2. @timlb
Started working on web in 1993
!
Primary interest:
!
What people do on line
!
What people want to do
What companies would like them to do
http://www.bryaneisenberg.com/hidden-secrets-of-amazon/
10. @timlb
The hope
“Half the money I spent on advertising is wasted…”
!
The web held out such promise
Built by people who were used to data
And we could track ‘everything’
So much better than off-line
We seized on this valuable idea
and used it to build an industry...
11. @timlb
The Analytics Arms Race…
“If we just had more data”
!
Huge efforts on multiple session attribution
Huge efforts on cross-device tracking
Huge efforts on visualisation
!
GA joins the arms race and offers
Multi-session reporting
‘User’ segments
Sequence segments
(User id tracking)
12. @timlb
The Analytics Arms Race…
“If we just had more data”
!
Huge efforts on multiple session attribution
Huge efforts on cross-device tracking
Huge efforts on visualisation
!
GA joins the arms race and offers
Multi-session reporting
‘User’ segments
Sequence segments
(User id tracking) …chasing rainbows?
13. @timlb
But…
The tools forget about the people
Classic example:
!
Tools optimise emails for what tools measure:
‘Opens’ and ‘clicks’
We CAN optimise for people/value with hold-out tests
!
(But the easy metrics are in the reports and testing is
extra work. So…)
14. @timlb
Conversion rate also flawed…
Not: conversions / real person
Not even: conversions / unique visitors
GA Conversion rate = conversions /
“sessions”
!
http://www.kaushik.net/avinash/universal-analytics-implementation-tips-
strategy-tactics/
15. @timlb
And so are sessions or visits!
30 minute time out: arbitrary convention
http://www.simoahava.com/analytics/the-
schema-conspiracy/
!
Look at your own data:
Funnel entrances on impossible pages
“Returning” visits on the same day
!
http://bit.ly/timlbGaSegmentSameDayTransaction
16. @timlb
Meanwhile...
“$40 billion is spent on digital advertising in the US on an annual
basis, while between $2 and 5 billion is spent on designing services.
So, we set a big expectation and then don’t deliver.”
- Chris Risdon of Adaptive Path
cited by Gerry McGovern January 2015
http://www.gerrymcgovern.com/new-thinking/advertising-and-marketing-are-broken-how-do-we-fix-them
17. @timlb
Suggestion: re-balance effort
Spend more time trying to understand our users as people
Look for the “why” as well as the “what”
Find out what people are trying to do
Learn what’s stopping them doing it
!
Re-balance some of the $40 billion from advertising to the $5
billion for ‘service design’
18. @timlb
#cboostdk
Key Point: The Strategy
!
Think about people:
Think about what people want to do on sites
not just how to can get them there
!
19. @timlb
#cboostdk
Key Point: The Strategy
!
Think about people:
Think about what people want to do on sites
not just how to can get them there
!
Recruit analysts for empathy as well as
curiosity
20. @timlb
The tactics
Use quick techniques to seek out clues, not definitive data
+
Use cheap and quick tools
+
Cross-reference or triangulate clues and data
=
Start optimising for people, not cookies
21. @timlb
Don’t just go fishing
Photo: Karwik https://
www.flickr.com/photos/karwik/
23. @timlb
Who are the people on our sites?
Personas...
!
If you don’t use personas, time to start
!
If you do use personas, do you find them in GA?
!
!
!
https://www.interaction-design.org/encyclopedia/
personas.html
!
http://www.bryaneisenberg.com/personas-magic-behind-
mirror/
24. @timlb
What sources of clues have we got?
GA
Where people came from:
Campaigns
Non-campaign channels
What they did:
Path (pv, landing, funnels, flows)
Goals
Content Grouping
Errors!
Custom data
Maybe clues about who
Demographics
Non-GA
Feedback surveys
Usability tests
Scroll & Click maps
Session recordings
Customer file demographics
26. @timlb
#cboostdk
Why feedback surveys are so good
Very fast
No need to wait for data to reach scale
!
Very direct insights
People tell you their intention and expectations
People tell you exactly what caused their problem (or what specifically what they
liked)
!
Words are from real visitors
Valuable understanding of visitor language and terminology
27. @timlb
Everywhere (before conversion)
!
For
● Widest range of insight
● Open to non-converters
!
Against
• Low response
• Active invitation 2% - 5%
• Visitor-initiated 0.1%
[source iPerceptions]
• Pop-up invitations can harm conversion
• Visitor-initiated systems have lower response
rate
• Probably skew towards complaints
How to work with feedback surveys
Where and how to ask
Horror show invitation HT
@JamesGurd
28. @timlb
Everywhere (before conversion)
!
For
● Widest range of insight
● Open to non-converters
!
Against
• Low response
• Active invitation 2% - 5%
• Visitor-initiated 0.1%
[source iPerceptions]
• Pop-up invitations can harm conversion
• Visitor-initiated systems have lower response
rate
• Probably skew towards complaints
How to work with feedback surveys
Where and how to ask
Horror show invitation HT
@JamesGurd
29. @timlb
Everywhere (before conversion)
!
For
● Widest range of insight
● Open to non-converters
!
Against
• Low response
• Active invitation 2% - 5%
• Visitor-initiated 0.1%
[source iPerceptions]
• Pop-up invitations can harm conversion
• Visitor-initiated systems have lower response
rate
• Probably skew towards complaints
How to work with feedback surveys
Where and how to ask
30. @timlb
What is the purpose of your visit today?
(Choices + “other” box)
!
Were you able to complete your task?
!
!
Any other comments?
How to work with feedback surveys
What to ask - mobile style
31. @timlb
How to work with feedback surveys
Where and how to ask
After conversion
For
● No risk to conversion
● Simple embedded form
● Can ask more questions
● Less skewed to complaints:
learn what people like
● Surprise: will include comments about previous
non-conversion visits
● High response
• 15% - 35% start survey
• 5% - 10% include text comment
Against
● Only converters
32. @timlb
How to work with feedback surveys
What to ask - post conversion
1 . Net Promoter Question
or
Customer Effort Score
!
2. Were you able to do everything
you wanted to do today?
!
3. Any other comments?
!
(If you’re not storing email address
with survey also ask for email
address if a reply is needed)
Net Promoter or NPS®
Net Promoter, Net Promoter Score, and NPS are trademarks of
Satmetrix Systems, Inc., Bain & Company, Inc., and Fred Reichheld
!
“How likely is it that you would
recommend this site to a friend or
colleague?”
Scale 0-10 where 0 is “very unlikely” and 10 is
“very likely”
!
0-6 are ‘detractors’
7-8 are ‘passives’
9-10 are ‘promoters’
Score = % of ‘promoters’ - % of detractors
!
Customer Effort Score
“How much effort did you personally
have to put forth to handle your
request?”
!
Scale 1-5 where 1 is “very low effort” and 5 is
“very high effort”
33. @timlb
How to work with feedback surveys
The Key Process: free text comments
No good just browsing comments!
!
MUST be systematic
!
● ‘Tag’ according to theme
● Ideally tagging done by same person
● ‘Tagger’ writes a regular report
● Report on trends
● Include ‘verbatim’ sample comments for each
point in report
!
Get comments from the contact centre too!
34. @timlb
Report in a consistent way
Show trends
Include relevant verbatim samples
How to work with feedback surveys
Beware confirmation bias
No process = opinion, not data = no insights
35. @timlb
Voice of Customer:
Not just about fixing bugs
Deeper Understanding people’s language & expectations
Matt Lacey from PRWD in an Econsultancy blog on a/b testing:
“In one example we saw a lift of 4% in sales across the site for a large European retailer,
simply by changing the messages in the USP bar. The key to this test was that the
messages that won had come directly from the mouths of customers, from user research
prior to developing the hypotheses for the test.” http://bit.ly/Superweek2014USB
!
Product Suggestions
People will ask for new products they expected you to stock
Competitor Intelligence
People will compare you to your rivals
36. @timlb
How to work with feedback surveys
Integrate with GA
• Many systems have features for sending data
to GA
• Custom dimensions/metrics
• Events (Values in standard reports)
• Warning: never send Personally Identifiable
Information to GA
Store anonymous survey-id for session-
level analysis…
!
Bonus: integrate with session recording
Suddenly...
...data about people in GA
!
(purpose of visit!)
37. @timlb
Session Recording
Real users!
Lots of them!
!
Vital tip: must be able to search for
recordings containing URL
!
Tip: SessionCam integrates with GA so you can
view sessions from within GA custom report
www.sessioncam.com
38. @timlb
Session Recording
Real users!
Lots of them!
!
Vital tip: must be able to search for
recordings containing URL
!
Tip: SessionCam integrates with GA so you can
view sessions from within GA custom report
www.sessioncam.com
39. @timlb
Scroll maps, mouse and click maps
Aggregate data
Quicker top-level view
‘People’ lost in crowd
!
!
Top: www.clicktale.com
Bottom: www.sessioncam.com
43. @timlb
What else: Usability Tests
Moderated
Recruit real users
Watch them yourself
Ask questions
Un-moderated
Panels
usertesting.com
whatusersdo.com
System only
www.openhallway.com
Photo: webusability.co.uk
Photo: tobii.com
userresearch.blog.gov.uk
www.gov.uk/service-manual/user-researchers
44. @timlb
#cboostdk
If you have addresses for registered users
(Definitely NOT in GA!)
!
Use agency to profile against standard data
!
Often a big surprise:
Large groups of people you would not expect
!
Assumptions about users change
Content of site needs to change
What else: Customer Demographics
45. @timlb
GA: what do we know?
Where people came from
!
Disappearing data:
Not provided
Dark social (Direct)
!
What’s left:
Campaigns
Sample referrers
What they did
!
More and more data:
Events
Enhanced Ecommerce
Custom data
Data widening
!
Who they are?
Demographic clues
!
46. @timlb
GA: Make Campaigns Work Harder
Campaigns are aimed at people
!
Make sure GA has data on ‘who’
Tagging Usually structured to suit channel management
Try to include data to segment visitors by intent
Stage in conversion journey
Kind of interest
Don’t forget promo/transactional email + paid/unpaid social
!
Paid Search
Good: Matched Search Query can be goldmine for intent
Not so good:
Although technically easier to structure around visitors
In reality has to be structured around ROI
47. @timlb
GA: Intent - Carmen’s Great Guide
Carmen Mardiros
Superweek 2014:
!
Visitor Intent: Smart clues for
understanding customer journeys
http://bit.ly/
SuperweekCarmenMardiros
48. @timlb
GA: Behaviour - what people did
(Flow reports)
(Navigation summary)
Content Grouping
Site search
Landing pages
Second page viewed
49. @timlb
GA: Segments and Micro-conversion
User segments?
Sequence segments
!
Micro-conversion
goals
!
Learn about ‘upper funnel’ research visits
!
Use with reverse goal report
!
Goals for content areas
http://bit.ly/SuperweekCarmenMardiros
51. @timlb
View of the Internet from my site
Rest of the Internet
Facebook
Google
!
Products
Checkout
Home Page
http://en.wikipedia.org/wiki/View_of_the_World_from_9th_Avenue
52. @timlb
View of the Internet from my site
Rest of the Internet
Facebook
Google
!
Products
Checkout
Home Page
http://en.wikipedia.org/wiki/View_of_the_World_from_9th_Avenue
Mobile??
53. @timlb
If you can do only one thing…
!
Adopt the strategy: think about people
!
Then start with one tactic: Ask people -
Put a feedback survey on your site
Final Key Point: Just one thing
Your
Agenda
Visitor Agenda
54. @timlb
If you can do only one thing…
!
Adopt the strategy: think about people
!
Then start with one tactic: Ask people -
Put a feedback survey on your site
Final Key Point: Just one thing
Your
Agenda
Visitor Agenda
55. @timlb
Thank you!
Pre-conversion
Usual suspects:
4Q - www.iperceptions.com/en/plans-and-
pricing/free
qualaroo.com
www.kampyle.com
!
Some others:
fluidsurveys.com
webengage.com
!
Beta/new:
www.hotjar.com (session recording too)
survicate.com
http://www.mayankho.com/
@martijnvv prototype at
Measurecamp London
!
Enterprise level
www.iperceptions.com
www.servicetick.com
Post-conversion
Usual suspects:
www.surveygizmo.com
www.surveymonkey.com
!
Some others:
fluidsurveys.com
webengage.com
!
Beta/new:
http://getsatalytics.com/
@DanBarker prototype at
Measurecamp London
!
General form builders:
http://www.wufoo.com
http://www.formstack.com
!
Enterprise level
www.iperceptions.com
www.servicetick.com
!
!
!
Feedback Systems (just some of them)