You Don't Know Your Visitors, So Stop Pretending
February 10, 2013
Part 1: What is analytics really?
Web analytics should hurt a little. Not just the pain of seeing your low traffic revealed in hard numbers, but the realization that you don’t really know your visitors.
We programmers tend to be preoccupied by our creations. We repetitively visit our sites, refining and building. Reading our own copy again and again becomes almost hypnotic. The sentences start to seem so right, so laden with all the meanings of our intentions and our history with the project.
Then we unveil our website to the public and shit gets real.
High bounce rates abound. Visitors mysteriously fail to get very far into the site. Visitors talk trash about the experience (if they talk about it at all). What can you do? Your secret weapon is properly designed analytics.
These type of data are known as “vanity metrics,” and they don’t really do you any good, except sometimes to boost your vanity. They do not help you improve any more than would a doctor who merely observes, “Looks like you’re sick. Best of luck!”
Real improvement comes from asking questions. The amount of technology involved varies. At its most primitive (and often most effective), learning about your website involves sitting someone down, giving them the barest outline of what your site is about, and watching them use it. Make sure you have something to write on while the user points out your ambiguous language, obscure navigation, and unintended connotations.
After doing this basic in-person testing and after ironing out the major flaws you are ready to enable more automated analytics. My favorite tool is Google Analytics, but it takes some work to get the most out of it.
Begin by identifying your goals for your site. Maybe you’re selling things and your goal is to maximize sales. Maybe you’re informing people, and your goal is for them to learn specific things. Make your goals specific and explicit. Let’s use an example.
Suppose you’re a technology consulting company. You’re looking for clients. But you’re looking for the right clients, those whose projects align with your skills and inclination. Let’s work backwards. Your primary goal is that prospective clients contact you. You can measure this, and I’ll explain some techniques later. For now notice there are other goals as well.
For instance you want to communicate your company’s specialty and strengths. Maybe you have a unique process of product design, maybe you’re exceptionally fast. In any case, you would like prospective clients to know basic things about you so that you’re not only likely to get contacted, but get contacted by the right people.
The next step is to make the goal specific in terms of visitor actions on your site.
- Goal: Visitors contact you.
- Telltale: Either they fill out your contact form and press the submit button, or they follow a “mailto:” link, or they select the email address on the page to copy it.
- Goal: Visitors know your firm’s process and specialty.
- Telltale: They visit a specific page about it and stay on that page for a given amount of time, or they watch a video about you at least most of the way through.
- Goal: Informed visitors contact you.
- Telltale: The visitor has fulfilled both of the previous goals during their visit.
Notice two things about these goals. One, their telltales don’t cover all cases, but they are useful nonetheless. For instance, a visitor may arrive on your site already knowing about you. They go straight to the contact form, so according to “the rules” they are counted as a contact, but not an informed contact. However, the telltales aren’t misleading in the other direction. If someone spends a long time reading about your process and past clients and watches your videos, then they are certainly informed.
The second thing to notice is that goals can be built out of other goals. Sequences of desired actions (like the informed contact) are known in web analytics as “funnels.” By identifying a funnel such as “the visitor learns about us and then contacts us,” you allow analytics software to look for the weak links. That is, the steps in the process where people most frequently quit. Our example funnel is very short, but think of a funnel for buying things online. It goes all the way from editing your cart to entering billing info, to …, to placing the order.
Once you identify goals and translate them into telltale actions on your page you are ready to write code to report the appropriate visitor actions to your analytics server. At least on Google Analytics, many actions are so basic that you don’t have to write any code to track them. Google will track them by default. However other more customized telltale actions require customized code.
Before discussing the specifics of Google Analytics code, let’s consider: what good do these goals do you?
First, starting with goals forces you to make the purpose of your site explicit. Ever visit brochure websites that are incoherent and smothered in inane social media icons? Why do people (usually organizations) make these? It’s because they don’t know what they want. In the absence of explicit prioritized goals a committee of designers will create a well-intentioned slurry of failure.
Second, measurable goals let you use science. Defining the telltale visitor actions that constitute a goal allows you to write code to reliably measure success. Data allows you to do experiments. You can finally answer questions definitively. “Does the front page carousel increase visitors’ time spent reading our blog articles?” “Does the contact link in the footer increase contacts?” When your design committee starts blathering, you just point to the charts.
Third, you can automate your experiments. Google Analytics offers a feature called “content experiments” that let you specify alternate versions of pages in your site. Google will choose which versions visitors see and report how various configurations affect your goals.