas, new campaigns, and new events to see if they can acquire more customers. In this process, they try spending money on advertisements, ad creative, ebooks, and elaborate events.

Traditionally, the measurement of whether or not marketing improved revenue was determined by managerial accountants and financial advisors (e.g. marketing spend increased, and revenue was up this quarter).

But as more tech-savvy marketers emerge, a new onus has moved from finance to marketing — for marketers to measure and optimize their own operations. It’s no longer about getting 5% more budget, staffing up to scale operations, and waiting until the reports show an increase/decrease in return. It’s about placing more accurate bets on where to spend money.

As businesses continue to emphasize the importance of CRMs in their operations, we see all kinds of measurement methodologies. Marketers run ads and tag people who interact with the ad, determining whether or not people converted through the advertisement. For inbound marketing and content, marketers are appending tags on people who download an ebook; determining if they convert into a customer or not.

By reading the above it seems pretty straight-forward, yet many marketing teams struggle to implement this simple, entry-level measurement methodology. Unfortunately, many businesses spend a lot of time running circles within their CRM; never really maximizing the value of their CRM.

But the above scenario is only the beginning of what we can begin to measure within CRMs — the real gold is in the correlations between activities.

It’s Time to Evolve: Understanding Attribution & Correlations

Think of attribution as giving someone credit for their work. In baseball, when a player hits a home run, we give that player credit for scoring the team a point.

We want to apply the same logic to our marketing — this is attribution.

Some marketers follow different styles of attribution. The technical terms are “first-touch attribution”, “last-touch attribution” and more. The names don’t matter so much as the meaning: which activities should you give credit for earning a customer?

Under a first-touch attribution model, the marketer gives credit for a conversion to the first activity a person interacted with (e.g. an advertisement on social media). Under a last-touch attribution model, the marketer gives credit for a conversion to the last activity a person interacts with before converting (e.g. the kickass copywriting in your automated email nurture).

These are an okay start, but what many marketers don’t know is that there’s another attribution model that introduces us to the notion of correlations: multi-channel attribution.

Multi-channel attribution is when a marketer gives credit for a conversion to a variety of different activities (e.g. the combination of an advertisement, social media post, ebook download, and automated email nurture). It becomes more complex to measure, but acknowledges a truth in marketing: the customer journey isn’t always a clear cut path.

This is where correlations come into play.

Under a multi-channel attribution model, marketers assume some combination of activities yields new customers. Excellent, but how do we measure the effectiveness of each activity? By comparing conversion correlations.

By tagging each person who interacts with an advertisement, content download, event, campaign, or organizational program (e.g. a software company’s beta program), marketers can collect the data to run reports on the effectiveness of each activity. By viewing only customers and reporting on the activities that customers interacted with, marketers will gain a strong sense of which activities are most significant to their customers’ journey.

For example, look at the following report:

Based on what the above report tells us, there appears to be a higher correlation between the Product Catalog and customer conversion than there is the Talk to a Sales Rep activity. This report also shows a low correlation between the Rewards Program and customer conversion, suggesting that the program is perhaps not promoted well, or that the rewards themselves need to be revamped to entice non-customers to take action. Lastly, we can see that the Advertisement does an okay job at acquiring interactions with customers.

In its collective form, this report tells us that the average person’s customer journey is likely to involve:

  1. The Product Catalog Activity,
  2. The Advertisement Activity

And perhaps this makes sense. Maybe this sample organization promotes their Product Catalog as an opt-in over social media, reaching and acquiring leads that then convert into customers — see how this starts to come together? And don’t forget the Rewards Program. It’s clearly the low-converting activity here, but perhaps it needs to be looked at again; to breathe new life into it.

The Future of Correlations & Conversion

As we progress further into data-driven organizations, computer scientists will explore the function and possibility of artificial intelligence (AI).

Just imagine artificial intelligence as a gigantic “if this, then that” function. The AI will continuously run, scanning the CRM for more data. As it finds patterns, correlations, and other unique events, the AI would be able to detect, predict, and compare certain correlations, sudden rises in demand, shifts in consumer preferences, etc. By layering in functionality for notifications, the AI would be able to trigger certain emails to send to CRM administrators, marketers, and sales representatives instructing action or alerting them of a pattern (e.g. the year-over-year revenue we usually make during summer has decreased by 18% when compared to 3 years ago).

Thanks for Reading, My Name is Colin MacInnis

I have an unhealthy obsession with marketing.

I’ve studied business, marketing, computer science, CRMs, and analytics in order to dissect and understand sales and marketing operations.

As a result, I now teach marketing at Cape Breton University and share my findings with fellow managers and business owners.

For more articles like this one, visit the Marketing Qualified Blog.