Sensis developed a search capability to drive presence in local search and open up its large business listing database. It selected an open source platform and built a total search solution with relevancy testing, quality management, and continuous deployment. Over time, Sensis evolved from a black box solution to a more transparent and manageable platform through problem identification, gold standard testing, and continuous tuning. Future plans include expanding the query engine and implementing machine learning.
3. • Sensis helps Australians
find, buy and sell
• From print directories to a
cross-platform lead generator
• Sensis publishes over 1.8
Million business listings
• Two of the top 10 visited online
sites in Australia
(WhitePages.com.au and
YellowPages.com.au)
Sensis
4. Business objectives
• Drive presence in the local
search market place
• Open up the largest database of
business listings in Australia
• Reduce the effort required from
local search developers Technology objectives
• Free to use, we are after the • Develop a total search platform
reporting • Relevancy testing as part of the
development lifecycle
• A framework to identify problem
spaces
• Manageable platform
• Continuous deployments
Project background
6. • Support for the search
capability team
• Structured vs non
structured data
• Deterministic vs black
box
• Non propriety code base
• Community backing
Platform selection
7. • A/B testing
• Machine learning
Optimized Lvl 5 • External collaboration
• Multiple contexts
• Online dashboards
• Test environments
Managed Lvl 4 • Dynamic search refinements
• Targets and metrics
• Defined team
• Regular monitoring
Monitored Lvl 3 • Static autosuggest
• Basic linguistics
• Adhoc processes
• Part time team
Adhoc Lvl 2 • Static dictionaries
• Individual led innovation
• No resources
• No reporting
Unmanaged Lvl 1 • Out of the box
features
The Sensis Search capability maturity model
*Courtesy of Pete Crawford & Craig Lonsdale
8. Location
Intent Chronology
• Name
• Type
Social Graph
• Product
• Spatial
Device
Individual
Context is key
9. Business Geo Service
Data
Solr Mashery
Business Name Query
Data Search
MongoDB Handler Service
Index API Publisher
Reporting
Type Query
Service
Handler
Historical
search
Data
Reporting
Events
Ontologies
Our architecture
10. Business Geo Service
Data
Solr Mashery
Business Name Query
Data Search
MongoDB Handler Service
Index API Publisher
Reporting
Type Query
Service
Handler
Historical
search
Data
Reporting
Events
Ontologies
Data staging
11. Business Geo Service
Data
Solr Mashery
Business Name Query
Data Search
MongoDB Handler Service
Index API Publisher
Reporting
Type Query
Service
Handler
Historical
search
Data
Reporting
Events
Ontologies
Search
12. Business Geo Service
Data
Solr Mashery
Business Name Query
Data Search
MongoDB Handler Service
Index API Publisher
Reporting
Type Query
Service
Handler
Historical
search
Data
Reporting
Events
Ontologies
API
13. Business Geo Service
Data
Solr Mashery
Business Name Query
Data Search
MongoDB Handler Service
Index API Publisher
Reporting
Type Query
Service
Handler
Historical
search
Data
Reporting
Events
Ontologies
API proxy
14. • Moved from a black box Yesterday Today Tomorrow
solution to a manageable
platform
• Deliver search improvements
without major code changes
• Understand how results were
calculated
• Identity problems scientifically
• Continuously tune and test
relevance
Evolution of search management
15. Specific gold sets for each
Path Analysis problem space:
used to identify Intent
Spelling & stemming
problems Location
spaces Phrase parsing
Features signed off
“Gold Sets” only when they make
used to define a positive impact to
overall quality quality score
score (TREC)
Problem spaces, quality management & tuning