You know you need web analytics to track your traffic.
What’s important, and a lot tougher, is answering this question: what insights into the behaviors and psychology of your users do your analytics provide?
This isn’t always obvious.
The truth is that there are plenty of mistakes to avoid in setting up, monitoring and evaluating your web traffic.
I keep this list on-hand to ensure that when I’m working with analytics I don’t fall into these common traps.
1. “Low bounce rate = win”
The bounce rate is one of those easy-to-swallow metrics that’s always “right there” in your analytics reports. It reflects what percentage of visitors left your site after viewing only one page.
The general rule of thumb is that a low bounce rate is a good thing – it means people are viewing more than one page and “engaging” your content.
But a bounce rate on its own tells you very little.
When you consider the full context of this metric it can often turn out negative.
Example: if you’re running a blog with a decent amount of subcribers, many people will click through to your post from their RSS feed reader. These visitors frequently are there for one purpose: to read your latest post and go about their day. That means it’s unlikely they’ll view more than one page (causing a high bounce rate).
2. “Low page views = fail”
Similar to bounce rate people often point to the number of page views per visit as a metric to measure your users’ level of engagement. However, this metric also requires a big grain of salt.
It depends again on the traffic source and user segment, but consider the purpose of your landing page. It could be a signal of a negative user experience.
Example: if your landing page highlights an offer you’ve been promoting off-site (on the web or with traditional advertising) and the checkout process requires two steps, a high page view number might be negative. One possibility: your users are confused – something isn’t clicking with their expectations, and they’re looking around.
3. Forget about pairing metrics
The tendency of #1 and #2 above boils down to viewing your traffic through the lens of a single metric. While it’s easy/quick to view and digest this information it rarely tells you enough about your traffic.
Pairing two metrics is a great way to take a deeper look at your data and segment your users.
Example: comparing time on site with the bounce rate for various referral sources affords you better insight into what your bounce rates indicate. A user who bounces after viewing a page for 2+ minutes isn’t necessarily coming away dissatisfied.
4. Don’t segment users
It’s easy to lump your visitors into the “our users” category and forget that user segments have different traits and goals in mind.
Segmenting your users is crucial for understanding the more nuanced picture of whether they’re reaching points of satisfaction.
Example: for UnstuckDigital.com two potential user segments are 1) blog readers and 2) potential clients. While there is always the possibility for overlap we tend to see blog readers arriving directly or through referral links whereas potential clients often arrive through search engines. We view these segments separately in our analytics.
5. Ignore the buying cycle
Most analytics platforms default to displaying data for the last 30 days. This is fine for many people, but you’ve got to keep in mind the buying/engagement cycle of your website and product/service.
Some companies sell products that represent a significant investment. For them it is rare to see a customer arrive at the website for the first time and buy within a few weeks. Instead customers normally research extensively before purchasing. The result is a buying cycle that regularly stretches longer than 30 days.
Stick to a measurement of time that makes sense for your sales cycle. This way when you launch an online marketing campaign you can see the conversions kick in down the line.
Example: an IT vendor sees a normal sales cycle of 60-90 days. They launch a new email marketing and pay-per-click campaign simultaneously. The 30-day analytics report shows their traffic spike, but what they don’t see is the conversions another 30-60 days later. This may seem obvious, but it’s easy to overlook – especially if the communication between sales and marketing is broken.
6. Bow down to the machine
When you’re faced with evaluating the success of your online efforts web analytics becomes your central focus. In many ways it’s the best information available for evaluating your online campaigns.
On the other hand it’s important to remember: web analytics technology isn’t perfect. There are “holes” in the way your traffic is tracked – and limitations as to what you can do about it.
Example: Bob visits your website through a pay-per-click ad. Bob wasn’t ready to buy so he left your site, making a mental note. Two weeks later, Bob returns – but he does so by searching in Google for your brand name. Then he buys your product. Most analytics packages (Google included) will credit Bob’s second referrer (brand name search) with the sale instead of your PPC ad.
7. Don’t filter internal visits
It’s easy to underestimate the impact internal traffic has on your website – sure, you and your coworkers check the company website a few times a week, but that’s not going to make a noticable difference is it?
Actually, internal traffic can seriously convolute analytics data. It’s important to ensure you’re filtering internal traffic so you don’t poison the well.
Example: We had a client a few years back whose analytics data were showing high time per visit numbers. When we looked into the issue we found out that everyone in the office had the company website set as their browser home page. Some people would pull up their browser and head out to lunch. Others left it open all day. The result was skewed and unreliable data.
Update: Internet Strategist (see below comment) suggested I add information on how to filter internal traffic. ROI Revolution actually has a great post on how to do this easily here.
{ 2 comments… read them below or add one }
Great post. Any chance we can talk you into a follow-up on do-it-yourself excluding your internal traffic from analytics data? This is something every person using Google Analytics needs done including those who don’t have the resources to pay to have their GA configured.
You might not have to write it yourself as I suspect you could find a great post elsewhere about this, maybe at ROI Revolution or some other analytics blog.
@InternetStrategist Thanks!
I took your advice and found a relevant post over at the ROI Revolution blog (a great analytics blog by the way). I’ve linked to it in the main post above.