Day one at Search Engine Strategies New York 2011 and the analytics tracks are in full swing. I started the morning presenting Introduction to Analytics and followed it up after lunch attending Matt Bailey's Analytics RX: Diagnosis and Recovery, where he discussed where to start and how to diagnose a site after you inherit it.
For those who need the short-short version: Introduction to Analytics was all about knowing what data you have and how to read it. Know your software. Every analytics package uses different terms and collects data differently. Knowing this, combined with a little common sense, can make all the difference in your testing and future planning.
Analytics RX can be summed up by a series of actions: Find all 404s and duplicate content issues. Segment your data and rank by performance. Assign goals and values and take your actions by following actions for where the opportunity for a greater return leads you.
The Intro to Analytics session included a room full of people, half of whom who were intentionally using some sort of analytics software. As an introduction session, there were many definitions and terms thrown out at those in attendance.
Lesson 1: Keep your terms straight
The attendees and I pondered the difference between a hit and a page view and had excellent post-session Q&A about the subtle difference between a new visitor and a unique visitor, as defined by Google Analytics. To set the record straight, a new visitor is a first-time visitor, a unique visitor can be either new or returning, but requires a specific period of time to be counted as such.
Lesson 2: How you collect your data matters
Lesson 3: Don't use "Caveman Analytics"
Merely printing off reports of the top-10 pages viewed or visitor counts and graphs of time on site "is so 1990s," said Matt Bailey. Those metrics, when taken alone, mean nothing. Segment your visitors and tie those reports in to your meaningful visitor segments.
Lesson 4: Have a Goal!
Matt presented an interesting analogy comparing visitors and other "caveman analytics" reports to shots on goal. Without specifying a goal and tracking which visitors actually convert, your segmented reports are like measuring a team's shots on goal, but not actually measuring how many goals were scored. No one ever talks about a losing team that had a lot of misses. They talk (and monetize) the actual goals scored.
Lesson 5: Page ranking is less important than conversions
If two pages rank for a keyword, the one that ranks higher is not necessarily the best page. In an example, Matt showed two pages, one ranked 2nd, one ranked 12th for the same keyword. While the one ranking 2nd brought three times more traffic, the one ranking 12th yielded more than three times the conversion rate. What is the take away here? Get the lower page to rank better, and the higher page to rank lower.
You'll hear it in many sessions: perform your SEM functions, test and repeat. As I said early this morning, Analytics is the testing part. Without proper analytics, you can not get meaningful test results and your retries and retesting will not give you the right results.