Metrics That Matter: A New Framework to Reveal Truly Actionable Insights

• Tweet
• Buffer

Marketers struggling with the explosion in data might take a lesson from Albert Einstein when he famously said, "Not everything that counts can be counted, and not everything that can be counted counts." My meandering thoughts for this column focus on the latter part of this statement, since I believe we've become much too metric happy: measuring too much and not necessarily the right stuff.

Why, you ask? Because the volume and variety of the data, tracking every step in the customer journey, is so daunting that marketers often don't know what questions to ask or which metrics to use. We know we need to make sense of the growing mass of data, yet too often there's not enough method to our madness. Meanwhile, buyers travel their own winding paths to purchase across ever-changing e-commerce websites and social platforms, smartphones and tablets, and in-store visits - leaving even more digital breadcrumbs behind.

What's measurable? Almost everything.

What's meaningful? That's another question altogether.

So let's look at the basics of what needs to be measured. As marketers, we want to know how our marketing activities perform across all channels - online, social, mobile, and offline. The key issue here is actionability. Simply put, actionable metrics should tell the marketer how to act based on the insights revealed by the data. Good key performance indicators (KPIs) help us understand why, when, and where customers buy and give us the insights to improve performance. Ineffective metrics do not definitively show us anything and may even be misleading. I'm not just talking about the obvious, one-dimensional counts of visitors, followers, subscribers, tweets, clicks, page views, email opens, posts, returning visitors, and the like. Even tried-and-true reach, engagement, and conversion KPIs may not give marketers what is needed to truly understand how multi-channel marketing efforts are doing - and what's needed to improve performance.

That's because marketing channels have become increasingly interrelated and interdependent. For instance, an email campaign can go viral and impact how your paid campaigns are performing. Negative sentiment from a competitor can lead to increased traffic and engagement on your web properties. Your social and mobile platforms add to the complexity as they contribute to changes in consumer behavior. Bounce rates, for example, from social sites tend to be highest across all channels as viewers consume content differently than they do on websites. A visitor may come to your site through Facebook and then jump right back into what she was doing, resulting in a high bounce rate for the social channel. Yet, that visitor may have gotten exactly what she needed during that brief interaction with your brand.

A New Framework

My experience in marketing and analytics makes me believe that we need a new framework - a different way of looking at the metrics that matter - in the midst of all this change. We need to differentiate what needs to be measured based on the insights these metrics offer. So here's my guide to which metrics to focus on and which ones to ignore, charting an evolutionary analytics process that will progressively deliver more impact on that elusive marketing ROI:

• Descriptive metrics. These metrics reveal what happened based on past actions and provide you with a "current-state-of-the-system" view. Good examples of these metrics include: site conversion rate, campaign conversion rate, cost per conversion (CPC), customer retention rate, customer referral rate, and average number of service calls per day. Descriptive metrics show up on marketing dashboards across enterprises, but unfortunately don't deliver understanding or actionability on their own. The dashboards with these metrics - even when they are unified, which they usually aren't - don't convey any system interdependency and correlation among the top-line metrics. Unfortunately, without a holistic understanding of the relationships among these various KPIs, descriptive metrics are of little use. Worse, they can mislead, becoming the wrong indicators of actual performance.
• Diagnostic metrics. What's needed is a method of going beyond the "current state" to show the actual levers of change. Marketers need to understand not just the "what," but also the "why" and the "when." To do this right you need access to real-time, granular data that can be sliced, diced, and segmented to discover hidden relationships, which otherwise easily escape detection. Diagnostic metrics are almost always rooted in multi-channel data, and are often overlaid with other KPIs to provide much-needed comparative perspectives. Examples of these metrics include: visits to purchase by customer acquisition sources (indicating best traffic sources in terms of likelihood of conversion), cost per referrer visits (best traffic in terms of cost per visit), social conversion rate vs. overall conversion rate, or even campaign CPA vs. social mentions and social sentiment. These metrics all help the marketing team better understand the complex relationships between seemingly disparate marketing channels. They make it possible to identify levers of change in one channel that can directly or indirectly impact another channel. Increasing a specific social marketing effort, for example, may lead to reduction in CPA for a specific paid media campaign.
• Predictive metrics. After identifying the levers of change, the following question becomes quite important: what happens if you move any of the levers? There is no need to guess or trust one's intuition. Predictive analytics can reveal the most likely outcomes based upon movement of the levers; this is accomplished by applying statistical models to all those granular data points to create predictions and forecasts. Using these models, the marketing team can forecast paid- and social-campaign performance, identify predictors-to-purchase by customer segments, improve real-time targeting, and predict how a wide variety of other factors will change. The models can help marketers identify the characteristics of the interactions between channels, campaigns, and performance and show how moving certain levers can result in trade-offs. Continuing the example above, while CPA might be reduced for one campaign, it could be increased for another. With this knowledge marketers could make decisions based on the dynamic interrelations of the entire system as opposed to being limited by a less holistic view.

Unlike descriptive metrics, which are not all that actionable, diagnostic and predictive metrics deliver powerful decision support and deep understanding of the levers of change. When combined effectively, they provide marketers the directional guidance to prescribe a better marketing mix for higher revenue and profitability. These metrics cut through the clutter of data and empower marketers to act now based on the right-time insights to create the desired future for your business.

Can there really be a better definition for metrics that matter?

Image on home page via Shutterstock.