I’ve found the authority on the topic and can confidently say, he’s the go-to guy regarding web analytics.

Avinash Kaushik is author of Web Analytics 2.0. He’s on Twitter, he has a blog…there’s no reason why you shouldn’t find him and connect to him to stay up to date on this ever-changing topic.

In his own words:

Web analytics is: 1. the analysis of qualitative and quantitative data from your website and the competition, 2. to drive a continual improvement of the online experience of your customers and prospects, 3. which translates into your desired outcomes (online and offline).

The importance of Outcome

This is why I love the guy. He stresses what I’ve always believed: don’t do it, unless you can’t measure it’s business outcome. At the end of the day, if there is no impact on SALES, COSTS, OR CUSTOMER LOYALTY, ask yourself why you are spending the time and money to collect data, weave it into a report, then pat yourself on the back.

Avinash also provides great clarity on key terms. Lets look closer at those.

Sources of Traffic

Avinash explains that if we tag our marketing campaigns correctly, then Direct Traffic represents free traffic because it comes from people who arrive via using bookmarks, typing in the URL, or other activities. It also represents traffic that is familiar the website – so may represent returning Visitors and existing customers.

What Sources are Telling us

Twitter traffic: tech-savvy people, possibly bleeding-edge kind; looking for instant real-time results/informations/answers

Stumbleupon and Digg: might be interested in recent stories or promotions, product launches, etc. “These sites represent a sense of valudation that your content is good and is being spread by others whom you don’t know.”

Google: these numbers will validate the hard work you put into your SEO and whether it is paying off. Traffic from, images.google.com, for example, is derived from all the tagging and relevant descriptions applied to each image on your site

Visits

• “Visits report that someone came to your website and spent some time browsing before leaving.”
• The visitor experience = “session”

Unique Visitors

• The number of people who come to your website
• “It is likely, but not always true, that each unique visitor is a unique person”
• Can be influenced by growsers that don’t accept cookies (or reject 3rd party cookies)

Bounce Rate

• Fairly standard metric in most tools
• Hard to misunderstand
• Actionable – identifies “low hanging fruit”
• Measures customer behaviour
• Here’s the best part, Avinash defines it as, “I came, I puked, I left”
• Technical definition: the percentage of sessions on your website with only one page view

Some thoughts on what to do with your bounce rate results, because afterall, it really is a very actionable piece of information:

• “Measure Bounce Rate for your website’s top referrers.” This means, those sites who send you traffic that doesn’t bounce (eg. not just random traffic that comes and leaves)
• “Measure Bounce Rate for your search keywords.” – both paid and organic. This will help you determine if you are optimizing in the right direction or not. Either the landing page isn’t compelling or the keyword opt isn’t working – action can be taken on either.

Avinash’s love for bounce rates doesn’t end there. Here’s how he further supports it as the best analytic around. It passes what he calls the Four Attributes Test:

Uncomplex: it measure single-page-view visitors. Or, “I came, I puked, I left.” It’s easy to understand, explain, and propagate. Enough said.

Relevant: it identifies where you are wasting marketing/sales dollars and which pages stink when it comes to delivering on the “scent.” Those two things apply to most web businesses. Bam! (-seriously, how funny is this guy?!)

Timely: bounce rate is now standard in pretty much every web analytics took and available in every report. Every day. Nice!

Instantly Useful: you can just look at it and know what needs attention. You see a 25-30% bounce rate for your site, and instantly you know things are fine. You look at a page with 50% bounce rate, and you know that pages needs attention. You see a campaign or keyword with a 70% bounce rate, and you know there is a fire.

Conversion Rates

‘Tis the best, isn’t it? I love conversion rates as they have pure and clear business application. It’s the kind of thing you can show to your boss and say, “SEE. I told you so.” Ok, perhaps word it a little differently to save your hyde.

Avinash describes this as Outcomes divided by Unique Visitors (or Visits). By outcomes, he means a submission of an order on your ecommerce site, but you could also be measuring leads as well.

So why the option for the denominator ? He explains…

Using Visits: you assume that every visit to your website is a chance to get someone to place an order and get someone converted.

Using Unique Visitors: you grant that it is OK for a person to visit your website multiple times before making a purchase (realistic for the web, let’s be honest).

Conversion Rate: a Methodology

Avinash provides a methodology that can be used to, “do a real root cause diagnosis of top key performance indicators.” Or what the hell is going wrong with one of your metrics. Here’s the scenario he provides:

Your boss comes into your office and tells you to improve Conversion Rate by 10% – which is huge! What do you do? Should you run out and spend a ton of money on affiliates email campaigns, or paid search ads? Should you run to identify the demographic profiles or people who visit your website? Ack!

Instead, start by identifying all the variables that could cause Conversion Rate to go up or down. After all, before you can go out and improve Conversion Rate, you need to identify all the influencing levers.

• Acquisition strategy (where you spend money to acquire traffic)
• Organic search keyword ranks
• Ease of checkout process
• Distribution of why people come to your site (primary purpose)
• Website “scent” (ability of your campaigns to deliver traffic to the most relevant pages
• etc.

He then recommends you collect data for each of the variables above, and by analyzing that, you can identify where there is room for improvement (eg. where conversion rates stink). Then do a CBA of where you can get the most bang for your buck.

Here’s his important take away on this exercise:

• This exercise is of tremendous value
• This exercise is hard
• You can’t improve what you don’t understand (duh!)

Final thoughts: let the data, not opinions, drive action.