SlideShare a Scribd company logo
Case Study:
Dynamic Pricing in
Ride-Hailing Platforms
Hamed Shams
Product Manager at
Abstract
Over the past decade, ride-hailing platforms revolutionized urban transportation. However,
one of the most significant challenges faced by these platforms is the constantly fluctuating
supply & demand volumes which causes serious issues in dynamically setting prices.
This case study examines surge pricing in ride-hailing platforms, with a focus on UBER's
hexagonal indexing system (H3). The study explores the challenges of implementing this
system in Snapp, particularly H3's inability to adapt to extreme fluctuations. More precisely,
having a prefixed hexagon resolution level for each city is not always optimal, as it would
result in considerable no-decision cases due to the lack of signals.
To address this issue, the author proposes a machine learning based approach and discusses
the impact of launching this solution into production. This product resulted in a 27%
decrease in no-decision cases, showcasing the effectiveness of the approach.
Introduction
Dynamic pricing is the study of finding the optimum price of products / services in a
frequently adjusting manner, which applies to both online services (e.g., e-commerce,
platforms) and brick-and-mortar stores.
Nowadays, technologies have paved the way for setting prices more accurately by
considering many factors, and also much more frequently (almost real-time).
With this specific definition, we can say “dynamic pricing” emerged in the 1980s, driven by
technological innovations. It was pioneered by the airline industry, which used factors like
departure time, destination, and season, to automate their pricing.
Ride-Hailing Platforms
Ride-Hailing Platforms
Passengers’ Expectation
• Lowest Fare
• Lowest Waiting Time
Pricing in Ride-Hailing Platforms
Drivers’ Expectation
• Maximum Earning
The Price is Upfront…
A fare estimation of the ride price before it start.
1. Distance Factor
Price = [ Base Factor
+ ( Price per Km * Km ) ]
e.g in Tehran
Price = [ 4000 + ( 750 * 8.1 ) ] = 10,000
2. Time Factor (ETA)
Price = [ Base Factor
+ ( Price per Km * Km )
+ ( Price per Min * Min ) ]
e.g in Tehran
Price = [ 4000 + ( 350 * 8.1 ) + ( 300 * 30 ) ] = 16,000
How about the Marketplace imbalance?
3. Surge Factor
Price = [ Base Factor
+ ( Price per Km * Km )
+ ( Price per Min * Min ) ] * Surge
E.g., Surge = 80%
Price = [ 4000 + (350 * 8.1) + (300 * 30) ] * 1.8
# Ride Request events
# Price Check events
Demand Rate =
# Accept events
# Ride Request events
Supply Rate =
Measured by:
7:00 7:10 7:20 7:30 7:40 7:50 8:00 8:10
Time Boundaries
Space Boundaries
The procedure of ordering a ride is a series of location-based actions over a
period of time. Therefore, Supply and Demand conversion rates need to be
defined over a specific unit of space and time.
To be more accurate, all the events need to be collected and processed with
respect to a predefined space and time unit; Otherwise, no conversion rate
can be calculated. Without a clear definition of space/time, nothing can be
interpreted nor compared.
Time and Space Boundaries
Time and Space Boundaries:
“No-Decision” Cases
A minimum number of “price check” events is required within a predefined
time/space unit, in order to consider the conversion rates coming from those
price checks as valid.
Therefore, whenever the number of available events within a time/space
boundary is lower than this minimum threshold, we consider it a “no-decision”
case, because we can not make any decision for adjusting the price.
Time and Space Boundaries:
How to Define?
The behaviour of supply and demand is directly tied to the number of
available drivers and passengers within a given area, which of course changes
over time, based on many factors, such as traffic patterns.
One of these two dimensions must be tightly defined (as a hard constraint) so
we can start playing with the other in a more dynamic manner, and hopefully,
achieve an optimum solution ultimately. By considering a fixed time window
(say 10-minute intervals) we can start defining the units of space; in other
words, indexing the geospatial data.
Geospatial Indexing in Surge Pricing
1. Defining some Districts Manually
2. Hexagon Hierarchical Indexing
Problems with Manually Defined Polygons
1. High Variations within each Polygon
2. Phantom Demand
1. High Variations within each Polygon
Imagine a football match is happening at the Azadi Stadium…
Azadi Stadium
Chitgar District
1. High Variations within each Polygon
increase in the demand of ”Chitgar” increase in the price of “Chitgar”
Azadi Stadium
Chitgar District
1. High Variations within each Polygon
Even for riders on the low-demand areas of this district, which is not fare.
Azadi Stadium
Chitgar District
2. Phantom Demand
Azadi Stadium
Imagine a driver realize the demand of “Chitgar” district has raised (surged)
Chitgar District
2. Phantom Demand
But, there may be no ride request since the real demand is on the other side
Azadi Stadium
Chitgar District
Solution:
A more granular and dynamic approach
https://www.uber.com/blog/h3
Advantages of Hexagonal Indexing
1. First and foremost, it is not manual!
2. More granularity in supply-demand (address problem #1)
3. Smooth gradients (address problem #2)
4. Symmetric shapes and equal partitions
5. Equal distance from the neighbors
Hexagon Resolution Level
H3 supports 16 different diameters. Each finer resolution, has cells with one
seventh the area of the coarser resolution. Hexagons cannot be perfectly
subdivided into seven hexagons, so the finer cells are only approximately
contained within a parent cell.
Hexagon Resolution Level
Critical Issue:
a single resolution can not work everywhere everytime
The behaviour of supply-demand is extremely fluctuating, which leads to
sudden rises and falls in the “price check” events. Therefore, by using a
specific (fixed) resolution level for each city, two things would happen:
• Hexagons are too small, so they rarely meet the predefined min threshold,
leading to no-decision cases.
• Hexagons are too big, resulting in the similar scenarios with the Manually
Defined Polygons. i.e., lack of granularity and smoothness.
H3 addressed this issue with the “K-Ring” function…
If the events within a hexagon does not meet the min ”price check” threshold
L1
H3’s Geo Expansion Function: “K-Ring”
Then, the hexagon will be expanded to its neighbors (L2)
L1
L2
L2
L2
L2
L2
L2
L1
K-Ring-1
H3’s Geo Expansion Function: “K-Ring”
L1
L2
L2
L2
L2
L2
L2
L1
L2
L2
L2
L2
L2
L2
L3
L3
L3
L3
L3
L3
L3
L3
L3
L3
L3
L3
L1
K-Ring-1 K-Ring-2
If the events are not enough yet, then it will get expanded another level (L3)
H3’s Geo Expansion Function: “K-Ring”
Problem Definition
While the Geo Expansion (K-Ring) approach has proven effective in mitigating
no-decision cases, our experiments have shown that it is insufficient to
handle extreme fluctuations. This often results in outdated surge factors and
unadjusted prices, causing an unbalanced marketplace.
The issue is particularly pronounced during rush hours in commercial regions,
where the system struggles to keep up with the influx of incoming "price
check" signals. Moreover, manually adjusting resolution levels for each city is
an impractical and potentially biased approach.
Therefore, there is a need for a more robust and automated solution that can
adapt to extreme fluctuations in a timely and effective manner.
Proposed Solution
To address the issues of outdated surge factors and unadjusted prices caused
by extreme fluctuations, an automated solution is required to minimize no-
decision cases and accelerate price adaptation. This can be achieved by
dynamically adjusting the resolution level in each city, based on the estimated
number of "price check" events in the next time interval.
To accomplish this, we developed a time series forecasting approach to
predict the number of "price check" events based on space and time. For
example, we can forecast the number of price check events in Tehran within
the next hour. To maintain confidentiality, we have omitted implementation
details from this document.
Business Impact
• About 27% decrease in the no-decision cases overall. The reduction rate
was much more considerable during midnight and rush hours.
• Resulting in a 6% increase in drivers’ earnings while improving rider’s
conversion. In other words, a more reasonable and fair pricing.

More Related Content

What's hot

Ant colony optimization
Ant colony optimizationAnt colony optimization
Ant colony optimizationAbdul Rahman
 
Ant Colony Optimization presentation
Ant Colony Optimization presentationAnt Colony Optimization presentation
Ant Colony Optimization presentationPartha Das
 
Bus Tracking Application in Android
Bus Tracking Application in AndroidBus Tracking Application in Android
Bus Tracking Application in Android
Abhishek Singh
 
Simulated annealing.ppt
Simulated annealing.pptSimulated annealing.ppt
Simulated annealing.pptKaal Nath
 
Traffic signal control management based on integrating GIS and WSN technology
Traffic signal control management based on integrating GIS and WSN technologyTraffic signal control management based on integrating GIS and WSN technology
Traffic signal control management based on integrating GIS and WSN technology
krushna kotgire
 
automatic vehicle location
automatic vehicle locationautomatic vehicle location
automatic vehicle locationAkhil Kumar
 
Particle filter
Particle filterParticle filter
Particle filter
Mohammad Reza Jabbari
 
Ant colony optimization
Ant colony optimizationAnt colony optimization
Ant colony optimization
Suman Chatterjee
 
Best School Bus Tracking System
Best School Bus Tracking SystemBest School Bus Tracking System
Best School Bus Tracking System
jananya213
 
Practical Swarm Optimization (PSO)
Practical Swarm Optimization (PSO)Practical Swarm Optimization (PSO)
Practical Swarm Optimization (PSO)
khashayar Danesh Narooei
 
Kroger Store Simulation Using Arena
Kroger Store Simulation Using ArenaKroger Store Simulation Using Arena
Kroger Store Simulation Using Arena
Dhivya Rajprasad
 
Resilient Design Using Queue Theory
Resilient Design Using Queue TheoryResilient Design Using Queue Theory
Resilient Design Using Queue Theory
ScyllaDB
 
Deone pranilfinalreport
Deone pranilfinalreportDeone pranilfinalreport
Deone pranilfinalreport
Pranil Deone
 
Simulated annealing -a informative approach
Simulated annealing -a informative approachSimulated annealing -a informative approach
Simulated annealing -a informative approach
Ranak Ghosh
 
Reinforcement learning
Reinforcement learningReinforcement learning
Reinforcement learning
DongHyun Kwak
 
Particle swarm optimization
Particle swarm optimizationParticle swarm optimization
Particle swarm optimizationHanya Mohammed
 
Metaheuristics
MetaheuristicsMetaheuristics
Metaheuristics
CherifRehouma
 
Swarm intelligence
Swarm intelligenceSwarm intelligence
Swarm intelligence
Velmurugan Sivaraman
 
Ant Colony Optimization - ACO
Ant Colony Optimization - ACOAnt Colony Optimization - ACO
Ant Colony Optimization - ACO
Mohamed Talaat
 

What's hot (20)

Ant colony optimization
Ant colony optimizationAnt colony optimization
Ant colony optimization
 
Ant Colony Optimization presentation
Ant Colony Optimization presentationAnt Colony Optimization presentation
Ant Colony Optimization presentation
 
Bus Tracking Application in Android
Bus Tracking Application in AndroidBus Tracking Application in Android
Bus Tracking Application in Android
 
Simulated annealing.ppt
Simulated annealing.pptSimulated annealing.ppt
Simulated annealing.ppt
 
Traffic signal control management based on integrating GIS and WSN technology
Traffic signal control management based on integrating GIS and WSN technologyTraffic signal control management based on integrating GIS and WSN technology
Traffic signal control management based on integrating GIS and WSN technology
 
automatic vehicle location
automatic vehicle locationautomatic vehicle location
automatic vehicle location
 
Particle filter
Particle filterParticle filter
Particle filter
 
Ant colony optimization
Ant colony optimizationAnt colony optimization
Ant colony optimization
 
Best School Bus Tracking System
Best School Bus Tracking SystemBest School Bus Tracking System
Best School Bus Tracking System
 
Practical Swarm Optimization (PSO)
Practical Swarm Optimization (PSO)Practical Swarm Optimization (PSO)
Practical Swarm Optimization (PSO)
 
Kroger Store Simulation Using Arena
Kroger Store Simulation Using ArenaKroger Store Simulation Using Arena
Kroger Store Simulation Using Arena
 
Final project
Final projectFinal project
Final project
 
Resilient Design Using Queue Theory
Resilient Design Using Queue TheoryResilient Design Using Queue Theory
Resilient Design Using Queue Theory
 
Deone pranilfinalreport
Deone pranilfinalreportDeone pranilfinalreport
Deone pranilfinalreport
 
Simulated annealing -a informative approach
Simulated annealing -a informative approachSimulated annealing -a informative approach
Simulated annealing -a informative approach
 
Reinforcement learning
Reinforcement learningReinforcement learning
Reinforcement learning
 
Particle swarm optimization
Particle swarm optimizationParticle swarm optimization
Particle swarm optimization
 
Metaheuristics
MetaheuristicsMetaheuristics
Metaheuristics
 
Swarm intelligence
Swarm intelligenceSwarm intelligence
Swarm intelligence
 
Ant Colony Optimization - ACO
Ant Colony Optimization - ACOAnt Colony Optimization - ACO
Ant Colony Optimization - ACO
 

Similar to Dynamic Pricing in Ride-Hailing Platforms

[215]streetwise machine learning for painless parking
[215]streetwise machine learning for painless parking[215]streetwise machine learning for painless parking
[215]streetwise machine learning for painless parking
NAVER D2
 
Applications Of Math In Real Life And Business
Applications Of Math In Real Life And BusinessApplications Of Math In Real Life And Business
Applications Of Math In Real Life And Business
Muhammad Ahmad Badar
 
Demand Responsive Transit Brokerages
Demand Responsive Transit BrokeragesDemand Responsive Transit Brokerages
Demand Responsive Transit Brokerages
Texxi Global
 
Comparison Study of Multiple Traveling Salesmen Problem using Genetic Algorithm
Comparison Study of Multiple Traveling Salesmen Problem using Genetic AlgorithmComparison Study of Multiple Traveling Salesmen Problem using Genetic Algorithm
Comparison Study of Multiple Traveling Salesmen Problem using Genetic Algorithm
IOSR Journals
 
Simulation of real estate price environment
Simulation of real estate price environmentSimulation of real estate price environment
Simulation of real estate price environment
Sohin Shah
 
Autonomous Traffic Signal Control using Decision Tree
Autonomous Traffic Signal Control using Decision Tree Autonomous Traffic Signal Control using Decision Tree
Autonomous Traffic Signal Control using Decision Tree
IJECEIAES
 
Assessing the Impacts of Uncertainty Propagation to System Requirements by Ev...
Assessing the Impacts of Uncertainty Propagation to System Requirements by Ev...Assessing the Impacts of Uncertainty Propagation to System Requirements by Ev...
Assessing the Impacts of Uncertainty Propagation to System Requirements by Ev...
Alejandro Salado
 
Flink Forward Berlin 2018: Amey Chaugule - "Threading Needles in a Haystack: ...
Flink Forward Berlin 2018: Amey Chaugule - "Threading Needles in a Haystack: ...Flink Forward Berlin 2018: Amey Chaugule - "Threading Needles in a Haystack: ...
Flink Forward Berlin 2018: Amey Chaugule - "Threading Needles in a Haystack: ...
Flink Forward
 
Pricing and commissions Webinar English
Pricing and commissions Webinar EnglishPricing and commissions Webinar English
Pricing and commissions Webinar EnglishCamilo Tellez
 
Site suitability analysis for constructing New ATM in Margao , Goa
Site suitability analysis for constructing New ATM in Margao , GoaSite suitability analysis for constructing New ATM in Margao , Goa
Site suitability analysis for constructing New ATM in Margao , Goasuyog patwardhan
 
Sessionizing Uber Trips in Realtime - Flink Forward '18, Berlin
Sessionizing Uber Trips in Realtime  - Flink Forward '18, BerlinSessionizing Uber Trips in Realtime  - Flink Forward '18, Berlin
Sessionizing Uber Trips in Realtime - Flink Forward '18, Berlin
Amey Chaugule
 
When and where are bus express services justified?
When and where are bus express services justified?When and where are bus express services justified?
When and where are bus express services justified?
BRTCoE
 
Streaming Processing in Uber Marketplace for Kafka Summit 2016
Streaming Processing in Uber Marketplace for Kafka Summit 2016Streaming Processing in Uber Marketplace for Kafka Summit 2016
Streaming Processing in Uber Marketplace for Kafka Summit 2016
Danny Yuan
 
C ost+behaviour+estimation
C ost+behaviour+estimationC ost+behaviour+estimation
C ost+behaviour+estimationKhalid Aziz
 
IRJET- Congestion Reducing System through Sensors, Image Processors and Vanet...
IRJET- Congestion Reducing System through Sensors, Image Processors and Vanet...IRJET- Congestion Reducing System through Sensors, Image Processors and Vanet...
IRJET- Congestion Reducing System through Sensors, Image Processors and Vanet...
IRJET Journal
 
Mobility Pricing: How to Harness Mobility Pricing to Reduce Congestion, Promo...
Mobility Pricing: How to Harness Mobility Pricing to Reduce Congestion, Promo...Mobility Pricing: How to Harness Mobility Pricing to Reduce Congestion, Promo...
Mobility Pricing: How to Harness Mobility Pricing to Reduce Congestion, Promo...
WSP
 
A COMPARATIVE STUDY OF DIFFERENT INTEGRATED MULTIPLE CRITERIA DECISION MAKING...
A COMPARATIVE STUDY OF DIFFERENT INTEGRATED MULTIPLE CRITERIA DECISION MAKING...A COMPARATIVE STUDY OF DIFFERENT INTEGRATED MULTIPLE CRITERIA DECISION MAKING...
A COMPARATIVE STUDY OF DIFFERENT INTEGRATED MULTIPLE CRITERIA DECISION MAKING...
Shankha Goswami
 

Similar to Dynamic Pricing in Ride-Hailing Platforms (20)

[215]streetwise machine learning for painless parking
[215]streetwise machine learning for painless parking[215]streetwise machine learning for painless parking
[215]streetwise machine learning for painless parking
 
CabModelWriteup
CabModelWriteupCabModelWriteup
CabModelWriteup
 
HackerOne_Report
HackerOne_ReportHackerOne_Report
HackerOne_Report
 
Applications Of Math In Real Life And Business
Applications Of Math In Real Life And BusinessApplications Of Math In Real Life And Business
Applications Of Math In Real Life And Business
 
Demand Responsive Transit Brokerages
Demand Responsive Transit BrokeragesDemand Responsive Transit Brokerages
Demand Responsive Transit Brokerages
 
Comparison Study of Multiple Traveling Salesmen Problem using Genetic Algorithm
Comparison Study of Multiple Traveling Salesmen Problem using Genetic AlgorithmComparison Study of Multiple Traveling Salesmen Problem using Genetic Algorithm
Comparison Study of Multiple Traveling Salesmen Problem using Genetic Algorithm
 
Simulation of real estate price environment
Simulation of real estate price environmentSimulation of real estate price environment
Simulation of real estate price environment
 
Autonomous Traffic Signal Control using Decision Tree
Autonomous Traffic Signal Control using Decision Tree Autonomous Traffic Signal Control using Decision Tree
Autonomous Traffic Signal Control using Decision Tree
 
Assessing the Impacts of Uncertainty Propagation to System Requirements by Ev...
Assessing the Impacts of Uncertainty Propagation to System Requirements by Ev...Assessing the Impacts of Uncertainty Propagation to System Requirements by Ev...
Assessing the Impacts of Uncertainty Propagation to System Requirements by Ev...
 
Flink Forward Berlin 2018: Amey Chaugule - "Threading Needles in a Haystack: ...
Flink Forward Berlin 2018: Amey Chaugule - "Threading Needles in a Haystack: ...Flink Forward Berlin 2018: Amey Chaugule - "Threading Needles in a Haystack: ...
Flink Forward Berlin 2018: Amey Chaugule - "Threading Needles in a Haystack: ...
 
Pricing and commissions Webinar English
Pricing and commissions Webinar EnglishPricing and commissions Webinar English
Pricing and commissions Webinar English
 
Site suitability analysis for constructing New ATM in Margao , Goa
Site suitability analysis for constructing New ATM in Margao , GoaSite suitability analysis for constructing New ATM in Margao , Goa
Site suitability analysis for constructing New ATM in Margao , Goa
 
Sessionizing Uber Trips in Realtime - Flink Forward '18, Berlin
Sessionizing Uber Trips in Realtime  - Flink Forward '18, BerlinSessionizing Uber Trips in Realtime  - Flink Forward '18, Berlin
Sessionizing Uber Trips in Realtime - Flink Forward '18, Berlin
 
When and where are bus express services justified?
When and where are bus express services justified?When and where are bus express services justified?
When and where are bus express services justified?
 
Streaming Processing in Uber Marketplace for Kafka Summit 2016
Streaming Processing in Uber Marketplace for Kafka Summit 2016Streaming Processing in Uber Marketplace for Kafka Summit 2016
Streaming Processing in Uber Marketplace for Kafka Summit 2016
 
C ost+behaviour+estimation
C ost+behaviour+estimationC ost+behaviour+estimation
C ost+behaviour+estimation
 
facility
facilityfacility
facility
 
IRJET- Congestion Reducing System through Sensors, Image Processors and Vanet...
IRJET- Congestion Reducing System through Sensors, Image Processors and Vanet...IRJET- Congestion Reducing System through Sensors, Image Processors and Vanet...
IRJET- Congestion Reducing System through Sensors, Image Processors and Vanet...
 
Mobility Pricing: How to Harness Mobility Pricing to Reduce Congestion, Promo...
Mobility Pricing: How to Harness Mobility Pricing to Reduce Congestion, Promo...Mobility Pricing: How to Harness Mobility Pricing to Reduce Congestion, Promo...
Mobility Pricing: How to Harness Mobility Pricing to Reduce Congestion, Promo...
 
A COMPARATIVE STUDY OF DIFFERENT INTEGRATED MULTIPLE CRITERIA DECISION MAKING...
A COMPARATIVE STUDY OF DIFFERENT INTEGRATED MULTIPLE CRITERIA DECISION MAKING...A COMPARATIVE STUDY OF DIFFERENT INTEGRATED MULTIPLE CRITERIA DECISION MAKING...
A COMPARATIVE STUDY OF DIFFERENT INTEGRATED MULTIPLE CRITERIA DECISION MAKING...
 

Recently uploaded

An introduction to the cryptocurrency investment platform Binance Savings.
An introduction to the cryptocurrency investment platform Binance Savings.An introduction to the cryptocurrency investment platform Binance Savings.
An introduction to the cryptocurrency investment platform Binance Savings.
Any kyc Account
 
Mastering B2B Payments Webinar from BlueSnap
Mastering B2B Payments Webinar from BlueSnapMastering B2B Payments Webinar from BlueSnap
Mastering B2B Payments Webinar from BlueSnap
Norma Mushkat Gaffin
 
The Influence of Marketing Strategy and Market Competition on Business Perfor...
The Influence of Marketing Strategy and Market Competition on Business Perfor...The Influence of Marketing Strategy and Market Competition on Business Perfor...
The Influence of Marketing Strategy and Market Competition on Business Perfor...
Adam Smith
 
Tata Group Dials Taiwan for Its Chipmaking Ambition in Gujarat’s Dholera
Tata Group Dials Taiwan for Its Chipmaking Ambition in Gujarat’s DholeraTata Group Dials Taiwan for Its Chipmaking Ambition in Gujarat’s Dholera
Tata Group Dials Taiwan for Its Chipmaking Ambition in Gujarat’s Dholera
Avirahi City Dholera
 
Organizational Change Leadership Agile Tour Geneve 2024
Organizational Change Leadership Agile Tour Geneve 2024Organizational Change Leadership Agile Tour Geneve 2024
Organizational Change Leadership Agile Tour Geneve 2024
Kirill Klimov
 
Auditing study material for b.com final year students
Auditing study material for b.com final year  studentsAuditing study material for b.com final year  students
Auditing study material for b.com final year students
narasimhamurthyh4
 
Company Valuation webinar series - Tuesday, 4 June 2024
Company Valuation webinar series - Tuesday, 4 June 2024Company Valuation webinar series - Tuesday, 4 June 2024
Company Valuation webinar series - Tuesday, 4 June 2024
FelixPerez547899
 
Observation Lab PowerPoint Assignment for TEM 431
Observation Lab PowerPoint Assignment for TEM 431Observation Lab PowerPoint Assignment for TEM 431
Observation Lab PowerPoint Assignment for TEM 431
ecamare2
 
ModelingMarketingStrategiesMKS.CollumbiaUniversitypdf
ModelingMarketingStrategiesMKS.CollumbiaUniversitypdfModelingMarketingStrategiesMKS.CollumbiaUniversitypdf
ModelingMarketingStrategiesMKS.CollumbiaUniversitypdf
fisherameliaisabella
 
The Influence of Marketing Strategy and Market Competition on Business Perfor...
The Influence of Marketing Strategy and Market Competition on Business Perfor...The Influence of Marketing Strategy and Market Competition on Business Perfor...
The Influence of Marketing Strategy and Market Competition on Business Perfor...
Adam Smith
 
Premium MEAN Stack Development Solutions for Modern Businesses
Premium MEAN Stack Development Solutions for Modern BusinessesPremium MEAN Stack Development Solutions for Modern Businesses
Premium MEAN Stack Development Solutions for Modern Businesses
SynapseIndia
 
一比一原版加拿大渥太华大学毕业证(uottawa毕业证书)如何办理
一比一原版加拿大渥太华大学毕业证(uottawa毕业证书)如何办理一比一原版加拿大渥太华大学毕业证(uottawa毕业证书)如何办理
一比一原版加拿大渥太华大学毕业证(uottawa毕业证书)如何办理
taqyed
 
Digital Transformation and IT Strategy Toolkit and Templates
Digital Transformation and IT Strategy Toolkit and TemplatesDigital Transformation and IT Strategy Toolkit and Templates
Digital Transformation and IT Strategy Toolkit and Templates
Aurelien Domont, MBA
 
Event Report - SAP Sapphire 2024 Orlando - lots of innovation and old challenges
Event Report - SAP Sapphire 2024 Orlando - lots of innovation and old challengesEvent Report - SAP Sapphire 2024 Orlando - lots of innovation and old challenges
Event Report - SAP Sapphire 2024 Orlando - lots of innovation and old challenges
Holger Mueller
 
Top mailing list providers in the USA.pptx
Top mailing list providers in the USA.pptxTop mailing list providers in the USA.pptx
Top mailing list providers in the USA.pptx
JeremyPeirce1
 
Brand Analysis for an artist named Struan
Brand Analysis for an artist named StruanBrand Analysis for an artist named Struan
Brand Analysis for an artist named Struan
sarahvanessa51503
 
FIA officials brutally tortured innocent and snatched 200 Bitcoins of worth 4...
FIA officials brutally tortured innocent and snatched 200 Bitcoins of worth 4...FIA officials brutally tortured innocent and snatched 200 Bitcoins of worth 4...
FIA officials brutally tortured innocent and snatched 200 Bitcoins of worth 4...
jamalseoexpert1978
 
Exploring Patterns of Connection with Social Dreaming
Exploring Patterns of Connection with Social DreamingExploring Patterns of Connection with Social Dreaming
Exploring Patterns of Connection with Social Dreaming
Nicola Wreford-Howard
 
Call 8867766396 Satta Matka Dpboss Matka Guessing Satta batta Matka 420 Satta...
Call 8867766396 Satta Matka Dpboss Matka Guessing Satta batta Matka 420 Satta...Call 8867766396 Satta Matka Dpboss Matka Guessing Satta batta Matka 420 Satta...
Call 8867766396 Satta Matka Dpboss Matka Guessing Satta batta Matka 420 Satta...
bosssp10
 
-- June 2024 is National Volunteer Month --
-- June 2024 is National Volunteer Month ---- June 2024 is National Volunteer Month --
-- June 2024 is National Volunteer Month --
NZSG
 

Recently uploaded (20)

An introduction to the cryptocurrency investment platform Binance Savings.
An introduction to the cryptocurrency investment platform Binance Savings.An introduction to the cryptocurrency investment platform Binance Savings.
An introduction to the cryptocurrency investment platform Binance Savings.
 
Mastering B2B Payments Webinar from BlueSnap
Mastering B2B Payments Webinar from BlueSnapMastering B2B Payments Webinar from BlueSnap
Mastering B2B Payments Webinar from BlueSnap
 
The Influence of Marketing Strategy and Market Competition on Business Perfor...
The Influence of Marketing Strategy and Market Competition on Business Perfor...The Influence of Marketing Strategy and Market Competition on Business Perfor...
The Influence of Marketing Strategy and Market Competition on Business Perfor...
 
Tata Group Dials Taiwan for Its Chipmaking Ambition in Gujarat’s Dholera
Tata Group Dials Taiwan for Its Chipmaking Ambition in Gujarat’s DholeraTata Group Dials Taiwan for Its Chipmaking Ambition in Gujarat’s Dholera
Tata Group Dials Taiwan for Its Chipmaking Ambition in Gujarat’s Dholera
 
Organizational Change Leadership Agile Tour Geneve 2024
Organizational Change Leadership Agile Tour Geneve 2024Organizational Change Leadership Agile Tour Geneve 2024
Organizational Change Leadership Agile Tour Geneve 2024
 
Auditing study material for b.com final year students
Auditing study material for b.com final year  studentsAuditing study material for b.com final year  students
Auditing study material for b.com final year students
 
Company Valuation webinar series - Tuesday, 4 June 2024
Company Valuation webinar series - Tuesday, 4 June 2024Company Valuation webinar series - Tuesday, 4 June 2024
Company Valuation webinar series - Tuesday, 4 June 2024
 
Observation Lab PowerPoint Assignment for TEM 431
Observation Lab PowerPoint Assignment for TEM 431Observation Lab PowerPoint Assignment for TEM 431
Observation Lab PowerPoint Assignment for TEM 431
 
ModelingMarketingStrategiesMKS.CollumbiaUniversitypdf
ModelingMarketingStrategiesMKS.CollumbiaUniversitypdfModelingMarketingStrategiesMKS.CollumbiaUniversitypdf
ModelingMarketingStrategiesMKS.CollumbiaUniversitypdf
 
The Influence of Marketing Strategy and Market Competition on Business Perfor...
The Influence of Marketing Strategy and Market Competition on Business Perfor...The Influence of Marketing Strategy and Market Competition on Business Perfor...
The Influence of Marketing Strategy and Market Competition on Business Perfor...
 
Premium MEAN Stack Development Solutions for Modern Businesses
Premium MEAN Stack Development Solutions for Modern BusinessesPremium MEAN Stack Development Solutions for Modern Businesses
Premium MEAN Stack Development Solutions for Modern Businesses
 
一比一原版加拿大渥太华大学毕业证(uottawa毕业证书)如何办理
一比一原版加拿大渥太华大学毕业证(uottawa毕业证书)如何办理一比一原版加拿大渥太华大学毕业证(uottawa毕业证书)如何办理
一比一原版加拿大渥太华大学毕业证(uottawa毕业证书)如何办理
 
Digital Transformation and IT Strategy Toolkit and Templates
Digital Transformation and IT Strategy Toolkit and TemplatesDigital Transformation and IT Strategy Toolkit and Templates
Digital Transformation and IT Strategy Toolkit and Templates
 
Event Report - SAP Sapphire 2024 Orlando - lots of innovation and old challenges
Event Report - SAP Sapphire 2024 Orlando - lots of innovation and old challengesEvent Report - SAP Sapphire 2024 Orlando - lots of innovation and old challenges
Event Report - SAP Sapphire 2024 Orlando - lots of innovation and old challenges
 
Top mailing list providers in the USA.pptx
Top mailing list providers in the USA.pptxTop mailing list providers in the USA.pptx
Top mailing list providers in the USA.pptx
 
Brand Analysis for an artist named Struan
Brand Analysis for an artist named StruanBrand Analysis for an artist named Struan
Brand Analysis for an artist named Struan
 
FIA officials brutally tortured innocent and snatched 200 Bitcoins of worth 4...
FIA officials brutally tortured innocent and snatched 200 Bitcoins of worth 4...FIA officials brutally tortured innocent and snatched 200 Bitcoins of worth 4...
FIA officials brutally tortured innocent and snatched 200 Bitcoins of worth 4...
 
Exploring Patterns of Connection with Social Dreaming
Exploring Patterns of Connection with Social DreamingExploring Patterns of Connection with Social Dreaming
Exploring Patterns of Connection with Social Dreaming
 
Call 8867766396 Satta Matka Dpboss Matka Guessing Satta batta Matka 420 Satta...
Call 8867766396 Satta Matka Dpboss Matka Guessing Satta batta Matka 420 Satta...Call 8867766396 Satta Matka Dpboss Matka Guessing Satta batta Matka 420 Satta...
Call 8867766396 Satta Matka Dpboss Matka Guessing Satta batta Matka 420 Satta...
 
-- June 2024 is National Volunteer Month --
-- June 2024 is National Volunteer Month ---- June 2024 is National Volunteer Month --
-- June 2024 is National Volunteer Month --
 

Dynamic Pricing in Ride-Hailing Platforms

  • 1. Case Study: Dynamic Pricing in Ride-Hailing Platforms Hamed Shams Product Manager at
  • 2. Abstract Over the past decade, ride-hailing platforms revolutionized urban transportation. However, one of the most significant challenges faced by these platforms is the constantly fluctuating supply & demand volumes which causes serious issues in dynamically setting prices. This case study examines surge pricing in ride-hailing platforms, with a focus on UBER's hexagonal indexing system (H3). The study explores the challenges of implementing this system in Snapp, particularly H3's inability to adapt to extreme fluctuations. More precisely, having a prefixed hexagon resolution level for each city is not always optimal, as it would result in considerable no-decision cases due to the lack of signals. To address this issue, the author proposes a machine learning based approach and discusses the impact of launching this solution into production. This product resulted in a 27% decrease in no-decision cases, showcasing the effectiveness of the approach.
  • 3. Introduction Dynamic pricing is the study of finding the optimum price of products / services in a frequently adjusting manner, which applies to both online services (e.g., e-commerce, platforms) and brick-and-mortar stores. Nowadays, technologies have paved the way for setting prices more accurately by considering many factors, and also much more frequently (almost real-time). With this specific definition, we can say “dynamic pricing” emerged in the 1980s, driven by technological innovations. It was pioneered by the airline industry, which used factors like departure time, destination, and season, to automate their pricing.
  • 6. Passengers’ Expectation • Lowest Fare • Lowest Waiting Time Pricing in Ride-Hailing Platforms Drivers’ Expectation • Maximum Earning
  • 7. The Price is Upfront… A fare estimation of the ride price before it start.
  • 8. 1. Distance Factor Price = [ Base Factor + ( Price per Km * Km ) ] e.g in Tehran Price = [ 4000 + ( 750 * 8.1 ) ] = 10,000
  • 9. 2. Time Factor (ETA) Price = [ Base Factor + ( Price per Km * Km ) + ( Price per Min * Min ) ] e.g in Tehran Price = [ 4000 + ( 350 * 8.1 ) + ( 300 * 30 ) ] = 16,000
  • 10. How about the Marketplace imbalance?
  • 11. 3. Surge Factor Price = [ Base Factor + ( Price per Km * Km ) + ( Price per Min * Min ) ] * Surge E.g., Surge = 80% Price = [ 4000 + (350 * 8.1) + (300 * 30) ] * 1.8
  • 12. # Ride Request events # Price Check events Demand Rate = # Accept events # Ride Request events Supply Rate = Measured by:
  • 13. 7:00 7:10 7:20 7:30 7:40 7:50 8:00 8:10 Time Boundaries Space Boundaries
  • 14. The procedure of ordering a ride is a series of location-based actions over a period of time. Therefore, Supply and Demand conversion rates need to be defined over a specific unit of space and time. To be more accurate, all the events need to be collected and processed with respect to a predefined space and time unit; Otherwise, no conversion rate can be calculated. Without a clear definition of space/time, nothing can be interpreted nor compared. Time and Space Boundaries
  • 15. Time and Space Boundaries: “No-Decision” Cases A minimum number of “price check” events is required within a predefined time/space unit, in order to consider the conversion rates coming from those price checks as valid. Therefore, whenever the number of available events within a time/space boundary is lower than this minimum threshold, we consider it a “no-decision” case, because we can not make any decision for adjusting the price.
  • 16. Time and Space Boundaries: How to Define? The behaviour of supply and demand is directly tied to the number of available drivers and passengers within a given area, which of course changes over time, based on many factors, such as traffic patterns. One of these two dimensions must be tightly defined (as a hard constraint) so we can start playing with the other in a more dynamic manner, and hopefully, achieve an optimum solution ultimately. By considering a fixed time window (say 10-minute intervals) we can start defining the units of space; in other words, indexing the geospatial data.
  • 17. Geospatial Indexing in Surge Pricing 1. Defining some Districts Manually 2. Hexagon Hierarchical Indexing
  • 18. Problems with Manually Defined Polygons 1. High Variations within each Polygon 2. Phantom Demand
  • 19. 1. High Variations within each Polygon Imagine a football match is happening at the Azadi Stadium… Azadi Stadium Chitgar District
  • 20. 1. High Variations within each Polygon increase in the demand of ”Chitgar” increase in the price of “Chitgar” Azadi Stadium Chitgar District
  • 21. 1. High Variations within each Polygon Even for riders on the low-demand areas of this district, which is not fare. Azadi Stadium Chitgar District
  • 22. 2. Phantom Demand Azadi Stadium Imagine a driver realize the demand of “Chitgar” district has raised (surged) Chitgar District
  • 23. 2. Phantom Demand But, there may be no ride request since the real demand is on the other side Azadi Stadium Chitgar District
  • 24. Solution: A more granular and dynamic approach
  • 26. Advantages of Hexagonal Indexing 1. First and foremost, it is not manual! 2. More granularity in supply-demand (address problem #1) 3. Smooth gradients (address problem #2) 4. Symmetric shapes and equal partitions 5. Equal distance from the neighbors
  • 27. Hexagon Resolution Level H3 supports 16 different diameters. Each finer resolution, has cells with one seventh the area of the coarser resolution. Hexagons cannot be perfectly subdivided into seven hexagons, so the finer cells are only approximately contained within a parent cell.
  • 29. Critical Issue: a single resolution can not work everywhere everytime The behaviour of supply-demand is extremely fluctuating, which leads to sudden rises and falls in the “price check” events. Therefore, by using a specific (fixed) resolution level for each city, two things would happen: • Hexagons are too small, so they rarely meet the predefined min threshold, leading to no-decision cases. • Hexagons are too big, resulting in the similar scenarios with the Manually Defined Polygons. i.e., lack of granularity and smoothness. H3 addressed this issue with the “K-Ring” function…
  • 30. If the events within a hexagon does not meet the min ”price check” threshold L1 H3’s Geo Expansion Function: “K-Ring”
  • 31. Then, the hexagon will be expanded to its neighbors (L2) L1 L2 L2 L2 L2 L2 L2 L1 K-Ring-1 H3’s Geo Expansion Function: “K-Ring”
  • 32. L1 L2 L2 L2 L2 L2 L2 L1 L2 L2 L2 L2 L2 L2 L3 L3 L3 L3 L3 L3 L3 L3 L3 L3 L3 L3 L1 K-Ring-1 K-Ring-2 If the events are not enough yet, then it will get expanded another level (L3) H3’s Geo Expansion Function: “K-Ring”
  • 33. Problem Definition While the Geo Expansion (K-Ring) approach has proven effective in mitigating no-decision cases, our experiments have shown that it is insufficient to handle extreme fluctuations. This often results in outdated surge factors and unadjusted prices, causing an unbalanced marketplace. The issue is particularly pronounced during rush hours in commercial regions, where the system struggles to keep up with the influx of incoming "price check" signals. Moreover, manually adjusting resolution levels for each city is an impractical and potentially biased approach. Therefore, there is a need for a more robust and automated solution that can adapt to extreme fluctuations in a timely and effective manner.
  • 34. Proposed Solution To address the issues of outdated surge factors and unadjusted prices caused by extreme fluctuations, an automated solution is required to minimize no- decision cases and accelerate price adaptation. This can be achieved by dynamically adjusting the resolution level in each city, based on the estimated number of "price check" events in the next time interval. To accomplish this, we developed a time series forecasting approach to predict the number of "price check" events based on space and time. For example, we can forecast the number of price check events in Tehran within the next hour. To maintain confidentiality, we have omitted implementation details from this document.
  • 35. Business Impact • About 27% decrease in the no-decision cases overall. The reduction rate was much more considerable during midnight and rush hours. • Resulting in a 6% increase in drivers’ earnings while improving rider’s conversion. In other words, a more reasonable and fair pricing.