Insights Analyst, Angus Carbarns, explores the rising role of machine learning and the confluence of SEO and CRO in creating truly effective user-first search experiences. He delivered this talk at the ManyMinds conference in London.
2. • I work at Dog Digital, a full-service
design and digital marketing agency
• My formative ‘digital’ years were spent
doing outreach at a big agency,
followed by stints in content and site
strategy at some small agencies
• I now help clients use analytics,
personalisation and CRO to solve their
customers’ problems
• As a psychology and sociology
graduate, I’m interested in the
‘what’(quant) and the ‘why’ (qual) of
human behaviour
@dogdigital
8. Primarily link-based, relatively
static, predictable ranking
model.
Multiplicity of factors
(including links), resulting in a
more personalised search
experience.
Driven totally by machine
learning? Likely adaptive to
personal needs and intents.
@dogdigital
9. 9
Image: Gianluca Fiorelli, Moz
“It (RankBrain) works when people make
ambiguous searches or use colloquial terms,
trying to solve a classic breakdown computers
have because they don’t understand those
queries or never saw them before.”
Greg Corrado, Principal Scientist Google
18. 18
• Identify the purpose of the landing page.
• Identify main content, supplementary content
and advertisements. Is it easy to identify the
main content immediately?
• Review main content with regard to the purpose
of the page.
• Determine the amount of useful main content.
• Determine the benefit of the supplementary
content.
19. 19
@dogdigital
“Google could see how satisfied users
were... The best sign of happiness was
‘the long click’… That meant Google had
successfully fulfilled the query. But
unhappy users were unhappy in their
own ways, most telling were the ‘short
clicks’”
Steven Levy, In the Plex
Source: Cyrus Shpard, M0z
23. 23
@dogdigital
“…When you add in more images and other
elements that make pages more complex, those
sessions converted less. Why? The culprit might be
the cumulative performance impact of all those
pages elements. The more elements on a page, the
greater the page’s weight”.
“Bounced sessions had median full page load
times that were 53% slower than non-bounced
sessions”
Tammy Everts, SOASTA
33. 33
• What are we trying to
improve?
• What issues are we aware of
(to guide step 2)?
• What does ‘good’ look like?
• Create a framework
• Traffic breakdown
• Device insights
• Analysis against key
Goals
• Qualitative analysis
• Synthesise key findings
• How can we improve Goal X?
• Hypotheses informed by data
• Prioritise based on projected
impact, difficulty, resource
• Test hypotheses
• Implement if successful
• Continually iterate and
deploy if not
• Challenge biases!