Analysis paralysis: choosing your marketing data wisely

Analysis paralysis: choosing your marketing data wisely

karl-heberger-coumn-sigIt’s 2019, which means the future is here! Unfortunately, we don’t have the flying cars we were all promised. Instead, we have all of the information the world has ever known (practically) at arm’s length every minute of every day. I’m not sure it’s a fair trade, but I guess that’s the problem with high expectations.

For marketers, the future presents a double-edged sword. On one hand, we now have more data than we could have ever dreamed of. Rather than making decisions based on anecdotal evidence or intuition, we can make educated decisions based on facts. This is great—we’ll never make a bad decision again! Or so we might think.

The reality is that marketers are now inundated with too much data. At any given time we have access to hundreds of reports in various adtech and martech platforms. Systems like Google Analytics, Facebook Insights and Hubspot make it easy to slice-and-dice data any way we’d like. Depending on how you use these systems, they can either be a portal to enlightenment or the world’s biggest time suck.

Take Google Analytics as an example. This very powerful tool measures how visitors interact with your company’s website. It contains over one hundred different built-in reports. Within each report, data can be segmented, filtered and compared in thousands of ways. Marketers looking for busy work need look no further.

Building reports within Google Analytics or any marketing dashboard is not a good use of time if you don’t know what you’re looking for. It’s easy to find metrics that are increasing or decreasing over time, staying steady or fluctuating with no discernable pattern. These reports can be perceived as good or bad. When perceived as bad, it often results in even more busy work as a team has to respond to emails like this:

“We had our intern run some Google Analytics reports and the results are not good. Why is direct traffic to the About Us page is down 50% year-over-year? And why are visitors using the Safari browser 12% more likely to bounce than visitors who are using Firefox? We need to fix this!”

I wish I was making this stuff up. It’s very easy to dig into the reports and find all sorts of metrics that can be perceived as unfavorable. I’ve worked with a number of people who seemed to really enjoy it. The problem is it’s not a productive use of time.

The goal of data exploration should be to develop insights that can be used to inform action. Not only does a decrease in direct traffic to the “About Us” page not matter for the success of a business, there also isn’t much you can do about it.

I used this silly example about traffic to a specific page, but realistically there can be an infinite number of conclusions about the data available in any of our marketing analytics dashboards. It’s important to establish which specific metrics are the most important to your organization’s success. We typically refer to these metrics as key performance indicators or KPIs. When analyzing performance, marketers should focus on pre-defined KPIs to uncover opportunities for improvements.

For example, a SaaS company might offer free software demonstrations to prospective customers. For them, a request for a free demo on their website is a KPI. When they analyze the performance of their website, they start by measuring the conversion rate by channel. This report uncovers the fact that site visitors who come from LinkedIn are the most likely to convert. Visitors who come to the site from a banner advertisement are the least likely to convert. This information easily informs an action to be taken—shift dollars from banner advertising to LinkedIn promotions.

The best way to learn from the data made available in a marketing dashboard is to use the scientific method. Data in its raw form does not provide insight. Marketers should conduct experiments that test various hypotheses. Over time, the knowledge gathered from these experiments can be used to inform more effective marketing campaigns.

A simple example would be to test two different offers on a campaign landing page. That SaaS company has had success with free demos, but what if they offered a free 60-day trial? By splitting the advertising to drive visitors to two different landing pages with just this single variable, they can learn which offer is more likely to convert visitors into leads.

Marketing dashboards are great tools every marketer has at his or her disposal. Unfortunately, they provide an unlimited number of meaningless reports that create analysis paralysis and waste a lot of people’s time. When used correctly, however, they can provide wisdom that informs decisions. I hope to see more that in the future.

Karl Heberger is chief strategy officer at Mason Digital, a full-service digital marketing firm. He can be reached at [email protected].