Contributed by guest expert Andrew McLoughlin for Colibri Digital Marketing
One thing we’ve learned as a top San Francisco digital marketing agency is that our efforts are almost totally meaningless if we don’t have a way to measure our impact. Like most businesses, ours depends on feedback, tracking results, testing new approaches and strategies, and being able to compare the results objectively.
For digital marketers, Google Analytics is one of the most important tools for tracking results and making sure your website is optimized for search engines. We’ve put together this crash-course below to give you the framework you need to understand the software and to begin to work it into your own digital toolkit. A small disclaimer, this isn’t meant to replace Google’s own analytics course and certification, but rather to supplement it.
Google Analytics and the Importance of Numbers
Imagine a scenario: It’s mid-spring. Your site has just published a new landing page which uses a slightly different phrasing from your usual branding. You’re linking to the landing page from your monthly newsletter and publishing links to it on your social media accounts. Three days later, your site has a measured increase in conversions (say, sales from your online store). It seems like the conversions correlate with the new landing page, with the altered branding, so you re-do your whole site to match the new tone and imagery. Suddenly, however, conversions drop, and revenue hits an all-time low. What went wrong?
The mistake in this example was matching up the correlation of the two events — the new landing page and the increase in conversions — and assuming, without evidence, that the new branding directly caused the spike in conversions. If you’d been tracking your traffic with Google Analytics, you’d have seen that the bounce rate for the new landing page was unusually high, and the time on page was unusually low. People hated the new branding, even people who liked your site enough to sign up for the newsletter. Your conversions spiked because they always spike around this time of year: people are buying gifts for Mother’s Day.
In this scenario, data analytics would have saved your site from the unfortunate self-sabotage of acting rashly without evidence. Having hard numbers, tracing their origins, and segmenting the data by a number of different variables is the only clear path to success.
How Not To Use Google Analytics
-Channels vs. techniques
Though our scenario above was telling, don’t mistake Google Analytics for the be-all, end-all marketing tool that some people take it for. Indeed, there’s one cardinal mistake that many inexperienced digital marketers make when using Google Analytics for the first time. Let us explain.
Basically, Google Analytics breaks data down primarily by channel. There are other ways to configure your reports, but the default, and the one most people start with, breaks traffic down into groups like “Organic” (people who found your site in the search results by querying relevant keywords), “Direct” (people who clicked a link, loaded a bookmark, or typed a URL), “Social” (people who came from social media referrals), “Paid Search” (people who clicked on a paid ad), and so on.
This default grouping of data makes it dangerously easy to think of each discrete channel as a discrete marketing strategy. That’s a serious pitfall, and it catches more and more digital marketers in its snare.
Google Analytics does a fine job of telling you about your site traffic, and user behaviors, and so on, but don’t treat it like a catch-all for your broader marketing efforts. The best advice we can give you is to use the tool the way it’s intended. Don’t try to hammer a nail with a screwdriver, and don’t expect Google Analytics to do the heavy lifting for your digital marketing strategy.
What It Can Do
So, with that out of the way, let’s talk about Google Analytics’s various strengths and applications. It can tell you where your visitors are coming from, by grouping your traffic into discrete sources. It can tell you what sorts of paths users follow once they get to your site, and the data can be segmented by a number of variables to get a more comprehensive picture. It can track conversions, goals, and user interactions, and, under the right circumstances, it can be used to see whether new marketing initiatives are having the impact you were hoping for.
Let’s envision another scenario. You want to find out more about your users’ demographics, specifically to help you decide whether it would be worth your time to invest in a dedicated mobile version of your site, or to redirect your marketing efforts. So, you chart the data, breaking down your users by location, device type (screen resolution can tell you whether it was a tablet, cell phone, or desktop), browser, language, and so on.
By analyzing the results, you discover that a majority of your visitors are primarily German-speaking tablet users, most of whom are based in Europe. It also seems that the majority of your users are repeat visitors, usually following a direct link.
So what do you do?
In this case, you’d be wise to invest in a German language version of your site, optimized for mobile. You’d also probably see considerable success from a newsletter, since such a large percentage of your users have demonstrated a recurring or consistent interest in your content.
By letting the data tell you more about what’s really going on, you will have learned more about your users, and you’ll be better able to meet their needs.
The thing with data, even comprehensive data, is that it isn’t useful to you without context. You can’t extrapolate from even a few points of data. This is where benchmarking comes in.
Basically, benchmarking is the process by which you track data over time, relative to circumstance, to get a feel for what a “typical” spread would look like, so that you have a way to evaluate the data for a particular campaign against an average. Whenever possible, we track data for at least three months, minimum, when coming up with our benchmarks, just to account for outliers (like mother’s day shopping!) Ideally, you’d want to track data against benchmarks from the same month of the year, the same day of the week, and even the same time of day. For instance, if your traffic typically spikes around 5am PST (that’s 2pm in Berlin) you’d want to track your results from the same time period when comparing the results of a new campaign or landing page. Most business sites generally spike mid-day on weekdays, while most leisure sites spike on evenings and weekends.
If your site has been around a while, it’s always good to track the year-over-year benchmarks as well. If your site spikes around Christmas, then don’t despair if your traffic for June doesn’t match your all-time-highs. The more you contextualize your data, the more useful it will be, and the more precise and correct the inferences you generate.
Tips and Strategies
-Don’t touch the raw data
From the moment you implement your analytics tracking codes, keep one view that is totally unfiltered and unmodified. If you don’t have one, make one now. Once you filter your data, it’s filtered retroactively. Google doesn’t save previous versions, so your alterations are permanent. If you accidentally implement a filter incorrectly, or erase certain data, it’s gone for good, so keep one backup of your whole account. You can always re-build filters, but you can’t restore data without an archive.
-Try everything in a test view first
Like the previous tip, when adding filters or creating a custom view, always make the change in a sandbox first. Use a test view to play around and to make sure that your filters are doing what you want them to. That way, you’ll have a chance to try again and perfect it before you make the change permanent.
-Don’t make a million filters when one will do
Say you’re filtering bot spam (and you should be!). You don’t want fake user data cluttering up your results, and there are plenty of hostnames that you’ll want to exclude. Remember, though, that you can write a simple regular expression that you can add to indefinitely for each new hostname you want to filter out. By making just a single filter, called “bot spam” or something, and using it properly, you can save yourself the hassle of trying to comb through hundreds of individual filters whenever you need to add or modify a particular thing. It’s best practice, and it will save you from a massive headache in the long run.
-Document everything by date
Google Analytics lets you keep a log of changes, revisions, new pages, and more, and you should be recording every change. I can’t stress that enough. Imagine noticing an unexplained spike in traffic that started two or three months ago, but is beginning to taper off. What change did you make to your site three months ago that might have triggered it? Were you publishing a newsletter, or covering a particular topic in your blog? Did you make a change to your site architecture, or re-structure your menus? Did you add a tracking code to a page that had always been there, that is now underperforming for one reason or another?
Without a log, and without context, your data, however tantalizing, won’t be useful to you.
That’s it for our quick Google Analytics crash course for digital marketers. We hope it’s given you more of a foundation in how to use this extremely versatile tool. If you’re already versed in Analytics, we hope that this post will inspire you to continue to develop good habits, and to make the most of your data, in your ongoing digital marketing efforts. If you’ve made it to the end of this post, then, on behalf of Colibri Digital Marketing, thank you for reading. You rock. We welcome your feedback, so give us a shout on Facebook or Twitter, and let us know what you think! And a sincere thanks to our kind hosts, for running this post.