AI for enhanced discoverability and user experience in online bookstores - Tech Forum 2024

Join Christian Roy and Jean-Benoît Dumais as they delve into the practical applications of AI technology in the world of online bookstores. This session highlights how AI innovations are being employed on leslibraires.ca to improve SEO, enhance discoverability, and optimize user experience. The session explains how the TAMIS software, which uses natural language processing techniques applied to books, is being used to achieve these results. The presentation offers actionable insights into how AI solves common business challenges in the e-commerce space.

Link to presentation video and transcript: https://bnctechforum.ca/sessions/ai-for-enhanced-discoverability-and-user-experience-in-online-bookstores/

Presented by BookNet Canada on October 3, 2024, with support from the Department of Canadian Heritage.

1. AI for enhanced discoverability and
user experience in online
bookstores
Tech Forum
October 2024

4. 1.About leslibraires.ca
2.Reminders on Metadata, Discoverability and User
Experience
3.Context and Opportunities
4.Practical usage of AI Generated Metadata on the
Web
5.Results and Insights

Questions and discussion
Plan

5. ??????

About leslibraires.ca

6. Cooperative operating under the banner Les libraires
●120 independent bookstores primarily in Quebec, the
Maritimes, Ontario, and Manitoba
●Business model is based on mutualization
●Bookstores can benefit from advanced, costly technology
otherwise inaccessible.
●Other benefits : TV campaigns, FCIB collective membership,
purchasing group with Canada Post to reduce shipping costs,
a key element in competing with online giants.

7. Cooperative operating under the banner Les libraires
●Mutualization encourages collaboration and innovation
●In terms of sustainability, our model promotes the optimization of
existing capacities and reduces duplication or wastage of
resources.
●Since the historic sales peak in 2020-2021 (COVID-19
pandemic), sales have stabilized at a level that is still four to five
times higher than before the crisis.

8. Innovative marketplace for Canadian independent bookstores and
publishers leslibraires.ca
●Collective site to pool a robust technological infrastructure and create fair
business conditions in the face of rising web giants.
●Allows bookstores and publishers to benefit from our investments and
extensive experience to stand out and enhance discoverability by leveraging
accessibility metadata and AI.

9. Occupying an online sales space is not enough !

●According to a More Canada study (2018), changes in how
Canadians discover, purchase, or borrow books are eroding the
presence of Canadian expression in our literary landscape.
●To counter this erosion, independent bookstores can play two
complementary roles: raising awareness among readers and
ensuring access to Canadian books.
●This mission, which is concrete and tangible in a traditional
brick-and-mortar bookstore, becomes a challenge online, in a
dematerialized context.

10. Occupying an online sales space is not enough !

●To make Canadian books more discoverable online, our bookstores
are innovating by using AI to enrich traditional metadata, a task not
performed by Amazon and other international players.
●The project’s discoverability component began with raising
awareness among publishers about the potential of using AI to
improve book discoverability.
●This step was crucial to obtain their consent for using their works'
texts.

11. ⚙

Reminders on Metadata, Discoverability and
User Experience

12. Metadata
Discoverability
User Experience
(UX)

13. ? ? ?
Metadata
Discoverability
User Experience
(UX)

14. ? ? ?
Let’s have a look at the typical building blocks of a
bookselling website

15. Book page with a lot of details

16. Search engine

17. Navigation

18. Marketing and merchandising

19. Personalized recommendations

20. Personalized newsletter

22. Metadata
All building blocks are powered by metadata

23. Together, they create a unique user experience
User Experience
(UX)

24. Some help “external” discoverability

Book pages, marketing and merchandising and navigation help through SEO
Newsletter reignite engagement
(external)
Discoverability

25. Some help “internal” discoverability

Marketing, merchandising and navigation showcase books.
Search helps users who fully or partially know what they want to find.
Recommendations stimulate serenpidity
(internal)
Discoverability

26. Metadata
Discoverability
User Experience
(UX)
The typical building blocks of a bookselling website are powered by
metadata and contribute to discoverability and UX.

Therefore metadata contribute to discoverability and UX.

27. ??????

Context and Opportunities

28. Context

Website redesign

UX research

SEO audit

29. Metadata: ??????
Discoverability: ??????
UX: ??????

31. Discoverability and UX are almost indistinguishable…

… because all those sites use the same metadata, in the
same way!

32. Can we do better?

A better use of the same metadata?
Or of more metadata? Or data, in general?

33. Book (meta)data

●created either on the product or consumer side of
the book chain
●can be more descriptive or more experiential

34. Product Consumer





Descriptive
Experiential
Catalog metadata (title, author, …)

(from publishers)
???
Likes
Ratings
Reviews

(from readers)
Marketing and merchandising
Recommandations
Prizes

(from booksellers, librarians, critics)

35. Opportunity

36. 2019 : Project TAMIS
https://bnctechforum.ca/sessions/project-tamis-using-ai-to-address-the-discoverability-challenge/

37. Named Entities Extraction

38. Places
2147 entities
Events
People
Organisations
Fort Saint-Louis
(Texas)
Manitoba
Chicago
cimetière de
Notre-Dame-de-
Lourdes
Gunfight at the
O.K. Corral
Prehistory
1920’s
Catholic Church
NHL
Association
francophone de
Kamloops
Louis Riel
Oprah Winfrey
Louis XIV
François-Xavier
Matthieu
Crazy Horse (chef
amérindien)

39. Places
2147 entities
Fort Saint-Louis
(Texas)
Manitoba
Chicago
cimetière de
Notre-Dame-de-
Lourdes

40. Product Consumer





Descriptive
Experiential
Catalog metadata (title, author, …)

(from publishers)
Entities

(form the book itself!)
Likes
Ratings
Reviews

(from readers)
Marketing and merchandising
Recommandations
Prizes

(from booksellers, librarians, critics)

41. Objectives

Differentiation

Better SEO for long tail
keywords


Context and
Opportunities

Website redesign

TAMIS
Concerns

Work with publishers

Help, but don’t
replace, humans

??????

42. ??????

Practical usage of AI Generated Metadata on
the Web

44. A consumer application
operated by
leslibraires.ca
Tool for professionals
offered by A10s

45. Named
entities from
TAMIS in
internal
search
engine

46. Entity
related
excerpts
in book
pages

47. Named
entities in
Schema.org
structured
metadata

48. Give access
to the full
TAMIS
database to
booksellers
and
leslibraires.ca
staff

49. ??????

Results and Insights

50. Positioning leslibraires.ca as an innovator, generating
media coverage

(brand awareness and website discoverability)

51. https://ici.radio-canada.ca/nouvelle/2040828/intelligence-artificielle-librairie-recherche-lecteur
https://www.livreshebdo.fr/article/quebec-une-intelligence-artificielle-pour-affiner-la-base-de-donnees-des-librairies
https://ici.radio-canada.ca/ohdio/premiere/emissions/c-est-encore-mieux-l-apres-midi/segments/rattrapage/470163/le-robot-lecteur

52. Search Engine Optimization, especially for long-tail
queries

(Canadian books discoverability)

55. Conversion rate increased by 30 to 45%

* combined results from UX research, website redesign, etc.

56. Stronger Publisher Collaboration

57. Booksellers interest in AI and innovation (through
concrete examples of how it can help their work, not
replace it)

58. Members of the coop understand the impacts

“We appreciate the efforts you put into initiatives and the prospect of new
achievements aimed at supporting and participating in the development of
Quebec's cultural bookstore sector. To maintain the cultural vocation we
cherish, it is evident that you understand the importance of having accessible
tools that allow independent bookstores to compete against large web
companies, among others.”

-Christian Laliberté, President and CEO, and Éléna Laliberté, General Manager of
Librairie Laliberté

59. ??????

Final Thoughts

60. Content discoverability in the web era concerns publishers.

●We developed a Reader Robot capable of extracting the
themes of a book so that it can describe itself beyond the
summary.
●By granting us licenses to generate enriched metadata from
their works, publishers have engaged in a process where AI
is leveraged by and for them and booksellers, rather than to
their detriment.
●To date, over 10,000 Canadian books benefit from this
enrichment, which influences their discoverability and
purchase.

61. The adoption of AI by and for publishers and booksellers

●It has not only impacted book discoverability and marketing
●It has also impacted digital literacy and meet the sector’s
demand for innovation
●It serves as a catalyst for other innovative initiatives and
marks the beginning of a virtuous cycle of technology
adoption.

62. Upcoming phase of innovation and enrichment

●Based on indicators related to
○reading difficulty
○reading intentions
○narrative arc forms
●It will enhance our personalized recommendation tools
●And foster more intuitive and tailored reading experiences

63. ??????

Questions and Discussion

64. ??????
Thanks!

[email protected]
[email protected]