SES Toronto : Web Analytics Track

First up is Bryan Eisenberg, co-founder of FutureNow.  ROI, Engagement, Attribution & More.

Attribution:  Apparently Google Analytics inflates search results. What’s working what’s not?  Uses example of a basketball team.  do the players that pass the ball get “credit ” for their points?  Non converting keywords actually may convert as the attribution maybe it influenced later searches.  Analytics will never be purely a science because it is not 100% exact.  It is used for trends.

How are things being reported?  Google only counts the “player” that made the basket, not the ones that contributed.  We are still operating in the dark if we do not understand visitor profiles and do not use personal identifiable data.

Everyone is concentrating on the conversion, without paying attention on the process before conversions.  If it takes 7 weeks to change a button color it is death for companies today.   There has to be flexibility in an organization to make changes and do tests.  An organization has to constantly test over and over.  A contiuous improvement project.

Creating a Marketing campaign, measuring its success is what pretty much everyone does.  However improving campaigns is what is needed.  Amazon.com is constantly adapting, changing and making the experience better.

The process to become more flexible is to have these resources:

Marketing & Analysis
Graphic Design, copywriting, creative resources
Technical

Conclusion:  Measuring Success = Money in the bank.

June Li from ClickInsight.ca: Turning Data into Dollars
Data and Reports don’t make you money.  They only are the beginning of the journey.

Its a conversion funnel:

Data and Reports -> Ask Questions -> Segmentation analysis -> Action -> $$$

The tools don’t provide meaning, the analyst does.  Without an analyst the tool is useless.   An analyst should be involved in the design of how the data is gathered.  Get invited to the design party:

Define Goals -> Design -> Execute -> Cleanse Data -> Analyze -> Decide -> Test -> Assess

What are the behaviors that will be used to measure success? What needs to have extra tagging if the tool isn’t set up to do it by default?

Segmenting Data:  The question to ask is:  “What’s different?”  between people that convert and tose that don’t.  See what is working and what is not.

We see 4 differently shaped funnels,  Web Metrics, Proven Methods for Measuring Web Site Success by Jim Stern.

Work backwards from the conversion event.   sources, keywords, geographies, days of week, what content is being consumed and is there a difference in these segments and how they convert.  Reports are to show if objectives were achieved, Analysis will show where more gains are possible.

We see a case study of a site redesign and go through the metrics of how the website has improve in terms of visits, new visits, bounce rate etc.  Organic conversions increased by 90% but other sources ncreased by 37% so this difference shows that something further can be done..  drille down to keywords and measuring the bounce rate to show what users may be expecting to see on the site.

The important thing is to take action, segmenting gives ou the clues to what you can be doing.  No Action = No Dollars.

Anne Marie Lorriman from Outrider: Search Query Audits — When targeting goes Awry

Exact Query Reporting

Issue #1 = The offside search Result

For example the word “Sonic” can be a toothbrush, a hedgehog, or a DVD publishing company

“Primus” Rock Band, Camping Supplies, or Telecommunications

We cannot anticipate what people are expecting with queries, so you have to go back and check the performance of these Keywords.  Broad match is ok but has to be looked at very closely.  The Quality Score will also go down if the clickthrough rate is not there.   Irrelevant clicks will also cost advertisers:  Monkey Love” is seen on an adwords ad for Ebay, so they are wasting their money for that term.

Issue #2 The hidden defect search result.  If you are bidding on the word “automobile” but only the word “car” appears on your page, your click rate will go up.   So they will create a “car” campaign and an “Automobile” campaign.

Conclusion = An Exact match Query Report with Negative keywords will create a more performant PPC strategy.

Richard Zwicky Enquisite: Analyzing valuable traffic

Data is worthless if not used properly.  Search Analytics is to measure what users have done before coming to your website, whereas Web Analytics encompasses internal traffic behavior.

Organic and Paid traffic behaves differently.  Where do people look on the SERPs page?

Currently Agencies spend 96% of their advertising money on 12% of total return.  Only 50 out of Fortune 500 companies actively do SEO.  When we analyze specific keywords referrals from organic search and segment by geolocation, etc. we can find low hanging fruit.  Example of a california website that wanted to rank for “hotel” in the UK.  They were ranking pretty much everywhere except Londong.  When they noticed this they actively tried to acquire links from local London businesses and the rankings increased for the London area.

Potential =  Keyword Volume/Search Engine Referrals x (Page views) x (Time on site) x (1- Bounce rate)

Then the Q&A discussed attribution of the originating source of a referrer.  I suggested the nooverride function in Google Analytics, but it was pointed out that this will also cause data for repeat visitor sources to be lost.

Leave a Reply

Your email address will not be published. Required fields are marked *