Marketing Analytics: 6 simple steps for interpreting your data
You’ve finally set up tracking on your site and have gathered weeks of information. You are now staring at your data saying, “Now what?”
Objectively interpreting your data can be extremely overwhelming and very difficult to do correctly … but it is essential.
The only thing worse than having no insights is having incorrect insights. The latter can be extremely costly to your business.
Use these six simple steps to help you effectively and correctly interpret your data.
Step #1: Be strategic
Don’t try to analyze it all; you’ll get lost in data and become discouraged and confused. Instead, narrow your focus to the metrics that will provide the most relevant insights.
Having primary and secondary KPIs for your site will help you begin to narrow your focus. Common KPIs include conversion rate, revenue, CPA (Cost per acquisition), number of leads, and clickthrough rate.
You will also want to study a few other secondary metrics that could affect your KPI. I’ll talk more about these secondary metrics in a moment.
Step #2: Understand your business and how the data relates
In Step #1, I recommended that you narrow your focus to just the relevant metrics. Your type of business, the industry you’re in, your target audience and your revenue model will all affect the relevance of different metrics.
Unfortunately, there are no cookie-cutter rules for choosing the right data on which to focus. While you will need to do research and have discussions within your company, here are some common KPIs used by different business types (as reported in the MarketingSherpa 2011 Landing Page Optimization Benchmark Report):
E-commerce: conversion rate, total revenue, average order value, orders completed, cart abandonment rate, and drop-off rate within the checkout process
B2B: visits, page views, and leads generated
B2C: conversion rate, clickthrough rate, and orders completed
Secondary metrics you should incorporate into your analysis will not only depend on your business, but also on what page you are examining. For example, you may want to look at next page path on a homepage (to determine what type of information and in what sequence users want), browser and device type on a checkout page (to expose any technical problems certain devices/browsers may be experiencing), or traffic sources on a landing page.
Step #3: Be objective
Once you determine what metrics you will analyze, analyze them objectively. Do not fall victim to the confirmation bias — a type of selective thinking that can lead to statistical errors when one favors or seeks information that confirms their hypothesis, ignoring information that refutes it. We can often get caught up in an “epiphany” and then go search for data that confirms our thought, disregarding data that may prove otherwise. This is a fatal mistake that could be very costly to your business.
Step #4: Look at the data from all angles
The easy part of data analysis is the question, “What is happening?” The more difficult, yet exceedingly more valuable, question is, “Why is it happening?”
Rarely is there one single way to interpret data. It is important that we look at the data from many angles to avoid tunnel vision that will lead us to make incorrect assumptions.
Let’s look at a common example — a high bounce rate. The “what” is that people are landing on your page and not clicking to a second page.
But the more important question is, “Why?”
Often times, people think a high bounce rate is because “people just don’t like my page.” Maybe. But there are other possible reasons:
- It could also be that your page load time is too high and Internet users are inherently impatient.
- It could be that you have great content, but you are lacking continuity between your pay-per-click (PPC) ad campaign and your landing page, so the wrong people are being sent to your page.
- Or it could be because it’s a landing page with no other links besides your call-to-action (CTA), so everyone who doesn’t click your CTA is counted as a bounce.
As you can see, different metrics can be analyzed in many different ways. So how do we know which “why” is true? Looking at multiple metrics together will help clarify.
Step #5: Analyze multiple metrics together
Looking at a single metric can usually provide a high-level insight. But to truly begin understanding how users interact with your page, and what makes them convert, you need to look at how different metrics relate to one another.
Take this metric for example:
An average visit duration of almost 40 minutes is very high. The source of this data must remain anonymous, but it is a B2B company selling a high-involvement purchase. The site has many pages with a lot of content, including videos. So at first glance, this could be interpreted as favorable — people are reading the content and interacting.
But once we look at this metric in tandem with the number of pages per visit, we see something different.
Our first hypothesis was that people were interacting with the site and doing their research; however, on average, they are only visiting 1.5 pages.
This should throw up a red flag in your mind; the likelihood of someone spending 37 minutes on 1-2 pages is low.
Then, when we add a third piece of information — most people land on the homepage — the likelihood is even further reduced, as there is very little content on this particular homepage.
When we looked into this further, we found that one of the links on the homepage opened a new tab. The homepage was still open in people’s browsers, but they had already moved on to a different page in a separate tab.
By looking at all of these metrics together, we were able to better understand the data, interpret the “why,” and avoid costly assumptions.
Another type of analysis to consider performing (something we do regularly at MECLABS) is looking at trends between different metrics — a correlation analysis.
For example, you may want to look at time on page in correlation with conversion rate. Does conversion rate go up as time on page goes up? Or does average order value increase as pages per visit increases? Does bounce rate increase as page load time increases?
Answering these questions can help you understand the thought process of your consumer and find out “what works” on your own site.
Step #6: Don’t treat everyone the same
It is always important to see how your site performs as a whole, but it is also important to take a more granular look to understand what (or who) performs best.
By segmenting users (e.g., creating an advanced segment in Google Analytics), you can learn how certain groups of visitors interact differently with your site. It will help you leverage what’s working for the high-performing groups as well as expose low-performing groups that you could spend less time and money on.
Google Analytics provides default segments and allows you to create custom segments, as well:
For example, you could create a segment based on the traffic channel they came from (organic search, PPC, referral, etc.), the specific PPC campaign they came from, a geographic segment, or what they did on your website (a specific link they clicked or the number of pages they visited).
Finally, and probably most importantly, you should look at the segment of people who completed your KPI. What was different about that group compared to those who did not complete the KPI? Did they visit more pages? Did a higher percentage of them come from a specific channel? This will help you understand what content people seek and the actions they take that increase their chance of converting.
In the end, data (big or small) is just numbers; it’s your job to give them meaning. I hope these six tips help. At MECLABS, we’re always learning, so we’d also like to hear from you. Share your insights in the 2013 Marketing Analytics Benchmark Survey and receive a free special report, Evaluating Website Optimization (a $97 value).