The Advantages of Being There Design at Microsoft Research India Kentaro Toyama Assistant Managing Director Microsoft Research India IWIPS 2007: “Actually.

The Advantages of Being ThereDesign at Microsoft Research India

Kentaro Toyama

Assistant Managing Director

Microsoft Research India

IWIPS 2007: “Actually Being There”

June 29, 2007 – Merida Mexico

Outline

The Challenge of India

The Value of Being There

The Five Stages of Design

Three Projects from MSR India– Warana Unwired– MultiPoint– Text-Free User Interfaces

The Last Stage?

Outline





The Last Stage?

India

People• ~1.1 billion people

– Over half under 25 years old• 22 languages• Annual incomes $100-$100M+• 28 states

Area• ~1/3 the area of United States

Technology• ~20M PCs, installed base• ~140M mobile subscriptions

– +7M each month

Sources: CIA Factbook, TRAI, CNN

Roads in India

India, a Personal View

My first trip to India (2004)

India, a Personal View

People• ~1.1 billion people

– Over half under 25 years old• 22 official languages • Annual incomes $100-$100M+• 28 states

Area• ~1/3 the area of United States

Technology• ~20M PCs, installed base• ~140M mobile subscriptions

– +7M each month

but, power held by fewtremendous energy and optimism

incredible diversity, EM microcosmreminiscent of European Union

impact of weather (ubiquity of agriculture)

huge interest in PCs, by everyonemobiles, mobiles, everywhere

Huge potential opportunity for computing industry.

But, there are new challenges that

neither India nor the industry have ever faced before.

Infosys campus, Bangalore

A small Internet café on a market street in a town near Bombay

Rural village with a VSAT Internet connection near Bhopal, Madhya Pradesh

Microsoft Research India• Established January, 2005

• Goals– World-class academic research– Contributions to Microsoft products and

businesses– Support growth of research programs

in India and elsewhere

• Six research areas– Cryptography, Security, and Algorithms– Digital Geographics– Mobility, Networks, and Systems– Multilingual Systems– Rigorous Software Engineering– Technology for Emerging Markets

• Currently ~50 full-time staff, growing

• Collaborations with government, academia, industry, and NGOs Microsoft Research India

Sadashivnagar, Bangalorehttp://research.microsoft.com/india

http://research.microsoft.com/india

Technology for Emerging Markets

Understand potential technology users in economically poor communities:– E.g., urban domestic labourers– E.g., rural entrepreneurs

Adapt, invent, or design applications of computing that contribute to socio-economic development of poor communities worldwide.

(Focus on research, not on shipping product.)

Computer-skills camp in Nakalabande, Bangalore(MSR India, Stree Jagruti Samiti, St. Joseph’s College)

Microsoft Research India

Interdisciplinary ResearchAishwarya Lakshmi Ratan

– Public Administration and International Development

Jonathan Donner– Communications

Nimmi Rangaswamy– Social Anthropology

Rajesh Veeraraghavan– Computer Science and

Economics

Indrani Medhi– Design

Kentaro Toyama– Computer Science

Randy Wang

Udai Singh Pawar

–

–

Computer Science

Physics

Society

Group

Technology

Individual

Society

Group

Technology

Individual

Innovation

Understanding

Impa

ct

Innovation

Understanding

Impa

ct

Rikin Gandhi– Astrophysics

Often hungry

Childrennot inschool

In perpetualdebt

Breadwinnerin formalsector

“Middleclass”

1

2

1

remainingdata

t0

Computers in Agriculture

Multi-mouse for EducationWell-Being Map Urban Consumer

Digital Study Hall

IT and Microentrepreneurs

Cost-Aware Data Transfer

Rohan MurtyComp. Sci., Harvard

Randy WangComputer Sci, UC Berkeley

Jonathan DonnerCommunications, Stanford

Udai Singh PawarPhysics, IIT Kanpur

Aishwarya Lakshmi RatanPublic Admin., Harvard

Nimmi RangaswamySociology, Univ. of Mumbai

Rajesh VeeraraghavanComp. Sci. & Econ., Clemson

Renee KuriyanEnergy and Res, UC Berkeley

Feature phones as “bar-code” readers for data-entry in rural microfinance

Information ecology of small businesses in developing markets

Cost-aware transfer of data across heterogeneous channels, e.g., for mobiles

DVD exchange over postal service and TVs as display for rural education

Study of dynanic middle-class consumers in urban emerging markets

Experiments with computing and communication systems in agriculture

The state’s role in rural IT projects, with a focus on Kerala’s Akshaya project

Transitions between states of wealth in emerging markets

Text-Free UI

Indrani MedhiDesign, Illinois Inst. of Tech.

UIs without text for users who are illliterate and may never have seen a computer before

Government and Rural IT

Outline





The Last Stage?

… to resolve contradictory generalizations:

• Resistance to new technology – But computers have glamour

• Poverty systemic and multi-dimensional– But households functional

• Stark lack of money– But willing to spend

• Information critical…– But rarely the bottleneck

• Computing needs are minimal– But there are opportunities!


Resistance to Technology…Many factors inhibit use of

technology:

• High cost

• Reluctance to depart from habits and traditions

• Fear of breaking technology

• Lack of awareness of technology’s functional value

• Barriers of education or literacyA child trying to explain to

her mother what is on a laptop screen.

But, Computers have GlamourExamples of interest in computing

technology:

• Retention rates at schools rise when the school has PCs.

• Rural PC kiosk owners see a rise in their confidence and status in community.

• Office service staff eager to learn about PCs and how to use them.

These examples have little to do with computer function.

A kiosk operator running a near Tiruvallur, Tamil

Nadu

Poverty is Systemic…Stable system makes escape

difficult:

• Lack of money means lack of time to do anything other than survive.

• Lack of time means less time for education.

• Lack of education means fewer job opportunities.

• Lack of job opportunities means lack of money.

“Shocks” to household create downward spiral, and there are always shocks:

• Health problem requires loan• Loan incurs interest• Interest payments prevent capital

accumulation

A government-sponsored mid-day meal in a Tamil

Nadu school.

But, Households still Functional“Good enough” solutions exist:

• Credit: All kinds of loans available

• Healthcare: Traditional medicines, primary healthcare services

• Agriculture information: agriculture extension, word of mouth, salesmen

Persistent Lack of Money…

Bangalore guideline for 45 minutes of housework a day: Rs. 150 (US$3)… per month!

Typical daily wage for agricultural labor: Rs. 60 per day (US$1.33; Rs. 30 for women)

Public-school teacher’s salary varies from Rs. 3000 to Rs. 8000 (US$67-178) per month.

Teachers on a school trip in Karnataka

But, Willingness to SpendLuxury and aspirational

consumption not unusual:

• Weddings costing Rs. 1 lakh (US$2200) in rural villages not infrequent (cf., avg. per capita GDP of ~US$700)

• Mobile phone ring tones popular even at Rs. 10 (US$0.20) per song

• Photography services to “enhance” photos popular. Cost range from Rs. 100 to Rs. 600 (US$2-12)

A Photoshop’ed photo of a village bride (Maharashtra)

Information is Critical…General lack of information hampers

quality of life:

• Hygiene and healthcare knowledge shallow or superstitious

• Poor fundamental and vocational education impedes career growth

• Very practical knowledge not readily available:– Government schemes for the

poor– Job information– Value of savings and investment

A 12-year-old enrolled in typing lessons at a rural

PC kiosk

But, Information not the BottleneckAccess to information not the

problem:

• Physical transfer of goods/cash often required. Transport infrastructure is poor.

• Levels of formal education very low, even with literacy. Education required to distinguish good information from bad.

• Other factors…– No faith in information source– Lack of time or money– Rigid mindsets

A petty shop owner in Tamil Nadu

Computing Needs Minimal…

Information processing rarely required…

• Little use of documents, charts, spreadsheets.

• Paper , pen, and manual calculation difficult to out-do:– Low cost– Lightweight, durable– Additional training not required

But, Technology can Help!

To draw interest of community.

To process and analyze aggregate data.

To streamline or improve existing processes.

Focus group on a potential technology-for-agriculture

project

Removal or reconciliation of preconceptions is the primary value of fieldwork.

General lessons are difficult to draw; contradictions abound.

Fieldwork helps to identify the specific constraints that apply to a given domain or application.

Women from Ariyapalayam, Tamil Nadu, husking corn

The Value of Being ThereRecap

Outline





The Last Stage?


Stage Knowledge Gained

Wonder Technology / Surface Problem

Exuberance Surface Solution

Realization Real Problem

Adjustment Real Solution

Identification User

Deeper

Intuition

Good design comes out of deep intuition into the user.

“Kids in the developing world need the newest technology, especially really rugged hardware and innovative software.”

– Nicholas Negroponte, from the One Laptop Per Child website (2005)

Exuberance

“The world's poorest two billion people desperately need healthcare, not laptops.”

– Bill Gates (WRI Conference, Seattle, 2000)

Realization

Outline





The Last Stage?

Warana UnwiredRajesh Veeraraghavan

Over 60% of population in agriculture

Mostly small and marginal farmers with 1-3 acres of land

Average income of $1-2 per day

Agriculture in India

Wonder

Sugarcane cooperative

70 villages, 70000 farmers

Asia’s first “Bridging Digital Divide” pilot (1998)

“Warana Wired Village Project”

Wonder

FactoryFactoryFTP

FTP

FTP

PC

Warana Farmer DB

Standard PC network

Weigh stations

Landline phone

PC enabledKiosks

54 kiosks in 54 villages

Cost: Rs.2.5 crores (US$500,000)

Rural PC Kiosks

Exuberance

• Check market price information

• Provide agricultural advice to farmers

• Conduct land-record transactions

• Surf the Internet

• And, do it all with a private business model!

Original Goals

Exuberance

Allow farmers to…

• Check market price information

• Provide agricultural advice to farmers

• Conduct land-record transactions

• Surf the Internet

• And, do it all with a private business model!

Original Goals

Realization

Allow farmers to…

Internal account MIS:

• Issue harvesting permit• Buy fertilizer through credit• Get paystub

• Query quantity of sugarcane harvested

Actual Use

Realization

High maintenance cost

Intermittent power

Network flaky

PC not optimally used!

Mounting Challenges

Realization

Can we preserve the functionality of the existing PC based system while making the entire system cheaper and more effective?

The Problem

Realization

The Solution: Warana Unwired!

SMS-enabledmobile phones

PC-basedkiosks

Adjustment

FactoryFactoryFTP

FTP

FTP

PC

Warana Farmer DB

Standard PC network

Weigh stations

Landline phone

PC-enabledkiosks

Original PC-Based Set-Up

Adjustment

GSM/CDMA

SMS network

FactoryFactoryPC

Warana Farmer DB

Standard PC network

Weigh stations

SMS-enabledphones

New Mobile-Based Set-Up

Windows

Mobile Remote APIs

SMS

SMS

SMS

Adjustment

24-hour access to services – 6000 SMS processed

80% of requests for getting sugarcane output

1238 unique farmer requests

Response time on harvesting data.– Original: 15 days PC: 2 days

Mobile: immediate

Telcos’ interest has perked up.

Neighboring cooperatives have expressed interest.

Warana Unwired – Results

Adjustment

System Cost/Farmer/Year

New PC System

394

Existing PC System

177

SMS Mobile (kiosks)

159

GPRS(kiosks) 139

SMS Mobile(without kiosks)

111

GPRS ( no kiosks)

91

Units: Rs COST DETAILS:

Common cost:Kiosk rent, Kiosk salary

SMS cost: 50 paise/SMS

GPRS per byte cost: 7000 times cheaper than SMS cost

High Maintenance cost: UPS battery, Hard disk, printer, monitor

No GPRS coverage

Low end phones do not support GPRS

SMS data plans are dropping

Savings over PCs 1 million Rupees /54 villages/1 year($22,000)

Costs

Adjustment

Adjustment

Farmer from pilot village expresses initial disbelief…

Once he sees it on the phone, he gets excited and says,

“Barabar hai, eh tho bahuth accha hai.”“The information is exact and it is very good.”

Farmer from another village demands access…

We tried to tell them that we were in a testing phase, to ensure that the system worked; the farmer replied,

“I saw messages are coming on the mobile phone. There is no problem. So where is the question of success?”

Farmer Response

Adjustment

So far…• Successful replacement of kiosks in

seven villages. System in operation since October 2006.

• Expansion to other villages in cooperative

To do…• Analysis of feedback and surveys

for concrete impact• Pilots with other cooperatives

Status

Adjustment

Outline





The Last Stage?

MultiPointUdai Singh Pawar, Joyojeet Pal (UC Berkeley), Kentaro Toyama

Education in India

Rural school in Chinhat, Uttar Pradesh

Wonder

300M children aged 6-18; 210M enrolled in school; 105M actively attending.

Mostly small and marginal farmers with 1-3 acres of land

Teachers poorly trained and frequently absent

No toilets

No walls

No permanent building

Terrible student-teacher ratio

Intermittent electricity

UPS broken

Frequent maintenanceof PCs required

Teachers not computer literate

Caste discrimination

Religious discrimination

Students hungry

Poor retention rates

Poor pay for teachers

Teacher absenteeism

Student illness

No supplies

No textbooks

Parents uninvolvedChild labour Teachers multitasking

Irrelevant curriculum

Heat

Many children per computer

Rural Education: Problems

Wonder

No toilets

No walls

No permanent building

Terrible student-teacher ratio

Intermittent electricity

UPS broken

Frequent maintenanceof PCs required

Teachers not computer literate

Caste discrimination

Religious discrimination

Students hungry

Poor retention rates

Poor pay for teachers

Teacher absenteeism

Student illness

No supplies

No textbooks

Parents uninvolvedChild labour Teachers multitasking

Irrelevant curriculum

Heat

Many children per computer

Rural Education: Problems

Exuberance

PCs in Rural Education

5-10% of primary schools in India already have a PC.

PC classrooms generally used to “babysit” students as teacher teaches other classes.

Typically, 2-6 PCs per primary school.

Exuberance

At school after school…

One PC, many children.

Realization

MultiPoint: Solution

Provide a mouse for every student

– One cursor for each mouse, with different colours or shapes

– USB mice• Have tried up to 20

– Content modified • Game-like environment

Adjustment

MultiPoint: Demo

Adjustment

MultiPoint: ResultsPreliminary user studies [ICTD2006]

• Questions

– Can students understand MultiPoint paradigm?

– How do children interact with MultiPoint?

– Does MultiPoint increase engagement?

• Methodology

– Trials:• 20 min single mouse• 20 min MultiPoint• 10 min free play

– 3 trials of 6-10 children

Before

Adjustment

MultiPoint: Early Results• Everyone wants a mouse.

– Girls more likely to share than boys.

• Kids understand MultiPoint immediately.

• All students more engaged for longer periods of time.– Even children without mice engage

longer.

• Self-reporting is positive.– Exception: one student didn’t like

MultiPoint because of competitiveness

Before

After

Adjustment

MultiPoint: Advantages

Incentives aligned

– Cost effective: One computer + 5 mice comes to ~$100 per child.

– Content authors can adapt to paradigm

– Government / administrators can claim better use of computers

– Teachers can keep more students entertained

– Students have more fun (cf., multi-player computer games)

Adjustment

MultiPoint: Current Work

Current work

– Software SDK for content writers to be released in August 2006

– Technical features to maximize educational value of MultiPoint

– More user studies to test pedagogical value

– Pilots with NGOs in India– Hoping to disseminate beyond

India

New hypothesis: Better for primary education than one PC per child?

Adjustment

Outline





The Last Stage?

Text-Free User InterfacesIndrani Medhi, Kentaro Toyama

Illiteracy

1-2 billion illiterate population in the world.

98% live in developing countries.

India’s rate of literacy (optimistically) estimated at

~60%.

Wonder

Target Users

Women from several Bangalore slums

Informal sector jobs

Income range: INR 800-2500 (USD 20-50) per month

Illiterate or semi-literate

Most have never seen a PC (those who have seen, only in their employers’ homes; but, not allowed to touch)

Wonder

Ethnographic Design

Exuberance

300+ hours with over 250 people from urban slums in Bangalore:

– Interviews– Participatory design – Rapid prototyping– Subject trials

Text-Free UI, Take 1

Design Principles:

– Pen or touch interface

– Liberal use of icons and images

– Voice feedback

– Care in details of graphics; semi-abstracted cartoons

– Aggressive use of mouse-over functionality

– Consistent help iconMonster.com for domestic labourers?

Maps for illiterate users?

Exuberance

User Studies, Take 1

Task: For a friend who is unemployed, find the best-paying job in her neighborhood.

Results: Subjects could manipulate the application, but only 30% completed the task, even with significant prompting:

Problem: Deeper problem in motivation and lack of cognitive model of how the PC worked.

Realization

Can any UI be converted into one that is usable by illiterate users?

ILLITERACY

FEAR OF TECHNOLOGY

LACK OF TRUST IN TECHNOLOGY

LACK OF AWARENESS OF WHAT TECHNOLOGY CAN DELIVER

New Problem!

Can a UI be developed to allow an illiterate, first-time PC user to access information he/she needs without any assistance or prompting?

New question:

Original question:

Realization

Full-Context Video

A full-context video explains the broader context of the application and how it works, in addition to instructional material about how to use the application.

Full-Context Video

Adjustment

Full-context video has clear value:

Without video first, only one out of 17 (6%) was able to complete the task at all, taking 11 prompts and 8.2 minutes

With video first, 18 out of 18 completed the task, with an average of 4.7 prompts and 6.5 minutes

Without Video (A)

With Video (B)

Total

Task Completed (out of 35) 8 35

Prompts reqd for completion 9.8 5.2

Avg completion time (min) 9.01 4.59

Sequence - AB

Task Completed - AB 1 17

Prompts reqd for completion 11 5.9


Sequence - BA

Task completed – BA 7 18

Prompts reqd for completion 6 4.7


User Studies, Take 2

Adjustment

Qualitative Results

Other observations:

Round-two subjects were incredulous that round-one subjects didn’t understand the application.

Impact of video not permanent for most subjects. Many wanted to see the full-context video each time, even after seeing it before.

Full-context video appears to increase motivation, as well as performance.

Those who saw full-context video were interested in providing feedback on the specifics of the UI.

Adjustment

Text-Free UI, Take 1

Design Principles:

– Pen or touch interface

– Liberal use of icons and images

– Voice feedback

– Care in details of graphics; semi-abstracted cartoons

– Aggressive use of mouse-over functionality

– Consistent help icon

– Full-context video

Adjustment

Outline





The Last Stage?

The story thus far…

Stage Warana Unwired MultiPoint Text-Free UI

Wonder agrarian society rural education illiteracy

ExuberanceInternet for

farmers!PCs in

classrooms!UIs without text!

Realizationlittle PC value for

most farmerstoo many children

per PCUI manipulation

not the issue

AdjustmentSMS-based

inventory querymultiple mice per

PCfull-context video

Identification ? ? ?

Why not PC kiosks?

Identification

Soup kitchens brand themselves another wayand serve a different kind of community.

Four-star restaurants brand themselves one way, and serve a particular clientele.

The importance of branding in serving food…

It’s difficult to serve both client groups in one physical location.

Why MultiPoint?

Identification

Games are fun, and games with other kidsare even more fun!

People are already accustomed to sharing hardware.

Why Full-Context Video?

If you were told that, if you put anything in this box and spelled

the object’s name 100 times out loud, it would come to life…

Identification

“Any sufficiently advanced technology is indistinguishable from magic.”

– Arthur C. Clarkewould you bother to do it?

Recapitulation

StageWarana Unwired

MultiPoint Text-Free UI

Wonder agrarian society rural education illiteracy

ExuberanceInternet for

farmers!PCs in

classrooms!UIs without text!

Realizationlittle PC value

for most farmers

too many children per PC

UI manipulation not the issue

AdjustmentSMS-based

inventory querymultiple mice per

PCfull-context video

Identification“Just the info I

need.”“We’ll share and play!”

“Demystify it for me.”

Cyclical

process

Outline





The Last Stage Starts It All Over Again!

Thank you!http://research.microsoft.com/research/tem

Questions? [email protected]

http://research.microsoft.com/research/tem



The Advantages of Being There Design at Microsoft Research India Kentaro Toyama Assistant Managing Director Microsoft Research India IWIPS 2007: “Actually.

Documents

bangaloremsr india

comindia technology

computing industry

microsoft products

stages of designthree

peopleover half

mobile subscriptions

emerging marketscurrently