Client: Emirates Group
Team: 2 UX Researchers of Carnegie Mellon University Innovation Group, 4 Data Engineers of TravelScout start-up
Mentors: Dr. Stuart Evans, Dr. Shantha Mohan
My Role: Collaborated with team members from TravelScout to set a road map for the project, led the visual design component of the project including various iterations of UI prototypes, getting feedback, acted as a liaison between Emirates executives and backend team, generated technical design specifications report for system architecture, conducted 3 rounds of stakeholder interviews and created presentation for final prototype.
Duration: 2 Months- Summer 2017
background
The biggest challenge for airline companies is that they do not have important background information for every individual customer. A sizable percentage of customers make bookings via call centers and travel agents, as opposed to making direct purchases through the airline websites. These third-party vendors, including online travel agents (OTAs) such as booking.com and expedia.com, do not provide customer details to the airline companies. The lack of basic customer information provides for a poor user experience, especially when there are changes to flight schedules or cancellations. Airline company representatives cannot reach out to customers in a timely manner, leaving customers anxious and irritated.
Apart from this, airline companies do not have access to rich consumer data that can be used to improve customer experience, acquire more customers and retain existing customers. The irony is that customer data is available on the internet, social media being one prominent source, but tying this data to passenger information is very challenging for airlines.
process
Initial research
Current data
Several people have their contact information (email IDs) publicly available on the internet. 300 million+ people use the professional network LinkedIn, and several have email IDs displayed on their profiles. Some also have contact information available on Facebook, Twitter and personal websites. Tying contact information of passengers (who book tickets via external agents) to their travel accounts will help airlines keep in touch with passengers.
People are active on social media, and log their interests (travel, food, sports, entertainment, etc.), who their friends and family members are, what locations they have visited, what events they are planning to attend, and so on. This information can be collected and associated with the traveler information airlines already possess - to provide a rich profile of travelers. This information can be used in a variety of ways. For example, it could be used to:
Customize travel for travelers - based on age, health problems, special care for infants or elders, language preferences, seat preferences, add-ons, cuisine and entertainment. Customer service representatives at airline call centers can use this data to offer relevant deals and add-ons to trips. Better customer experience will increase retention of customers, and the Net Promoter Score of airlines.
Help marketing understand what type of users react to what type of campaigns, so they can sharpen their campaigns and acquire more users and sell more to existing users. Marketing analysts can use this data to create marketing campaigns that result in greater return on investment for the company.
Processing several data points is a technically challenging problem which can be addressed by building a platform to do the same in a scalable and cost-efficient manner, we have the opportunity to replace or facilitate data science and engineering teams at large airline companies.
Competition
The travel industry is competitive and highly fragmented, with a number of players addressing niche use cases. Most players have chosen the B2C direction, and there is relatively less competition in the B2B space. OTAs such as Expedia and Priceline, who act as prime flight booking platforms for travelers, possess a lot of information about travelers’ travel preferences and travel history.
There are a lot of “big data analytics” companies - but only a few of them focus on the travel industry. Sqreem DB provides behavioral population data of towns, cities or countries. Dataminr transforms Twitter stream and other public data sets into actionable alerts to clients in the Finance, Public Sector, News, Corporate Security and Crisis Management space.
Big Data for Humans (launched in 2014) offers a data-science-as-a-service product to help the retail and travel sector understand “who their customers are and what they want and identifies how to sell more to them every day”. They focus on automating customer insights for marketing. Enterprises such as AirAsia use their product. This shows that there is a market for analytics products in the travel space.
We see that there is a strong opportunity for us to provide airline companies with missing pieces of crucial information: customer contact details and social data about customers.
Our focus is to uniquely identify and provide a deep understanding of every user. We are not focusing on web analytics to understand customer usage metrics available through existing analytics tools such as Google, Adobe Analytics and Mixpanel.
Target Personas
• Frequent Flyers
• First time traveler
• Business traveler
• Business Frequent Traveler traveling with family
• Rare traveler ( eg. once an year )
conceptualizing the opportunity
Shadow bot - user IDENTIFICATION and tracking
Use Cases
A user logs into Emirates site. Emirates site will insert some cookies into his browser. Shadow bot tracks the user’s activity on Emirates site using first party cookies.
User logs into Facebook. Emirates partners with Facebook to launch campaigns and shadow bot accesses the cookies updated by Facebook or embed some tracking codes in the ads which Emirates shows on Facebook pages and we can track the user’s activity on the Facebook page. Here we are tracking user’s activity using third party cookies or tracking pixels.
User visits Walmart.com. Walmart’s web server inserts some cookies into the user’s browser. But Emirates is not launching any campaign on Walmart.com. So, this cookie’s information is not going to be made available to Emirates. To track the user’s activity on Walmart.com, we should figure out if we can somehow push the web beacons or tracking code onto Walmart.com.
Swarm Intelligence - Platform Architecture
The Swarm Intelligence architecture would process the data from the helix database and the data we gather from the Shadow Bots.
Design decision : NOT to store the data gathered
Pros : Security/Legal constraints are met as we do not store user data without their permission on a third party server.
Data from Emirates : The data provided by Emirates is currently in a Helix Database. The architecture of this database may not scale well to accommodate large number of requests.
There are multiple parts to this system. We will look at each part based on the different scenarios.
Data from social media/web footprint: Since we are gathering data from social media, we have two possibilities with the data gathered
Design decision : NOT to store the data gathered
Pros : Security/Legal constraints are met as we do not store user data without their permission on a third party server.
Cons : Performance (Latency). We need to identify spikes in real time which requires intensive processing, doing it without storage is not going to be ideal.
Design decision : To store the data gathered
Pros : Storing the data will help reduce the latency in processing user’s information.
Cons : User’s social media data is highly critical information and needs to be handled properly. We should explore the legal considerations in depth.
Data from Emirates : The data provided by Emirates is currently in a Helix Database. The architecture of this database may not scale well to accommodate large number of requests.
visualization
A dashboard indicating the key metrics from the Swarm Intelligence is displayed. This is to aid the users understand the swarm activity.
Low Fidelity Mock-ups For Dashboard Displaying Swarm Intelligence Results:
The home page provides sign up information and latest travel packages that Emirates has to offer.
Consumer insights tab provides well documented information about every single customer, which can be used to provide personalized recommendations
The recent events tab entails the global travel trends and results from the platform’s analytics subsystem, which in turn will be helpful in identifying spikes in interests and optimizing pricing and travel offers accordingly.
Tracking tab will provide general demographic information about Emirates Airlines’ rich customer database.
Customer Facing Chat-Bot Assistance Mock-up Using IBM Watson:
The chatbot interface provides bespoke recomendation to customers based on swarm intelligence results.
User Journey Steps:
1. The customer logs in to Emirates Airlines website.
2. The shadow bot harnesses information about the customer from Helix database, social media network and other sources to understand the customer behavior and preferences.
3. The data is processed and intelligent tailor made recommendations are presented to the customer.
4. Customers is immediately satisfied by receiving personalized recommendations that drastically reduces steps required to book a ticket.
5. Furthermore, the chatbot embedded in customer dashboard helps him/her navigate throughout the entire process.
6. Any technical difficulties can be addressed by the chatbot or relayed to a customer service personnel.
7. After completion of the transaction, the customer receives a confirmation and the data about the customer preferences are stored on the database for analysis and future reference.
8. At the airport, the platform provides real time navigation to the customer through a mobile application and provides information such as the shortest possible route to reach at the terminal and the gate and other micro services such as baggage arrival and hotel bookings.
Project outcomes and looking forwards
Personalized offers
An in-depth analysis of data about every customer can be used to generate tailor made offers. This will not only enhance the customer satisfaction but will also increase their loyalty toward the Airlines. Data harnessed from Emirates’ Helix database and social media network can be used to provide a very personalized experience for the customers.
Improving Marketing Efforts
An immediate outcome generated from previous point by collecting detailed data from individual customers and presenting them special offers increases the chances of getting a favorable response which will enable Emirates to measure how customers think and behave for future marketing activities. This can also be used to analyze spikes in customer interests towards a particular destination, say Brazil for upcoming FIFA world cup and generate recommendations accordingly.
Pricing Strategies
By tracking travel demand patterns around the globe, Emirates can identify key customer segments and decide the pricing accordingly. A spike in purchases of travel tickets from place A to B can be reciprocated with a surge in pricing to maximize the revenue.
Increasing Customer Satisfaction
- Real-Time Baggage Status: The swarming based platform can be linked with a mobile application which helps track baggage location and provide pertinent recommendations and directions to the customer. It’s important to note that the popularity of the app would depend on how efficiently it is promoted.
- Wearable Technology & IoT: Internet of Things based smart beacons spread across the airport can be used in sync with the platform and mobile application to assist customers navigate their way to lounges, food and retail and to the boarding gates. In addition to that, customers will be able to save the location of where they park their car and track their children with smart bracelets even when they are flying alone.
- More Insightful Customer Behavior: Speech analytics tools can be used to harness useful information about various nuances in customer traits, which combined with data from various online channels like social media and Emirates database can provide important metrics to guide service personnel to provide the best solution in every scenario.
