Automatic dosa making machine project

IMB 439

U. Dinesh Kumar, Professor of Quantitative Methods and Information Systems, prepared this case for class discussion. This case is not intended to serve as an endorsement, source of primary data, or to show effective or inefficient handling of decision or business processes.

Copyright © 2013 by the Indian Institute of Management Bangalore. No part of the publication may be reproduced or transmitted in any form or by any means – electronic, mechanical, photocopying, recording, or otherwise (including internet) – without the permission of Indian Institute of Management Bangalore.

U. DINESH KUMAR

DOSA KING: A STANDARDIZED MASALA DOSA FOR EVERY INDIAN

On August 31, 2013, the famous Madras Club in Chennai was buzzing with visitors searching for a place to relax over the weekend. I met R. Narayanan, an entrepreneur, in one of the small rooms in the Madras Club. We placed our order for tender coconut water and coffee. Without wasting much time, I asked Narayanan to share the story of one of his initial ventures, the Dosa King. The Dosa King was aimed at providing hygienic and standardized masala dosas across India. A dosa is a savoury pancake made from a fermented batter of rice and black lentils; a masala dosa is a dosa with a potato masala (spiced potatoes) filling. Masala dosas (Exhibit 1) are usually served with several accompaniments, such as coconut chutney and sambar (a kind of spicy lentil and vegetable stew). According to Narayanan:

Dosa was by far the single most popular snack across the length and breadth of India; yet, eating places that served dosas in many parts of the country, especially in north India, were limited in number, largely due to the lack of understanding about the “formula” for the dosa as well the lack of an “expert” dosa maker.

Narayan established Indian Foods and Fermentations Ltd. (IFFL) in 1992 to offer a total solution to the problem stated above with an objective to develop an electronically controlled operator assisted dosa machine (ECOAD). According to market research conducted by IFFL in 1992, 20 million dosas were consumed in India on any given day. Narayanan and his colleagues at IFFL decided to come up with a complete dosa delivery system. The following requirements for the dosa delivery system were set by his team.

1. To design a dosa-making machine that could deliver consistent dosas at 30-second intervals, with uniform weight, crispiness, color, size, and taste.

2. To develop a dosa batter in dry or wet form with a shelf life of 4 weeks under refrigeration, which could be transported and used in any part of the country.

3. To develop a potato masala that could be automatically dispensed from the dosa machine in consistent quantities as desired.

4. To deliver the dosa to the customer in a convenient form such as a roll, so that the customer could eat the dosa on the move.

Narayanan was confident that the Dosa King would change the landscape of fast food restaurants in India.

BACKGROUND

Narayanan held a Post-Graduate Diploma in Management from the Indian Institute of Management Calcutta and a Bachelor’s degree in Commerce from Chennai. Narayanan worked for Coca-Cola and Nestlé before setting up Indian Foods and Fermentations Ltd. After several years in the food industry, Narayanan saw an opportunity in the fast food business in India. In 1992, India did not have any western fast food chains such as McDonalds, Burger King, or KFC. Restaurants were managed as small business entities by a few entrepreneurs and there were no major restaurant chains across the country. India went through a period of major economic reform in 1991 and the economy started growing much faster after the nineties. Narayanan expected the demand for fast food to grow in all major cities of India. Equipped with his experience at Coca-Cola and Nestlé, Narayanan was confident that he could change the restaurant industry in India. This strong belief led to the creation of the IFFL in 1992.

This document is authorized for use only by AJ Roberto (aj.roberto23@gmail.com). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

Nur Demir
Dosa King: A Standardized Masala Dosa for Every Indian Page 2 of 8

USER REQUIREMENT ANALYSIS FOR DOSA KING

The IFFL’s market research showed that on any given day, 90% of the studied adult males would eat some snack outside their home. Adult males in the 16–45 years age group formed the core target audience. The Dosa King was intended to be a mass distributed item, and the number of retail points across the country was meant to be in excess of 10,000 within 5 years from its launch in 1992. The IFFL realized that their largest stakeholder would therefore inevitably be the retailer or franchisee who invested in dosa machines, batter, and masala, and also interfaced with the customer.

Standardization

Although the dosa was originally a snack/breakfast item, it also served as a midday meal in many parts of the country. Moreover, in many parts of the country, the dosa was an item meant for “out of home” consumption. Since the ingredients, the formula, and the expertise required for making a dosa were all variables, there was no standardization in the business. Customers had a sense of what they considered a good dosa, but were unable to describe the same. The size of the dosa was another variable that lacked standardization. Accompaniments to the dosa (sambar, chutney, and so on) also lacked standardization. Paper masala dosas had already fashioned a need among customers for crispiness. Customers, therefore, expected an initial crunch when biting into their dosa. The dosa also had to achieve a consistent brown color as per customer expectation.

Final customer satisfaction depended on a combination of organoleptic factors (including visual appeal), all of which was to be achieved through an automatic dosa vending machine. The accompaniments served with the dosa would satisfy the variety-seeking behavior of the customers; therefore, it was decided to vary the accompaniments constantly to “pull” customers into the store.

Delivery of dosas

Data collected at a few restaurants revealed that the arrival of customers for a snack/mini-meal followed a Poisson distribution, and dosas would have to be delivered at intervals of 30 seconds. Although it was intended to be a dosa- making-and-dispensing machine, the customer expectation from a dosa machine would be virtually instantaneous delivery of dosas.

Size of dosas

The marketing department at the IFFL had visualized the dosa as a “stand up and eat” affair at a pricing of INR 10 for a masala dosa weighing 175 grams (1$ = Rs. 62, in November 2013). The dosa was, therefore, to be served and eaten in the form of a roll and the accompaniments were to be poured on top of the vertical dosa roll, similar to pouring ketchup or chutney over other fast food items. The size of the dosa was to be 6 inches in diameter. The dosa was to be delivered out of the machine in the form of a roll, so that it could be easily scooped into a wrapper and handed over to the customer.

Accompaniments

The customers expected to have the sambar and the chutney served along with the dosa, not poured over the dosa. The customers also expected these accompaniments to be available in liberal quantities and on demand.

Liquid chaser

The customers preferred to have either a hot beverage (tea or coffee) or a soft drink after having the dosa; in any instance, water was expected to be served after/during consumption. Customers expected these beverages to be available at the same premises where the dosas were sold.

SYSTEM REQUIREMENTS

The IFFL conducted a market survey to capture the “voice of the customers” for the dosa delivery system and

This document is authorized for use only by AJ Roberto (aj.roberto23@gmail.com). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

Nur Demir
Dosa King: A Standardized Masala Dosa for Every Indian Page 3 of 8

identified various system requirements, which were as follows.

Dosa machine

1. The dosa machine was to have a footprint not exceeding 6 sq. ft. It should be mobile enough to be pushed out of an eatery to occupy the footpath without substantially hindering pedestrian traffic.

2. The dosa-making process should be visible to the customer in order to emphasize the hygiene factor. The customer would have visual access to the entire process from one side and the operator of the machine would have access to the inner working of the dosa machine on the other side.

3. The countertop on which the dosa machine was set up would serve as the counter for customer transactions. 4. The accompaniments would be placed at a separate counter away from the dosa machine so as to avoid

congestion at the dosa counter itself. These accompaniments would be made available in a variety of exotic flavors, which would be the selling pitch.

5. The entire machine should be electronically controlled, with an interface that would permit the automatic upload of the logs of the maintenance engineers’ preventative maintenance visits. The login and logout times were to be recorded, with a password required for access.

Dosa batter

The dosa batter was a fermented product. Once the fermentation process commenced, not only would it be in a permanent state of motion, but there would also be physical changes. The product (i.e., the batter) would become more viscous; over time, the liquid components would separate from the solids. When making dosas in a home setting, before spreading the batter on the tawa (pan/griddle for making dosas), the batter would be mixed using a ladle. Dosa batter had a shelf life of approximately 16 hours; after this, it would sour and would be unfit for further use.

In the 1960s, the Central Food Technological Research Institute (CFTRI) based in Mysore had developed a dosa batter that was chemically derived. This was primarily meant as a “convenience product”; however, in reality, the taste of this product was never accepted by consumers.

By using preservatives, University Department of Chemical Technology (UDCT) in Bombay found it possible to slow down the fermentation process, vacuum-pack the dosa batter (in wet form), and transport it to the final place of use in refrigerated trucks. This batter could be preserved for up to 4 weeks under refrigeration.

Potato masala

Potato masala (spiced potatoes) was an important ingredient of the masala dosa. Potato masala should ideally contain onions and all the required masalas at acceptable levels. The cooked potato masala could be packed in retort pouches. The product would then have a shelf life of approximately 6 months; this potato masala could be transported and stored in ambient conditions. Potatoes could also be cooked and drum-dried for instant reconstitution into a potato mash. However, the required masalas, onions, and other condiments would need to be added locally.

Accompaniments

The sambar (lentil and vegetable stew) would need to be pre-prepared. Some form of dehydration of the sambar mix or a part thereof was required. The sambar could then be reconstituted at the franchise. In most parts of the country, the experience of having a masala dosa was incomplete without coconut chutney. In several parts of the country where coconuts were not easily available, dal-based chutney would be used. This kind of chutney was considered a poor substitute for coconut chutney and was not readily accepted by the customers. In South India, onion chutney was also served with dosas; in some instances, mint and chili chutneys were served.

Standardization

The entire purpose of bringing a dosa machine to the marketplace was to standardize the product to a level where a customer would get a masala dosa of the same quality wherever in India he/she travelled. In fact, standardization

This document is authorized for use only by AJ Roberto (aj.roberto23@gmail.com). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

Nur Demir
Dosa King: A Standardized Masala Dosa for Every Indian Page 4 of 8

was the unique selling proportion (USP) for the IFFL since different restaurants produced dosas with differing size, taste, and crispiness. The IFFL’s objective was, therefore, to create an error-proof system for every single attribute associated with the dosa delivery system.

Beverages

Vending machines for coffee, tea, and soft drinks were available on attractive terms to the IFFL. However, the contract terms were not franchisee-friendly, and unless the IFFL had a very successful working model in place, none of these potential partners were willing to change their terms of reference. Setting the price of a beverage higher than the cost of a masala dosa was an unacceptable choice.

The IFFL saw this as a market opportunity to create a new brand of soft drinks. Reliable soft drink dispensers existed in the market, and it was possible to tie up with the manufacturers of an apple juice concentrate (for instance), Rasna, or some such non-aerated drink.

Packaging

According to Narayanan’s original idea, the masala dosa was meant to be eaten “on the go”. The issue was one of delivery. How was the dosa to be carried and eaten? In what wrapping would the IFFL offer the dosa? How should the chutney be dispensed?

Other issues

Franchising implied standardization on all counts, which included store layout, the product, and delivery process. Moreover, the process of cooking the dosas would produce steam, which would rapidly blacken the ceiling unless some mechanism for trapping and channeling out the steam was not built into the system. The layout of the store and the location of the dosa machine would become important. Once standardized, the left-to-right orientation of the dosa machine had to be strictly maintained and adhered to.

PRODUCT DEVELOPMENT CHALLENGES

Project funding

The project was funded by venture capital (VC), with minimal resources required from Narayanan. Options that would have been acceptable to a large organization with deep pockets were simply not available to Narayanan. In 1992, the venture capital industry in India was in its nascent stage. In fact prior to 1988, the Indian government had no policy on venture capital funds.1 The first guideline to legalize venture capitals was issued in 1988 aimed at allowing state-controlled banks to create venture capital funds. The guidelines were very restrictive, the funds were restricted to less than INR 100 million and the recipient firm should be involved in relatively new and “untried” technology.2 In the case of IFFL, the VC funding had a debt to equity ratio of 2, resulting in heavy financial burden on IFFL. Further, the risk appetite of the VC slowed down at the later stages of the project and a public issue became necessary. Financial institutions that had sanctioned the loan made the disbursement conditional to the appointment of merchant bankers to the public issue. The lead-time for such an appointment meant an additional delay of 4 months, and subsequently, the public issue was successfully completed and all funds were obtained only 6 months later. At the same time, it became necessary to “show results” to keep the VC disbursement in place. Therefore, the solution sought was a combination of the prevailing funds available. Further, the application of the available funds was prioritized based on those that could “show results”.

1 R Dossani and M Kenney, Creating an Environment: Developing Venture Capital in India, Working paper series on Berkeley Roundtable on the International Economy (BRIE #143). Downloaded at http://iis-db.stanford.edu/pubs/12010/Dossani_Kenney.pdf 2 Ibid

This document is authorized for use only by AJ Roberto (aj.roberto23@gmail.com). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

Nur Demir
Dosa King: A Standardized Masala Dosa for Every Indian Page 5 of 8

Dosa machine

The first VC funds were utilized for the manufacture of the dosa machine. Narayanan himself being a non-engineer, quickly managed to assemble a team of engineers to support him in the development process. At the time, the quote given by any design house varied from a few lakh rupees to INR 10 million. However, the total VC funds that had been assigned for the basic design were only INR 1 million.

The design was assigned to a company manufacturing coffee vending machines, which replicated the entire dosa- making process through a mechanical robot designed to imitate the human movements involved in the process. A robot poured the dosa batter on to a stationary tawa (steel griddle) and then heated the same to the desired level. The development of this machine took 9 months. The manufacturer delivered a plain dosa-making contraption that proved a dosa could be made and delivered in 90-second intervals without human intervention.

Based on this prototype, the project financing commenced. With funds in hand, the development activity moved to an engineering house. The concept of a rotating drum with a well on top was proposed – the wet batter could be dropped into the well on top of the drum. Concurrently, the original robot concept was also designed, since that was the design shown to the financial institutions.

The size and color of the dosa, the quantity of the dosa batter, the quantity of potato masala, and so on were not scientifically researched decisions. Each of these was taken for granted as the decision-making progressed. Narayanan took decisions based on intuition and the design process followed. Narayanan recalled:

It was only at this stage – as part of our various experiments to get the right color and the right crispiness – that we realized that the batter was a bigger problem than the dosa machine itself. We were unable to get the same size, weight, or color out of the same batch of batter and on the same day. It was at this stage that we realized that the batter would need standardization before the machine could be designed for any degree of accuracy. There was no way that we could go back to either the financial institutions or the VC and tell them that there was an un-quantifiable variable that had suddenly appeared on the horizon.

Therefore, an optimized dosa was created. While work proceeded on developing a standardized dosa batter, development of the dosa machine proceeded on the basis of this optimized dosa. The task of the team responsible for batter development would be to match the quality of the optimized dosa. The optimized dosa was created by using identical material daily; trials commenced at a fixed time every day. Several options – including the dilution of the batter as the viscosity increased – were attempted before arriving at the “fixed-time commencement of trials” formula. The potato masala that was used during the trial period was made manually.

Dosa batter

This decision regarding the dosa batter had a bearing on the plant location as well. It would be necessary to transport the material from the factory to franchises all over the country. Narayanan believed that the bulk of the market in India would be the non-traditional dosa-eating areas of northern and the western parts of India. A centrally located factory was critical if the franchises were to process wet batter with a limited shelf life. The factory location also depended on proximity to raw materials such as rice, dal, and potatoes.

Climatic conditions would also influence the fermentation process, unless the IFFL created standard conditions in which the fermentation would occur. The UDCT believed that a wet batter could be preserved after fermentation through chemical preservation. The IFFL needed to analyze the fermentation process and arrest the process by drying the product once it reached the right consistency and bacterial count. Since the fermentation process took over 12 hours, they needed to search for the right enzyme(s) that would consistently accelerate the process.

Potato masala and accompaniments

The potato masala turned out to be another major challenge for the IFFL. Potatoes from different parts of the country tasted different; further, the potatoes harvested at different times from the same location differed in taste. The IFFL decided to dehydrate the potatoes along with the masala and other ingredients; this would be reconstituted

This document is authorized for use only by AJ Roberto (aj.roberto23@gmail.com). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

Nur Demir
Dosa King: A Standardized Masala Dosa for Every Indian Page 6 of 8

at the franchises. In the long run, the IFFL wanted to use locally sourced potatoes and also develop a process that would ensure consistency. The IFFL also decided to dehydrate the sambar made in Mumbai and Chennai formulations and to develop a range of exotic chutneys that would be changed frequently. The IFFL wanted to negotiate with Coca-Cola, PepsiCo, Brooke Bond, Nestlé, and Rasna to address an important customer need – most people wanted a hot/cold beverage along with their dosa.

PROJECT RISKS

The IFFL realized there were several risks associated with the project. The dosa machine was being designed to deliver dosas from a batter that had not been standardized. The delivery of the dosa in a vertical format (which was to be eaten like a roll) assumed that a crisp enough dosa could be delivered that would not sag while holding the potato masala in place in the center. The consumers’ acceptance of dosas served with exotic chutneys alone (i.e., without sambar) was assumed, although the development of sambar was underway as a backup. It was assumed that a machine should be developed to make dosas weighing 90 grams each, although the industry standard was 175 grams. This decision was taken because the 175 gram dosa could not stand upright. The shell of the masala dosa could not hold such a large weight at the center of the dosa and still remain upright.

DECISIONS, DECISIONS, AND MORE DECISIONS

A factory location study was carried out to determine the optimal location of the factory; several locations were analyzed. Since transportation of the wet batter was critical, Nagpur was chosen as the location. The main criterion for the selection of the location was the availability of potatoes, dal, rice powder, and other ingredients. Development of the original dosa machine design undertaken with the soft drink machine manufacturer continued; a new drum-design dosa machine was under development simultaneously.

The UDCT was commissioned to develop a wet batter, while the enzyme route to accelerate and dry the batter was also being researched. A retort packaging facility was sourced from a Mumbai-based manufacturer, while a lab-scale drum drier was ordered to commence trials on the dried potatoes. The final order for the machines was kept open, partly as a result of delayed financial clearances and partly owing to the wait for the final information tally. The development of the chutney(s) was outsourced to a Hyderabad-based food manufacturer; the arrangement was to procure the finished and packaged product(s) from this manufacturer.

All options with regard to the accompanying beverages (where readymade solutions were available) were kept open. The IFFL made several other decisions related to the project:

1. It was decided to enter the test market with the original specifications for dosa size, weight, color, etc. since the final machine design was still under development. Any changes were to be made after the responses from the test market were analyzed.

2. By this stage, it became clear that the drum-design dosa machine was the preferred machine design; the original design was abandoned. All development efforts were now focused on the new design.

3. It was decided that dried dosa batter with or without enzyme would suffice for the test market. The trials run by the UDCT were simply not falling in line during “repeatability” trials. Moreover, the preservatives used in the batter gave a “non-food” smell to the dosa.

4. A mix of onions and spices preserved in brine was developed for the potato masala. This was to be added to the potato flakes that resulted from the dehydration process.

5. Four different accompaniments were used in the test market.

TEST MARKETING

The test market became the final validation and verification ground for the dosa delivery system. The test market results were as follows:

1. The dosa simply did not perform in the test market. All attempts to deliver crisp and consistent dosas completely failed. There was no consistency in two consecutive dosas.

2. Customers were expected to eat the dosa “on the go”; however, the actual dosas were limp and they sagged. Therefore, a mid-term switch to paper plates was attempted, which resulted in further reduction in traffic.

This document is authorized for use only by AJ Roberto (aj.roberto23@gmail.com). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

Nur Demir
Dosa King: A Standardized Masala Dosa for Every Indian Page 7 of 8

3. The demand for sambar accelerated. Once the compromise regarding the plate had been made, this compromise (additional sambar) was also made.

4. The dosas were in fact being delivered every 90 seconds; however, every third dosa would turn out damaged.

5. The retorted potato masala could not be tested as the product gave out a smell of chemical preservatives. Further development of this potato masala was stopped following the disastrous start in the test market, and the product was withdrawn.

6. Frequent power cuts resulted in interruptions to the entire dosa-making process and caused further destruction of the material being cooked.

IFFL’S RESPONSE TO TEST MARKET RESULTS

The IFFL responded to the test market results with the following decisions:

1. The size of dosa was revised to a larger size that looked visually acceptable at 110 gram. 2. The enzyme to accelerate the process of making the dosa batter was finalized and a consistent dosa batter

was achieved. 3. The combination of potato flakes and a brine mix for the onions was tested among consumers and was

found to be acceptable. 4. The software on the electronics was redone for the various problems that were inherent, including a sensor

that would start the process only when the machine had reached the right temperature. The location of the sensing device was optimized. In the earlier version, the sensing device was positioned at the center of the drum. The sides of the drum did not achieve the desired temperature, which resulted in the edges of the dosa sagging. By positioning the sensors on the edge of the drum, the crispiness of the dosa was better controlled.

5. The accompaniments were not really tested for acceptance, since the customers’ rejection of the limp dosas was nearly unanimous. The decision was to use the same combination of exotic chutneys for the final launch as well.

6. The breakdown of the machine was frequent owing to several reasons. However, the lack of a steady stream of customers did not permit sufficient number of readings for a clear understanding of the reasons for such failures. A degree of denial on the part of the engineers who had designed the machine also led to incorrect conclusions regarding the reasons for the breakdown.

7. The footprint of the machine was already in excess of 8 sq. ft.; this needed to be reduced to 6 sq. ft. 8. Delivery time of the dosa needed to be accelerated to 30 seconds.

Armed with this information and faced with increasing pressure to launch the Dosa King as well as the drying up of funds, accelerated development commenced for the final launch.

FINAL LAUNCH

A locally-fabricated drum drier for the potato flakes was deployed since the final drum drier would be delayed by at least 12 months owing to a delay in placing the order, which was an outcome of non-disbursement of funds by the financial institutions. During this interim period of development, several commitments regarding the launch date had been made to these institutions to ensure that funds would be available on time. A “launch or perish” mood set in at the IFFL and a final deadline of 6 months was fixed.

Narayanan recalled:

Despite clear indications that the machine’s performance was unsatisfactory, we went ahead with the launch with a maintenance crew of nine people against an initially planned maintenance crew of two who were provided with pagers to enable prompt service in the event of a breakdown. It is important to remember that cellphones were yet to penetrate the market in that time.