Analytics for SEO
10 likes3,766 views
This document provides tips for getting an SEO analytics program beyond the "stone age" by focusing on opportunity, competition, and attribution. It recommends using analytics tools to find opportunities by analyzing popular pages, keywords, and competition. Proper attribution is important to understand what is truly driving conversions and avoid costly mistakes. Log files can provide a better picture of attribution by showing all customer visits and referrals before a conversion.
Analytics for SEO
- 1. How to: Get your SEO Analytics program beyond the stone age.
- 2. First, a little fine printA lot of the techniques in here aren’t mine. They’re brilliant stuff by brilliant people. I provide links and credit where relevant.Also, analytics changes all the time, so by the time you read this, it may have all changed, blah blah etc. etc.Finally, this is based on a presentation I gave at SMX West 2010. I’d be a moron if I didn’t say ‘thanks’ to everyone at the conference who came and watched, invited me to speak, and didn’t fling rotten vegetables when I periodically fell on my face.Are you actually reading this? Seriously? There’s nothing to see here, I swear. Move on to the next page. All is well. This is what we call ‘greeking – text used to fill space when we don’t want to fill it with real writing. I’m just randomly writing down what comes to mind. And if you’ve read this far you REALLY NEED TO MOVE ALONG NOW OK?!
- 3. This is how most people respond to analytics: It scares them.
- 4. This is how most people respond to analytics: It scares them.But, as I like to say: If you bury your head in the sand, all you can do is talk out of your arse.
- 5. You’re better served pulling your head out of the sand and taking advantage of all that analytics can do for your SEO campaign.
- 6. You’re better served pulling your head out of the sand and taking advantage of all that analytics can do for your SEO campaign.Just focus on:OpportunityCompetitionAttribution
- 7. Opportunity is the gap between what you’re getting right now from organic search, and what you could have.Competition tells you how hard it’s going to be to close that gap.Attribution is what you use to determine exactly what generated sales or other conversions.
- 8. You have to remember those three principles!!!! Here’s a sentence. Read it three times and I guarantee you’ll never forget them:
- 9. The squirrels saw a great opportunity to eliminate competition by tearing up the garden – if Jane used incorrect attribution, she’d kill the raccoons.
- 10. Told ya.
- 11. Another way to look at it:Opportunity and competition are about optimization.Attribution is about keeping your job.
- 12. Here’s how you can use analytics for the optimization part:
- 13. Most folks, when they start an SEO campaign, toddle happily over to their favorite keyword tool, generate a huge list, and start optimizing.
- 14. That’s not the right way to do it. A monkey could do it…
- 15. …but monkeys know better.
- 16. I’m not saying keyword research is bad.But, if keyword research is the sole driver of your SEO campaign, you miss some of your biggest opportunity gaps.
- 17. As well as some huge competitive vacuums where you can easily beat your rivals.Competition
- 18. In about 9 steps, you can analyze opportunity and competition, and upgrade your SEO campaigns.Note: All of the techniques you’re about to see assume you have an analytics toolset that was written after 2001.
- 19. Step 1: Look at the most-visited pages on your site.
- 20. Step 2: Look at the keywords driving traffic to each of those pages.
- 21. Step 3: Choose a good optimization target. In my case, ‘google analytics tutorial’ is a no-brainer: Lots of traffic, solid time on page, and good goal conversion.
- 22. Step 4: Now I’ll go look at a keyword analysis tool, and compare my monthly traffic on that keyword to the monthly search volume. I’ve got a huge opportunity gap
- 23. Step 5: Figure out what’s driving that search traffic. In my case, it’s a little video search result that’s showing up at the bottom of a standard Google search results page. So, we’ve got an opportunity. But can we compete?
- 24. Step 6: Determine competition. SEOMOZ gives us a nice shortcut with their Keyword Difficulty Tool.
- 25. Step 6a: You can also use Aaron Wall’s excellent SEO for Firefox plugin, which shows you the link strength and other data for every ranking page in a search result.
- 26. Step 7: Optimize the ranking page. I should probably change the article title to ‘Google analytics tutorial 1’, removing ‘video’.
- 27. I can also link to newer content, to try to get the newer stuff ranking.
- 28. A few Google Analytics reports I find super-helpful when I’m doing opportunity gap and competition analysis:
- 29. Show where you ranked for a given phrase when the visitor clicked.Andre Scholten/Nikki Raehttp://bit.ly/serpclicks
- 30. Show clicks from page 2 of the search results. That’s a fantastic way to find opportunities. If you’re getting traffic and you’re on page 2, imagine what you’ll get on page 1.Will Critchlowhttp://bit.ly/2ndpageclicks
- 31. And a report I came up with: Show page with no clicks from search.Ian Lurie (me)http://bit.ly/seoclickless
- 32. You can find all of this plus a bunch of other tips on my Google Analytics Cheatsheet:The Cheatsheethttp://bit.ly/GACheat
- 33. Now you’re using analytics to find SEO opportunities. Woo-hoo!But you still have a problem: How do you know what’s driving conversions and what’s not?
- 34. That’s the single biggest problem in web marketing today: Attribution.Most analytics tools will track the first or last click before a sale. They ignore all of the clicks in between.Look at the scenario on the next page.
- 35. Many shoppers will come to a site several times, from several sources, before they purchase or otherwise convert.PPC click on ‘indoor trainers’PPC click on ‘wattage trainers’SEO click on ‘tacx trainers’SEO click on ‘tacx flow’Banner ad clickDirect navigationClick from phredheaven.comSEO click on brand nameDirect navigation$In this case, Google Analytics would tell me direct navigation or PPC drove the sale. The visits in between are all lost.
- 36. This is the kind of problem that can create costly mistakes, or cost you your job.If I were using traditional analytics in the previous example, I’d assume my work to rank #4 for ‘tacx trainers’ was totally wasted. I might stop optimizing for it, break the conversion chain and lose valuable sales.
- 37. The answer: Learn to love your log files!They already have the data you need.With log files, you can get a solid attribution picture. No revolutionary technology needed.
- 38. You’re going to need access to your server log files to do the procedure in this section.You can usually download log files from your hosting provider’s ‘control panel’, or get ‘em from your IT team.
- 39. Step 1: Scrub the logs. You don’t need all the data they contain.Remove all records except real pages.Remove all records generated by bots.Then sort by cookie, IP address and date.
- 40. Step 2: Find your goal page. In my case, it’s the ‘colloquium’ page on my blog:
- 41. Step 3: Group records by IP address and/or cookie data. You can now see all of the visits by a particular customer – not just the first and last visits.
- 42. Step 4: Make a pretty graph for your boss.
- 43. There are lots of tools that can help you with this kind of analysis. With a log file of fewer than 20,000 lines, you can use Microsoft Excel.Splunk is a log file analyzer on steroids. Find it at Splunk.com.A scripting language like PERL or Ruby can do the trick, if you know the language.You can also import the data into SQL Server or MySQL and crunch the data there.Enquisite has a nifty logging tool built into Enquisite Optimizer that makes this a lot easier.Coremetrics has some great attribution features built in.
- 44. That’s it. One last reminder of those three essential words…
- 45. The squirrels saw a great opportunity to eliminate competition by tearing up the garden – if Jane used incorrect attribution, she’d kill the raccoons.
- 46. If you liked this, you might want to check out my blog, www.conversationmarketing.com.= @portentint = me
- 47. Or, if this seems important but made your head hurt, you can hire my company, Portent Interactive, to do it for you: www.portent.com. = @portentint = me
Editor's Notes
- #19: The reports I'm going to show you now are all in Google Analytics. Assuming your analytics tool was created some time after 2002, you should be able to get the same data. If you don't have analytics, I'll go with my standard strategy: Kick your webmaster in the ass.
- #20: First step: Determine opportunity. This one's easy. Go into your analytics tool. Find the busiest pages on the site.
- #21: Then look at the top keywords generating traffic to those pages. Look for keywords that are 'clear winners'. I could go into all sorts of statistical significance stuff but I can't do it in 11 minutes, so forget it.
- #22: Got your top keywords? OK, now check conversion rate or keyword quality. You can do that by looking at the conversion rate (duh) or, if you aren't tracking conversions (shame) by looking at bounce rate and time on page for that term. The bounce rate is the percentage of visitors who arrive at the target page via the target keyword and then leave your site, never to return. Time on site is just what it sounds like.
- #23: NOW look at the keyword tools
- #24: Almost there, I promise.Now, check your site's rankings for these keywords. If you don't rank in the top 3, on page 1, then you have an opportunity gap. We know that, because we know roughly how many people will click on each listing in the SERPs (Search Engine Rankings Pages). Here, I’m getting clicks for a lousy little video blended search listing, at the bottom of page 1.So, we have an opportunity. But do we stand a chance? Or will we come up short?
- #25: Or, if you're really lazy, you can use something like SEOMOZ's keyword difficulty tool. Some of us are old-fashioned though, and prefer to see how the sausage is made.
- #26: Link strength of the top 5-10 pages. You can do this using Aaron Wall’s SEO4Firefox, for starters. For more detailed data you can grab it by hand using a tool like SEOMOZ’sLinkscape.
- #27: Now I need to optimize the ranking page, and maybe start funneling some link authority around to better take advantage.
- #28: Now I need to optimize the ranking page, and maybe start funneling some link authority around to better take advantage.
- #30: A few clever custom reports that can help with this analysis:You can also use a snazzy Google Analytics filter by AndreScholten and Nikki Rae to show you the rankings when that term got clicked. It's the coolest
- #31: You can also grab all keywords that generated clicks from the 2nd page of search results. How helpful is that? Really damned helpful! If you're getting traffic from page 2, imagine what you could get from page 1. This is true opportunity gap stuff.<blockquote>Note: I have all the reports I mention in this presentation in a Google Analytics Cheatsheet that you can grab at <a href="http://bit.ly/gacheat">http://bit.ly/gacheat</a>. Don't sprain your fingers trying to write/type this all down. I'll also post a relatively close transcript of this presentation to my blog tonight.</blockquote>
- #32: Lots of sites have pages that get zero search traffic. It's OK. It happens to everyone now and then.You can use this simple pivot table report to get a list of zero-search-traffic pages in Google Analytics:Why do it? Because the more pages you have getting measurable search traffic, the easier it is to find opportunities. Voila.
- #33: All of these links are in the Google Analytics Cheatsheet, by the way. My only shameless plug. I figure it’s OK since it’s free.
- #36: Most analytics tools just credit first or last clicks. That’s bad. You can easily end up missing important referrers. As well as losing your job as SEO.What you need is a model for accurate attribution: Something that lets you see each resource that generated a click by that person.
- #38: Learning to LOVE YOUR LOG FILES.They’re the sole repository, at this point, of attribution data you can get your hands on and really investigate. Make no mistake – accurate attribution ain’t easy (yet). But it’s definitely worth doing, and if you’re serious about SEO and search you need to learn your way around your log files anyway.
- #39: Step 1: Get the log files. If someone says no, well…
- #40: Step 2: Scrub the log file.
- #41: Then you get a spreadsheet. Use conditional formatting and you can quickly find folks who hit conversion pages. None of my clients want me publishing their logs at SMX, for some strange reason, so this is my sample data. But have a look – with first click you get google image search for mr. 155. With last click you get Google India. For Ms. 238, you end up with ‘Facepunch’ or ‘Organic’, but PPC generated 2 clicks. Facebook did one. Cracked did 1.
- #42: Then you get a spreadsheet. Use conditional formatting and you can quickly find folks who hit conversion pages. None of my clients want me publishing their logs at SMX, for some strange reason, so this is my sample data. But have a look – with first click you get google image search for mr. 155. With last click you get Google India. For Ms. 238, you end up with ‘Facepunch’ or ‘Organic’, but PPC generated 2 clicks. Facebook did one. Cracked did 1.
- #43: Collate, get a count, and you can at least show what’s been involved in sales. And make a pie chart that’ll clearly convince your boss you’re right. How can you not be? It’s a pie chart!
- #45: Step 1: Get the log files. If someone says no, well…
- #49: <h2>That's it</h2>That's it - 11 minutes on the dot. If you want to get started quickly, start with opportunity and competition. Then move to attribution.I'm happy to answer questions during the Q&A, or you can contact me at these addresses: