Evaluate your internet marketing strategy. What to measure and why.
This article is heard toward small to medium sized businesses in moderately competitive markets. However, the fundamentals of evaluating online marketing efforts are largely the same for any business. It is important to recognize that an effective Internet marketing plan must take the competitive landscape into account. The following questions should help get you thinking about what you need to know as your online marketing efforts evolve.
- How sophisticated are your competitors with regard to Internet marketing?
- How aggressive are they in their Internet marketing efforts?
- Do they hire agencies, or manage Internet marketing in-house?
- Does your business serve a local market? Regional? National? International?
- How many primary competitors are there?
- What’s your position in the market? Leader? Underdog?
- How much time do you have to devote to Internet marketing?
The first two questions are perhaps the most important. Marketing sophistication and the level of effort devoted to effective Internet marketing will level the playing field in spite of virtually everything else.
Internet Marketing: Primary Metrics
I’ll begin with what I call the primary metrics. These are the data shown on the first page when you log into Google Analytics. Here’s what they are and what mean.
- Sessions: This is the total number of visits to your website for the time period displayed. It includes repeat visits, so a visitor returning to your website will count as multiple sessions.
- Users: This is the number of actual visitors that contributed to the session total.
- Pageviews: This is the total number of browser page requests that resulted in a page being displayed.
- Pages / Session: The average number of pages served during each session. If you multiply this number by the number of session, it should match “Pageviews.”
- Avg. Session Duration: Average time, in minutes and seconds, of all sessions.
- Bounce Rate: A bounce is a simply a single page session. Not all bounces are bad. More on this later.
- % New Sessions: The percentage of sessions by visitors that hadn’t previously been to your website. Google Analytics also displays a pie chart representing this ratio.
All of this information is historical in nature, and only provides totals and averages. It is helpful for identifying trends, but to understand what’s happening on your website, you need to dig deeper.
Validate your website visitor data
From the beginning you need to know the information you rely on is valid. The vast majority of websites I work with for the first time take everything fed to Google Analytics as legitimate. The truth is it’s probably not even close. Here are a few things that screw up your data.
Spambots: These are computer programs, often called robots, spiders, crawlers or some other name. Not all robots are bad, but spambots scour the web looking for any online form to fill out with bogus information. Their goal is to create links to crappy websites by them into the comments section of your blog or other publicly visible content.
They do this in a nefarious attempt build the number of inbound links (backlinks) to some website. Entire companies, mostly in Eastern Europe and China, are devoted to selling backlinks this way. It’s obnoxious and doesn’t work, but comment spam is everywhere.
Tip: You should always moderate blog comments before allowing them to be published on your website. At the very least moderate those that include links.
What sucks is each attempt is reflected as as a visit. Usually, visits from spambots will have a bounce rate of 100% and a session duration of 00:00. They badly skew your data, and will often overwhelm legitimate traffic. As an example, a typical website with 1000 legitimate visits might have a bounce rate of 35%.
Add 500 visits from scumbag spambots and your traffic will be reflected as 1500 sessions with a bounce rate of 56.6%. Without the phony data from the spambots, the top level data looks pretty good. But, if you believe the numbers that include the spambot data, you’re likely to try to correct a problem that doesn’t exist.
Referral Spam: These are jerks who bombard your website from a URL they want you to visit. The idea is that if you see a sharp spike in traffic, you’ll want to investigate the source. Once you know where the traffic is coming from, it is very natural to want to check it out.
What’s actually happening is your are being tricked into visiting websites you wouldn’t know about otherwise. Beyond that, these referral spammers screw up your data much like the spambots do. Your traffic will be artificially high, and your bounce rate, time on site and pages per visit will all be wrong.
I’ll touch on filtering Google Analytics in a minute, so here’s a short list of URLs I filter out always.
There are more, but these eight routinely account for 60% to 70% of website visits. Many times I’ve seen referral spam account for virtually all of a website’s traffic.
My clients often think they have high bounce rates and low session durations, but plenty of traffic. So, they pull their hair out tweaking pages and moving content, but it makes no difference. Here’s a real-world example of what’s happening:
Unfiltered Website Traffic Data:
Avg. Session Duration: 00:00:06
Bounce Rate: 88.14%
Filtered Website Traffic Data:
Avg. Session Duration: 00:01:54
Bounce Rate: 39.38%
In this example, the client came to me to solve a bounce rate / session duration problem, when in fact, he suffered from weak traffic volume. It was the exact opposite of what he believed. It’s maddening and disheartening, but once we were able to deal with real numbers, we could work on the right things.
Hacking Attempts: Depending on the type of attack, hacking attempts will also skew your data. It should be obvious by now that bogus visits tend to push sessions up and session durations down.
Brute force attacks sometimes result in legitimate traffic not even getting through. What’s happening is the hacker has a network of corrupted computers trying to guess passwords or take advantage of a known exploit.
In some cases, they’ll just hammer your website with thousands of visits a second so no legitimate traffic can get through. I won’t bother with the details right now, but recognize that hacking attempts screw up your data.
It should be obvious, but you can’t ignore hacking attempts. Data and traffic issues aside, given time, a hacker will get through. That never ends well, so if you suspect you are being attacked, call or email us and we’ll help you deal with it.
Internal Traffic: A website that is being actively worked on will skew your number in the opposite direction. Web designers and content curators often spend hours on pages and hop through the website testing user interface changes.
Early on, that may be your only traffic. It might warm your heart to think people are visiting every page on your website an hour and a half at a time, but it’s not really real. Again, filtering is the answer. In this case, you would filter the IP addresses of anyone that actively works on your website.
Also, if your website is WordPress, you can use a plugin to disable Google Analytics tracking code for certain types of users when they’re logged in. If they aren’t logged in, it’ll still mess with your data though, so it is best to filter by IP.
Job Searches: Obviously, legitimate visitors to your careers page are desirable. I mention them here because job searchers are most likely not buying anything. It really depends on your business and the nature of the visit.
It may be that the website visitor arrived as a genuine potential customer. Then, decided to explore career opportunities. In this case, the visit certainly has conversion potential. In our agency, that’s unlikely, so unless we want to assess our recruiting efforts, we filter those visits into another view.
Next, I’ll provide details about keyword tracking and tools you can use to evaluate whether or not you are executing an effective Internet marketing strategy. Drop me an email, and I’ll be sure to let you know when “Effective Internet Marketing: Part II” is published.