A few months ago I attended a Guardian Master Class on Data Visualisation hosted by Tobias Sturt, Head of creative at Guardian Digital Agency and Adam Frost, Data Visualisation Manager at Guardian Digital Agency. Both were incredibly friendly, down-to-earth and made the class very enjoyable. I highly recommend checking them (and their work) out. Data ...
On stage we have a line-up so big it's like a Blazin' Squad gig.
Mikkel de Mib (our MC) - deMib
Richard Baxter - SEOGadget
Christine Church - KeyRelevance
Lasse Clarke Sorgaarde - MediaCom Denmark
Kevin Gibbons - SEOptimise
First up, Richard baxter...
Data is problematic in that it needs to be organised into actionable segments. How do we get from a data-set to an information architecture?
Can we group keyword types together into an actionable "journey" as such?
Baxter demonstrates an example looking at "buy used audi s3 black london" and segments the token types e.g "buy" is an action, "new" is a condition (and so forth). However, before we can get on to this segmentation principle we need to understand the demand curve. (Specific to product).
Using such data (typed) gives a richer understanding, providing we have the analytical tools to do so; which leads us to excel and a couple of recommended commands.
Baxter shows a spreadsheet with segmented query tokens per column, e.g. Make, colour, action etc.
This query checks an array for strings that can be found common to groups, which can then be charted to give insight into demand per token type.
Increase data sets - Improve your perspective
- Download all of your analytics data
- Use the suggest API
- Capture rankings
- Build sweet data models
Why is this awesome?
Such data reveals more about discover mechanics e.g. "how" do people search for my product.
Christine will be focusing on tools and is a huge proponent of the use of varied tools for a greater perspective and varied data-sources.
Google KWT = Christine points out that in Sept 2010 GKWT dropped "partner sites" data at this time.
Christine gives an overview tool of some of her favourite tools, from Google tools to SEOMoz.
Lasse Clarke Sorgaarde
Preso is entitled "Searchenomics", so hoping for some interesting data observations.
Setting the scene Lasse points out that first and fastest is not the winning technique in today's market-place.
Using a supermarket shelf analogy Lasse compares the decision funnel to position on shelf. Essentially we need to funnel people to "eye level" goods. Purchase intent is a key principle to understand.
Understanding the relationship of term type to customer intent is key to driving ROI - which is really the most sensible objective.
Kevin Gibbons - SEOptimise
Kevins Preso "10 Steps to Target Long Tail Traffic"
How do we optimise for this traffic when Google themselves say 20 - 25% of queries have never been recorded before! How on earth do we understand/predict/optimise for them?
Kev shows us a case study for Audley Travel, for whom they increased longtail traffic by 181% - such traffic being good quality with high propensity to convert.
1. Find common trends KW add-on - good example looking at GTrends "how, why, when, who, where"
2. Answer industry-specific FAQ's
3. Pick out "themes" in GA
4. Use advanced segments in GA e.g. only visits from 4+ token keywords
5. Use PPC data (impression and SoV) - even if just for a month as a data collection exercise. (Broad and phrase to get best data-spread.)
6. Use multiple tools (Hitwise data is explicitly mentioned)
7. Estimate average CTR, as this varies wildly in the tail.
8. Use excel to predict traffic values
9. Filter KW into thematic groups
10. Don't overthink it and go too long tail - work smarter.
Now for the Q&A
A question from the audience about PPC data tests - diffs with seasonality in short-terms tests. How can other data sources be used to normalise?
Kev agrees this is a problem and recommends Hitwise for best data sample for normalising and also points out to exclude brand terms (inc. misspells).
Q - What about webmaster tools data? Does it include paid search?
Kev - No
Richard - Avoid Google Keyword Tool for anything other than a yardstick measure.
Q - How important is first to market on a new KW e.g. "2011 car insurance"?
Richard - There does seem to be some advantage to being first with content.
Christine - Recommends to back up with other marketing techniques e.g. twitter etc.
Kevin - It's important to get this out as soon as possible. Citing an example of a Billy Connelly TV program where they pushed out client content months in advance, beating ITV themselves when the show was aired.