Moderator: Jake Baillie
Richard Zwicky, CEO, Enquisite
Craig Hordlow, Chief Strategiest,
Paul Botto, Head of Analytics Sales, Google
Paul Pellman, CEO, Click Forensics, Inc.
Measuring success in the online world. Richard starts off by talking about the traditional world of analysis, the 1.0 world, and the current 2.0 world.
- Page views
- Hits lack of standard measurements
- Undefined ROI
- Lack of controls
- Outside of PPC there is an undefined ROI that analytics can’t track effectively.
- Web 2.0:
- Provides deeper insights, the 2.0 world represents a new and better way:
- Transparent measurements
- Defined, precise ROI
- Follows your workflow
- Keyword potential = (1-keyword volume/search engine referrals)(pv)(time)(1-bounce rate)
- In relation to all other search traffic to the site and conversions:
- Referral rates – the number of referrals for a given term
- Conversion rates – the number of conversions a particular term generates
- Page views – derived from specific terms
- Time – in seconds that user through a specific term spends on-site
- Bounce ratio for the phrase
2.0 is information the way you want it (picture of a map overlaid with information)
Use the map overlay to see geographic distributions. Understand how a searcher’s physical location impacts your search results and business.
Web 2.0 analytics allow you to look at very narrow sections of your traffic with respect to all of these metrics. A modern analytics package lets you see data over time and metrics for really specific segments. For example:
hotels, in USA, from Google, page-2 traffic
In this real-world example, looking at a map overlay of only page 2 traffic shows New York and area has a lot of page two Google rankings. Data grids don’t show you this stuff. 2.0 is the mashup of different types of information to gain new or different insight.
- How we collect data might change
- Channels to reach customers will change
- How we use the data will change (static and dynamic content creation, reputation management, social media, next?)
- What we try to accomplish with the data won’t change (branding / sale)
- How we define value will evolve
Craig is up next:
Craig is going to talk primarily about what he calls “Motive analysis” – this is great because he’s really talking about searcher-intent as derived from the examination of search queries. For some reason this level of analysis often goes overlooked in SEO, and to a lesser extent in PPC.
Are you really, really listening? Marketers have an obsession, keywords: the ‘top’ of whatever. What about the ‘bottom’? The purpose of motive analysis is to dig deep, way beyond ‘top’ this and that.
Motive analysis has five steps:
- Identify visitor motive through nuances of the search query
- Segment the queries by motive
- Analyze the performance of each segment
- Identify motive disconnects
- Modify the disconnect on the entry page so it now speaks to that user
Basic necessities: website, spiders, analytics, and a brain
Ten motives of search:
- Product /service (athletic shoes)
- comparison/quality – top, best
- adjective qualifier – white shoes
- intended use – running shoes
- vendor/manufacturer – Gucci shoes
- location – san Francisco shoe store
- action request – buy shoes
- instruction – repair broken heels
- definition – what are pumps?
- problem – shoes for wide feet
Motive analysis – ranking for these terms is not enough, you must connect with the visitor, software cannot do this for you, people have many combinations of motives.
Use personas: powerful for understanding the user use the 10 motive of search to identify personas. Have a lowest common denominator persona in mind.
least common denominator persona = your mother
- shopping addict – brand, adjective qualifier
- local enthusiast: geo “Toronto”, intent “hiking”
- bargain hunter: intent (walking shoes), price (discount)
Review all referring keywords – start with the websites referring keywords VS entry page.
Begin grouping the keywords at a high level (cheap, discount, sale, bargain, low, clearance all in the same group). High level thematic grouping.
microsegmentation is grouping by highly specific phrases, which enables you to message back with equal specificity. The days of the one word search term are gone.
tip: generate a list of filters and segment. Google analytics example: filter by singular or combinations of keywords
- product name
Save these reports and schedule to run every month
The motive disconnect. When you find a motive that is not performing, it’s a motive disconnect. Usually when your content, value proposition, or business model are not aligned with your visitors.
What do we do about it? Bring it back to the entry page. Ask yourself:
- Is the message the user wants ont he page?
- is it succinct and persuasive with more info easily available?
- is it even feasible foryou to be relevant?(maybe you’re not cheap)
- is it visible enough
- should the message be on its own page? (cluttered pages, pages with lack of focus)
- Rebuild the entry page, build a new entry page.
- He wants us to build our long-tail (more pages, more!) without letting a motive disconnect occur. Which basically means crafting landing pages that concur with what your searchers are finding you for. I would argue true long tail comes form less structured or less predictable queries
Paul Botto, from Urchin is up next
How do the advanced segmentation features of google analytics relate to what we’ve been talking about?
- Isolate and analyze subsets of your traffic post-data capture
- compare segments and key performance metrics side by side
- analyze your traffic with predefined or customized segments
You can define a segment of visitors who ‘visit this page, then this page, then that page, but don’t buy’ as a comparison shopper. Save that as a segment and view all of those people and only those people.
(GAIQ) Google analytics individual qualification
The course is free, and the qualification is a nominal fee to take the test. This is a new qualification. It’s pretty challenging stuff, so even power users might want to look over the course (I know I will be, though I’m only a battery-powered-user).
Things are not pageview based anymore, it’s an experience like a movie more than a book. Webpronews for example has hundreds of videos. How do you track that? We used to sneak a pageview under the video called (press play on video). The report doesn’t translate well the experience the user had.
Clickforensics similarly has a software demo broken into 12 slides and a ‘schedule a demo’ button – how do we track or learn from the analytics on a page like this? Eeeeevent tracking! Yes, event tracking is the key to tracking these specific non-traditional navigation stuff.
Object, Action, Label
The object is the magazine viewer tool. Actions are turn the page, go to more tools, subscribe, send to friend, etc – The label would be: website magazine august 2008 edition
Google Analytics has a dedicated ‘event tracking’ section now. With top page objects, top object actions, and top object labels. No artificial pageview boosts.
They partnered with adobe, right in flash and flex, you can drop your GA code right into it. Build it into the process flow (like, always tag the play button, and every 10 second mark on the videos). It’s all drop down menu-driven stuff, but you have to get your flash and flex developers implement it.
Paul Pellman is up next
An Austiner who’s freezing his butt off, because today, it’s damn cold for Austin.
Within each keyword you buy there are different levels of quality based on the origin of the traffic. Invalid traffic is more than just click fraud:
Low quality: low conversion rates, cpa too high
Invalid traffic: outside geo, broad match conflict, unacceptable type (mfa, adult)
Click fraud: Viewing invalid traffic over time shows spikes that usually indicate malicious attacks.
Trends: Like spam, bots and botnets are becoming more sophisticated
Almost seventy percent of all sites in google adsense and yahoo are made for ads or parked
22 percent of all invalid activity comes from outside th targeted geo areas
Competitor clicks deplete ad spend
Brand infringement violations use established brands to send customers to competitor’s sites
Q and A
Comment more than a Q: Nobody talked about crazyegg, quicktale, and pagealyzer – allow you to see the actions of individuals in a page, how long do they take to fill out forms, poor man’s heat maps
A: beyond what happens on a page, identifying the opportunity for potential on a site as a whole is the next phase.
The vast majority of ppl using analytics are barely scratching the surface of what we have now – it’s not a technology problem, it’s a methodology problem. They want out of the box, ooh pretty reports, what needs to be updated is methodology.
Q: i get about a 17% invalid click rate – is this recognized by google analytics?
A: google analytics is not the right tool to analyze invalid clicks – but it is tied to adwords, so we pull that data out automatically. We update your analytics report to show exactly what you actually paid for.
Q: object action labels, are they available in google analytics now?
A: event tracking is invite only at the moment
Q: the tracking cookie in analytics is six months, isn’t that a bit long?
A: there are several cookies actually, placed with different time frames associated, so this could be a big long discussion. There is logic behind choosing those time frames, and adding the ability to change them is worth considering, but for now the choices are a best fit for analytics users as a whole
Q: its my understanding that the google analytics API is still in very private beta – is it opening up any time soon?
A: i’m not here to announce anything, if you give me a card i’ll get back to you. It’s an a exciting one, the API should be exciting for a lot of people here.