Analytics RX: Diagnosis and Recoveryby Lisa Barone on 03/22/2011 • No Comments | Internet Marketing Conferences
And we’re back from lunch! Did you grab food? Coffee? Something? You have to keep up your strength, you know, if you hope to make it through the next few days. I scarfed down some pasta. And a mini black and white cookie. That cookie is probably the best thing I’ll eat all week. But…um…anyway. Yeah, the conference.
Up next we have a solo presentation from one of my long-time favorites, Matt Bailey. I haven’t seen Matt in ages so there was a joyous hug when I entered the room earlier. Matt’s good people. He’s also smart people. Which brings us to this session. Ta-da!
Matt said he’s feeling a little naked. It’s his first time not presenting in a suit. He also doesn’t have a room at the hotel yet. Sounds like someone’s having an AWESOME trip. ;) Hopefully he doesn’t end up sleeping in the speaker room tonight.
Matt asks everyone to sit closer together…so that they don’t die from frostbite. It’s a wee bit chilly in these back rooms. I blame whoever was complaining about it being warm earlier. I should have beat him with my laptop when I had the chance. Oh…we’re starting. Okay, then!
We’re going to talk about what to do when you get a site for the first time. What do you do? What are the first few reports you should run? How do you know what’s going on with the Web site and how healthy it is?
If you are generating reports aka caveman analytics of page views, hits, top 10 pages, monthly visitors, time on site, etc, you’re not going to diagnosis or recover anything. We can’t keep doing this. We have to go far beyond the surface level of analytics. The way to get good analytics out of Google Analytics is to make it fit your business.
1. Find Errors First: Find the 404 errors using analytics. The Webmaster tools only tell you what Googlebot found. Finding 404 errors in your analytics tells you errors people found. Go into your top content and then look for your 404 page to see how much it comes up. You might be pleasantly surprised and you might be shocked. The advantage here is that you’re seeing the errors that people saw. This affects the spidering of your site. It can help your content come together and the customer experience. Look through the error 404 pages that show up and look at the content detail. You can look at the bounce rate, how many people were there, etc. You can click on the navigation summary and it will tell you where people are finding it and what people are doing when they find it. They’re probably leaving the site. You can adjust to areas as they’re critical.
2. Detect Duplicate Content: Sometimes we see this in the Webmaster tools as well, but often we use analytics to find duplicate content. Filter top pages by URL strings to find duplicates by looking at the strings of content. You can tell a lot just by the URL string.
3. Segment your data: After you get through finding your errors, segment the data to find out how well you’re doing in different business areas, product areas, general keywords, personas, etc. You can only get really good, down-to-earth detailed information by segmenting your data. Don’t segment your data by things like page views, base it on performance. You want to measure the score, not shots on goal. Ooo, nice sports analogy. That doesn’t give you any information you need to take it to someone and tell them you have to change the Web site. You’ve got to associate the data to other data that gives a perspective and adds context. Keep adding more and more context until it answers the question, “Why”. The more you ask why, the more you’re digging into the data.
Question-asking is the single greatest tool humans have – Neil Postman
We don’t ask enough questions as analysts in order to get a grasp on where the business is going. We need to ask more questions. Asking why distinguishes a great analyst. Basically, we all need to act more like toddlers. This should NOT be a problem for search marketers. Hi. Twitter.
The best way to talk about segmentation is to talk about Star Trek.
Analytics according to Captain Kirk
Crew of 450
54 total deaths = 13.7 percent mortality rate
- 11 percent – Yellow shirt
- 9 percent – Blue shirt
- 79 percent- Red shirt
Why? If Red Shirts beam down to the surface with Captain Kirk, 57.5 percent of the time they’re not coming back. 42.5 percent of time if there’s a fight on board, the red shirts are getting vaporized. 16 percent if Kirk meets an alien woman.
Say you have an electronics Web site. You have a visitor who comes and is trying to locate a digital camera. They’re looking for price, brand, size, battery life, etc. On the other side of the site, someone is looking for an MP3 player. You need to understand that you can’t classify these two people with a single conversion rate. It doesn’t tell the story. It actually ignores the story that you have two people looking for two different things on two different sides of the Web site. Each group needs a new conversion rate. People see your Web site differently based on what they were looking for. By segmenting our data we can better understand experiences by similar groups of people based on intent, what they wanted to see, and their result. Don’t measure success by rankings, measure rankings by performance.
Before you can even do that, you have to think about your goals and have them established in your analytics.
You can only measure success when your goals are clearly defined and when they’re in your analytics. Then you measure by the exact thing that makes you money as a business. If you’re not measuring by goals that make your money, you’re not measuring anything.
4, Performance provides direction. You can be selling a lot of low revenue product or you can be selling a lot of high revenue products. You can find revenue based on visits in your analytics. Beware of your margins. You can also get into certain content to see what keeps people on the site and what content kicks people off the site. Identify the pages that keep people on the site vs the pages that send people off the site. Don’t just measure by exit rate, you can measure by revenue. You can see the pages that are producing revenue and the pages that are subtracting revenue.
5. Segment by entry points: Just having a number one ranking does not bring along profitability.
Primary Keyword had an conversion rate of 2.2 percent. The client had two pages in the rankings for the same word. Their homepage ranked at number 3, the category page ranked at #12. They found that though the home page brought more people, folks entering the category page saw what they wanted and they were converting at a much higher rate. They chose to de-optimized the home page to push the category page further up in the rankings because it had a higher conversion rate and a lower bounce rate. Just because a page is ranking best for a particular term doesn’t mean that’s the page that SHOULD be ranking. You want to look for keywords and landing pages. Which page should be ranking for which keywords? To find out, measure by performance.
6. Identify bounce rates: A lot easier to figure out when you start segmenting. People bounce when they don’t find something specific to what they were searching for. That’s how you solve the mystery. He talks about the Weighted Sort filter in Google Analytics. When you click it, all the data gives you a mix of visits to bounce rate. You can identify a certain keywords with a number of visits and bounce rates. By using Weighted Sort you can more narrowly target which bounce rates and keywords are providing a high bounce rate and determine what needs to be done.
7. Follow the money: The more you follow the money the more effective you are as an analyst. That’s how you get people’s attention. People with Cs in their title respond when there’s a dollar sign. Focus on the end result of making the company money. As an analyst you need to make things easier to understand. When you get into analytics you have to track your goals. If you don’t have goal value in your analytics how are you measuring success online? Goal value is critical to understanding what’s going on otherwise everything is just shots on goals. When you have all those values in place it turns from measuring goals, hits visitors to measuring MONEY.
8. Take Action: Companies that brought in a full-time analyst have realized a 1200 percent return on investment. And that’s because they followed the money. He loves what he does because analytics make money. Those social media people next door aren’t making money. Heh. When someone is in the analytics tracking the money and breaking down the experience, you can’t help but being successful.
How to build a segment
Think if keywords as a big bucket of information. Start segmenting data by keywords. You have big buckets and small buckets. Your big bucket may be “watches”. Your small bucket is terms like “repair”, “vintage”, “Rolex”, “discount”, etc. This should mirror your keyword research. In Google Analytics, click to go to Advanced Analytics. It will expand to fill half your screen. Go all the way to the left side of your screen and it’ll say “set up custom segments”. THEN you’ll be asked Dimension or Metric. A Dimension is a keyword group – it’s a dimension of the traffic you want to segment. The Metric is KPIs around that group. With the Dimension you can say you want everyone from PPC. With Metric you can say”…that stayed on my site for more than five minutes”. If you’re setting up your first level of segmentation, click off Matches Exactly and go to Contains. Type your keyword in where it says value.
After you do that, scroll to the bottom of the page where it says “Name This Segment”. You must name it or it won’t work.
By looking at small things, you’re evaluating multitudes of customer experiences on your Web site.
And we’re out of here!
About the Author
Lisa Barone co-founded Outspoken Media in 2009 and served as Chief Branding Officer until April 2012.