Analytics in Action - Community
1 like39 views
The document summarizes how Airbnb uses data science to scale its business. It discusses Airbnb's two-sided marketplace model and how it processes large amounts of data from user profiles, reviews, and bookings. The data science team at Airbnb uses techniques like predictive modeling, A/B testing, and natural language processing. The team is organized to partner closely with other teams and directly apply insights from data science to influence product decisions and growth. Lessons from Airbnb include empowering data scientists, thinking of data culture holistically, and allowing individual interactions to become more efficient through data.
Analytics in Action - Community
- 1. Analytics in Action Community Management February 2019 http://DSign4.education
- 2. Introduction ©2016 L. SCHLENKER Agenda Introduction Communities Community Management Case Study - AirBnB Lessons Learned
- 3. • What proof do we have today that many organizations have lost their sense of community? • What does the author infer by communityship? • How can we facilitate the transformation or organizations towards communityship? • In what ways can Data Science contiribute to businesses becoming places of engagement? Rebuilding Companies as Communities ? Introduction Mintzberg, H., (2009) Rebuilding Companies as Communities
- 6. ©2006 LHST sarl How does work get done? Networks
- 7. Characteristic Value Degree Centrality Number of links Betweeness Centrality Role of brokerage Closeness Centrality Vector of visibility Network Centralization Centralized vs Decentralized Network Reach Importance of first 3 levels Boundary Spanners Linked to Innovation Peripheral Players Potential Gateways Networks
- 8. The Conversation Prism v2.0 • You are at the center of the prism • The first layer of circles displays the activity of learning and organizing engagement strategies… • The second ring maps specific authorities within an organization to provide a competent and helpful response. • The third ring represents the continual rotation of listening, responding, and learning online and in the real world. The Conversation Prism Technology
- 9. Integrated Community Management Conversations that make sense PEOPLE Dynamic profiles of all your people, with info captured from anywhere. Includes followups & targeting. WEBSITE Multiple page types & user profiles. Build custom responsive designs using NationBuilder Theme Sync. SALES Customizable finance pages with goal tracking, social sharing and personal fundraising COMMUNICATIONS Email & text blasting, free phone number with voicemail, and deeply integrated social media. ©2016 LHST sarl
- 10. A digital strategy rather than a website The Customer Experience WEBSITE MEMBER DATABASE COMMUNICATIONS FINANCES ©2016 LHST sarl
- 11. A 360 degree view of each participant The Customer Journey FINANCES ©2016 LHST sarl PROCESS PROFILE RESULTS
- 12. Secure, hosted website Subscriptions Emailblasting Member database Social media integration Process management Mobile engagement Web pages built for action Dynamic profiles Key components
- 14. • What is the organization’s business model? • Why does the organization focus on data? • Which data science techniques does the organization favor ? • What is the link between data science and decision making? • How is the Data Science team organized? • How does the organization use Data Science to propel growth Case Study Questions Technology
- 15. Data Infrastructure at Airbnb • 2500 Employees with rapid growth – the company has opened a dozen international offices simultaneously while explanding their product, marketing, and customer support teams • The site draws 100M users browsing over 2M listings • 10 million nights booked for more than 25 million people across 192 countries and 34,000 cities • A valuation of $25.5 billion as of June 2015 AirBNB Technology
- 16. 2-sided markets • AirBnB matches people who are looking for accommodation with people who are willing to rent out their place • A two-sided marketplace • Network effects, strong seasonality, infrequent transactions, and long time horizons • A valuation of $25.5 billion as of June 2015 Business Model Technology
- 17. How we scaled Data Science • AirBnB uses data science to understand individual experiences and aggregate those experiences to identify trends • Data reflects a decision made by a person. • Recreate the sequence of events leading up to decisions to identify what customers like and don’t like • They translate the customer’s “voice” into a language more suitable for decision-making Why Data Science ? Technology
- 18. How we scaled Data Science • Predictive Analytics - use insights to influence decisions • The Team begins by scanning the context of the problem, putting together a full synopsis of past research. • That synopsis translates to a plan, which encompasses prioritizing the levers we intend to utilize • As the plan gets underway, the team designs a controlled experiment through which to roll the plan out. • Finally, they measure the results of the experiment, identifying the causal impact of our efforts. • The Data Science team partners directly with engineers, designers, product managers, marketers, and others. Impact on Decision Making Technology
- 19. • The unstructured information about the rooms, room owners, locations of the room is sorted and analysed using Hadoop • AirBnB uses host guest interaction, current events and local market history to provide real time recommendations • AirBnB serves approximately 10 million requests a day and processes one million search queries • This represents 20 TB of data created daily and 1.4 petabytes of archived data Data Sources Technology AirbnB uses R to scale data science
- 20. • Airbnb has invested in building an internal R package called Rbnb • A set of collaborative solutions to common problems, standardizes visual presentations • The package includes more than 60 functions • It is used to move aggregated or filtered data into R, impute missing values, compute year- over-year trends, and perform data aggregations. • The objectives are to solve problems like automating the detection of host preferences and using guest ratings to predict rebooking rates Data Tools Technology AirbnB uses R to scale data science
- 21. • 70 Employees, 30+ Engineers, USD 5 million • 80% are proficient in R - 64% use R as their primary data analysis language), • Airbnb organizes monthly week-long data bootcamps for new hires and current team members Data Science Team Technology Scaling Data Science at AirBnB
- 22. • A/B Testing • Image Recognition and Analysis • Natural Language Processing • Predictive Modeling • Regression Analysis • Collaborative Filtering Data Science TechniquesIntroduction
- 23. • What is the organization’s business model? • 70 employees, 30+ Engineers, USD 5 million • Which data science techniques does the organization favor ? • What is the link between data science and decision making? • How is the Data Science team organized? • How does the organization use Data Science to propel growth Data Science Team Technology AIRBNB: LESSONS ON DIGITAL, START-UPS, BIG DATA
- 24. • Scaling data science is enabling guests and hosts to learn from each other • They think about the culture of data in the company as a whole rather than individual teams. • Individual interactions become more efficient as data scientists are empowered to move more quickly • Empowering teams is about removing the burden of reporting and basic data exploration to focus on impactful work. Lessons learned Technology AIRBNB: LESSONS ON DIGITAL, START-UPS, BIG DATA
- 25. • Libert, K., Your Network's Structure Matters More than It's Size • Manville, B., You Need a Community, Not a Network • Mintzberg, H., (2009), Rebuilding Companies as Communities, HBR • Neuman, R. (2016), How we scaled data science to all sides of Airbnb , Venture Beat • Pugh, K., (2013), Designing Effective Knowledge Networks, Sloan Management Review Bibliography Next Steps
Editor's Notes
- #23: The Data Science team at AirBnB uses A/B testing by exposing the users of their website, to various recommendation and ranking algorithms. The behaviour of the users is then correlated with the actual ratings or reviews they leave, which helps them test the effectiveness of the algorithms. AirBnB does analysis on photos to find out which ones work best for their users, what features in the photos make them most sought after and what kind of photos on the website get more number of clicks To interpret the true feelings of users, AirBnB uses natural language processing technology that analyses the review boards or the messages boards through sentiment analysis Using predictive modelling, AirBnB can create market specific forecast with multiple variables. Data mining at AirBnB helps the hosts to predict the best possible rates for their rentals. AirBnB uses regression analysis technique to find out which features of a particular listing have a major impact on the bookings made. Using collaborative filtering, the users (hosts) and the items (trips) data can be used to understand the preference for items by combining historical ratings through statistical learning from related hosts.