A very easy trap to fall through when performing website analysis is to compare metrics from different sources of traffic against one another. No one in their right mind would ever try to compare the Nielsen ratings of their prime-time TV ad during American Idol against the number of phone calls from the back-page black-and-white ad in their local town's post-gazette. We don't compare different sources of audience in the offline world, and we shouldn't do it in the online world either. But, it happens all the time – a lot of folks like to log-in to their favorite web analytics solution and compare the number of leads between their organic and their paid initiatives, or the pages per visit between their direct traffic and their referrals, leaving them with more questions than they had before they started!
On top of that, there are hundreds upon hundreds of metrics, reports, data sets, graphs, tables, queries and statistics to sort through; making what should be a simple task completely overwhelming. The goal of insightful analysis should be to obtain easy and actionable information, not to make your life miserable by being flooded with rows and rows of endless data.
So, stop comparing apples to oranges to pineapples to plasma TVs, and use the following guide to learn about the different sources of traffic your website receives, and what key metrics to pay attention to along the way.
What it is — Direct traffic is traffic that comes to your site "directly", without the assistance of an organic search result, a paid ad of some sort, or a link from another website. Normally, visitors are considered to be "direct" when they type in your website's URL directly into their browser's address bar. In some cases, direct traffic can also mean traffic from bookmarks and browser default homepages.
Measure: Percent New Visitors — Direct traffic is traffic that is already familiar enough with your brand and your website’s URL. They type it in directly, or access your site from a bookmark because they have been to your site before (and liked what they saw), or they heard or read about you in a TV ad or a print magazine. Measuring the percentage of new visitors that access your website directly can give you a very good idea of how strong your brand is becoming, and how effective your offline marketing initiatives are in driving new eyeballs to your website. Use bounce rate – the percentage of single-page visits to your site – to determine if New Visitors not only show brand-name recognition, but also show interest in your website.
What it is — Referral traffic is any traffic that reaches your website via a link on someone else's website. For example, if you have a directory listing on dmoz.org, and someone clicks on that link to reach your website, you will see a visit from dmoz.org in your referring sites report of your analytics package of choice.
What to measure: Opportunity — Referring sites' reports are often a mixed bag of traffic goodies for your website, and it's within this mixed bag of delicious referring traffic where you can find golden opportunities. Do you receive a lot of traffic from images.google.com? Chances are you have a good SEO implementation strategy in place, making good use of the image ALT attribute. Wikipedia.org a big source of referring traffic? Kudos to the article contributors who have included some links to your website along the way. Social Media sites giving you a few visits? Perhaps it's time to spruce up your Social Media strategy. A friend's website sending you lots of traffic from a link you shared with him last month? Don't forget to send him a Christmas card.
Don't worry about: Conversion Rate — Any leads or sales that you happen to receive from a referring link should be considered as a birthday present. Chances are very likely that a visitor clicked on the link to your website out of pure curiosity. Maybe your link was blue and the other links on the page were purple, and that's why they got to your site. Perhaps your link has a star or starburst next to it, which someone found cute. Whatever the case, be thankful that your link is apparently in a good spot on someone else's site, be it a friend's website or the latest piece of furniture you’re trying to sell on Craigslist, and that it led to a conversion. Don't bog yourself down in trying to figure out how to get more people from other website's to convert on yours.
What it is — All clicks on listings within the "natural" search engine results from Google, Yahoo, Bing.com, and ASK.com (among many others) are counted as "organic". These listings come from the billions of indexed pages that the major search engines have on demand when a user searches for anything online. By default, this includes clicks on sponsored ads, which are found above and to the right of organic search results, but most every web analytics tool allows you to configure your destination URLs to avoid mixing up paid and organic visits.
Measure: Keywords — This is as obvious as it sounds, but it's not enough to just know what keyword a visitor used to find your website – you should also understand the type of keyword that they used, and which one of your website pages a search engine served up to your visitor. Did a visitor find your website by casually browsing the web with a short-tail keyword, like "camera"? Or did they appear ready to buy by searching for the long-tail term "canon powershot sd1000 elph"? It's crucial that you have a solid SEO foundation with well-targeted content-rich pages, and your analytics package can help you determine if visitors are reacting positively or negatively to pages that search engines serve to them as a result of their query.
Don’t worry about: Entrance Pages / Entrance Paths — Even though it’s important to have well-targeted pages, you can’t lose sleep over what pages the visitor to your website is going to enter through. You don’t determine what your homepage is – the search engine algorithms determine what your homepage is going to be, and that is something that you cannot control. All you can do is ensure that every page is keyword-rich, content-oriented, and appropriately targeted to one key phrase – you will spend lots of unnecessary time trying to figure out how to get your visitors to only come in through your actual front door.
What it is — Any clicks from Google AdWords, Yahoo Search Marketing, MSN AdCenter, or any other pay-per-click initiative can be configured to track separately from your organic traffic. Check with your vendor to see how to do it – depending on your analytics platform, you may have to either add in some additional tracking code on your website or append some query parameters at the end of your destination URLs.
What to measure: Top Landing Pages + Bounce Rate — Fill in the gap between AdWords or YSM and your budgetary decisions with evaluating your pay-per-click landing pages, and how the visitors from your paid marketing efforts engage and respond to your message. You may have a high Quality Score and a high click-through rate, and everything may look fine on the surface, but use a Top Entry Pages / Top Landing Pages report to dig deeper and weed out potential problem pages. It's not enough anymore to just keep increasing your click-through rate – you must also increase the success rates of your promotional efforts as well.
Don’t worry about: Time on Site — A Goal of a pay-per-click campaign could be to increase the number of product sales or to increase the number of RFQs every month. Here, your ROI, ROAS, Bounce Rate, and Conversion Rate are all vital measurement statistics in determining the success of your efforts. Time on Site – and especially the average time on site – is great for informational purposes only. It’s not going to be able to provide you anything that you can take to your marketing manager or webmaster to optimize for, and it’s not going to help you decide if your dollars are being spent wisely. Even though time on site is a popular metric in web analytics, take it with a grain of salt when it comes to your cost-per-click marketing.