Extending IA past taxonomies - a case study of Debenhams Retail IA
1. Extending IA past taxonomies - a
case study of Debenhams Retail IA
Mags Hanley
Annabelle Reid
UX Australia, Brisbane
2. Introductions
Mags Hanley
• 20 years a Cybrarian,
Project Manager,
Information Architect,
Product Manager & Teacher
• Provided IA and UX
expertise to the project
Annabelle Reid
• Product Search and Image
Manager at Debenhams
• Responsible search and
navigation of transactional
products for
debenhams.com
3. What is Information Architecture?
IA is the art and science of organising and
labelling web sites, intranets, online
communities and software to support usability
and findability.*
IA Institute ‘What is IA?’
http://iainstitute.org/en/learn/resources/what_is_ia.php
4. Progression of information
architecture
1998
Site maps,
wireframes,
search and
navigation
2001
Taxonomies
and metadata
2004
Folksonomies
and tagging
2010
Domain
modelling and
semantic web
2015
The whole
thing + SEO
5. Most UXers understanding of IA
1998
Site maps,
wireframes,
search and
navigation
2001
Taxonomies
and metadata
2004
Folksonomies
and tagging
2010
Domain
modelling and
semantic web
2015
The whole
thing + SEO
6. Agenda
Covering
• Debenhams and the IA challenges
• IA model for Debenhams
• Navigation example
• IA and business interactions
• Continuing work
9. UX vs. IA in Debenhams
• User Experience is focused on the design of
the customer experience across the web site
and multi-channels.
• Information architecture (IA) is the structure
(product data, labels and categories) on which
the experience can be built on top.
• It closely aligned with Product Information,
Search and Categories all managed by
Annabelle
10. Little black dresses – Search vs. browse
• 35 ‘little black dresses’ found via search
• 114 ‘Little black dresses’ and 1 ‘Little black dress’ in the
categories
Problem: Inconsistency of categories and not integrating
search and categories together.
11. Where are the mugs?
• One of the top search
terms is ‘Mugs’
• But only 20% of people
search, so where is it
living in the navigation?
• It is under ‘Dinnerware -
loose
Problem: Our categories
names do not reflect the
words our customers use.
12. Bedding – what size are you looking for?
• Most people shop ‘Duvet
covers’ by size.
• Which one of the many
sizes on the site meets
the customers’ needs?
Problem: There is no
standardisation of size in
bedding. ‘Duvet covers and
pillow cases’ have 54
variants of size in the ‘Size’
attribute; many of which are
not sizes.
13. What we did in the project
Project
plan
Current
state
report
User
research &
business
goals
report
Content,
data &
search
report
IA and
conten
t
model
Dresses
categorie
s tested
with
users
Facets for
Dresses
section
Metadata and
URL rules for
SEO
Lo-fi
navigation
and
product
pages
prototype
Full Dresses
IA
Governance
and
operational
model
Next steps
DesignDefine
Project
Kick-off
Current state
analysis
Business
insights
Content
data &
search
User
Research
Develop
Dresses IA
Online User
Testing
Facet
developmen
t
Metadata
& SEO
Lo-fi
prototype
User
Testing
Governance
&
operational
model and
next steps
Develop IA
and content
model
14. IA model for Debenhams.com
Findability of the products
'Aboutness' of the products
Google.com
'RJR John Rocha blue dress'
Debenhams.com
Style Edit >
AW14: 6 Fabulous Trends >
Blue notes
Debenhams.com
Home page >
Dresses >
Brand: RJR John Rocha >
Colour: Blue Facebook.com
Debenhams - the official page
Brand: RJR.John Rocha Colour: Blue Length: Knee length Product: Dress
Print: Floral Sleeve: Long sleeveNeckline: Boat neck
Style: shiftOccasion: Work dress
21. IA and business interactions
Information
architecture
User research
Technology
Product data
SEO
Change
management
Analytics
22. IA and SEO
• Previous state
New categories were sent by
Trading Manager to Digital
Marketing for ‘SEO’ing.
Resulting in no changes to the
categories.
• Contribution to IA
Ability to identify language
patterns from across the the
whole of the internet, not just
the Debenhams web site.
• Changes to working
Instead of throwing the
categories over the wall to
Digital Marketing, work
consistently with IA and
Product team on category
labels.
• Impact
Digital Marketing and SEO
requirements are taken into
consideration as part of the
process.
Information
architecture
User
research
Technology
Product data
SEO
Change
management
Analytics
23. IA and user research
• Previous state
Very little user research
performed to inform the IA
• Contribution to the IA
Makes the IA user centred.
Using multiple research methods
to understanding how people find
products, including:
• Information seeking research
• Card sorting
• Path analysis
• Search log analysis
• Changes to working
Using the usability methods to
inform and break down the
business only considerations for
the IA.
• Impact
We now have methods to use to
design the IA better with user
feedback rather than just opinion
of the business
Information
architecture
User
research
Technology
Product data
SEO
Change
managemen
t
Analytics
24. IA and change management
• Previous state
Many parts of the organisation
felt they owned the IA. This
included Trading, Digital
Marketing, Buying and
Merchandising and Product data.
• Contribution to the IA
All the groups contribute to the
IA, as they represent the parts of
the business. For example B&M
tell us what the new trends are,
so we can describe the products
correctly.
• Changes to working
Ownership of the IA is within the
Product data team and is
empowered to make decisions
about the IA.
All departments can contribute,
but one function is responsible.
Take the departments through
the process, rather than impose it
on them.
• Impact
Changes can be made in
Debenhams rather than paralysis.
Information
architecture
User
research
Technology
Product data
SEO
Change
managemen
t
Analytics
25. IA and Product data
• Previous state
There was a product loading
team managed by Annabelle.
The focus was loading
correctly and secondly
describing it.
• Contribution to IA
IA is both categorisation and
structure of content; so
product data IS the IA of the
Debenhams site.
• Changes to working
Worked with the product data
team to ensure they
understand how to make the
data more user friendly,
correctly described and
reflects the trends in the
products.
• Impact
Better description of content
Information
architecture
User
research
Technology
Product data
SEO
Change
managemen
t
Analytics
26. IA and technology
• Previous state
The technology determined
what could be done in
navigation and search, with
slow releases.
• Contribution the IA
Limits the way that the
navigation works, how search
is tuned and how products can
be found on the site.
• Changes to working
As part of the IA model,
instead of living with the
technology, requirements for
search and navigation are as
important as e-commerce or
design.
• Impact
Let the IA have impact in the
site rather than just changing
labels.
Information
architecture
User
research
Technology
Product data
SEO
Change
managemen
t
Analytics
27. IA and analytics
• Previous state
IA was not measured by
analytics; only ad hoc reports
when someone asked
• Contribution to the IA
Makes the IA accountable.
We used click-through rates on
the navigation to try and
understand the patterns of
use.
• Changes to working
During implementation for
dresses, there was a
realisation that there was no
benchmark for the categories
and therefore knowing what
worked was hard to determine
• Impact
Analytics for benchmarking
and design are baked into the
process.
Information
architecture
User
research
Technology
Product data
SEO
Change
managemen
t
Analytics
28. Continuing work
• Since Mags left in the last 6 months, Debenhams has
made a lot of progress.
• The navigation has changed; made responsive and
updated and continuing to be refined.
• The search engine (FredHopper) is being upgraded
within the next few weeks
• A PIM has been commissioned so that the data is not
enhanced for digital, but created by the Buying and
Merchandising teams
• Changes to the team structure including an IA and
changes to the product team and ways Debenhams’
teams work together