Happy Days!

The stepson of IT will now have his day. Till 1998, when RelQ

first set up shop, testing was very much the untouchable, the ignored, the

lonely. He is still unglamorous, but changing trends will soon make him

fashionable, many vouch. And India will be its unmistakable lounge.

The US and Europe first came to India for manual testing. Then

it was for automation. Now, it's time for complex software testing and

value-adds. And it's also time for growth as the strong offshoring current

promises to wash ashore 70% of the world's outsourced testing market of $6.1

bn. The changing trend, we referred to, is mainly in terms of customer mindset.

They are able to save 70% of their costs by outsourcing testing to third party

vendors. And in a highly fragmented market, where putting a cap on the market

size is little more than a 'guesstimate', some analysts have quoted just the

current Bangalore market at $400 mn-the city has around 20,000 test engineers.

Some of Indian biggies have ramped up their testing teams

significantly over the last one year sniffing good business, even setting up

separate units after realizing the value derived by its clients from independent

testing services. So long the forte of small companies, the entry of big guns

into independent testing is a real test of character for many, experts say. The

changing market dynamics have already resulted in consolidation-Disha merged

with Aztec and Cognizant was in talks with RelQ not too long ago-increased

investments, not just in processes but also into training, and frenzied searches

for quality manpower. (There is no curriculum on testing in our engineering

colleges.)

And there are assertions galore. The smaller companies feel the

big guys have no depth of testing focus, busy as they are sniffing the next $5

bn opportunity. The big brothers tell the young cannot scale-it has to be

acquired to grow or remain small forever. On the other hand, the mid-sized

players, with both the focus and the ability to scale, simply smile, imagining

being the winners.

Managing growth, it emerges, is now the single biggest challenge

before Indian testing players, as they try to balance customer expectations with

realistic automation plans. There is a great business for everybody, but someone

might just loose his shirt.

Small Is Cute, But...



Some of the large-scale companies into software development and services-into

maintenance, application development and outsourced R&D-added one more

product line to their services, independent testing. It is one way of

capitalizing on their existing customer base to get more wallet share. But what

happens to smaller companies like Stag Software and others, who are into pure

play testing?

It sees increased competition (when Stag currently pitches in

for a project, it usually has four to five players on the competition table) but

none that it cannot overcome. "Specializations should happen in terms of

domains and there are companies even today that specialize in specific verticals-banking,

usability testing, load testing-and they will continue to do so," says

founder Ashok T. Somewhere down the line, he feels, if you are a smaller

company, the pitch will be in terms of what you have to offer in a particular

domain. The answer will be in becoming a niche player in a certain business

vertical or in a test vertical.

When bigger players capitalize and smaller players feel the

heat, a certain amount of consolidation will happen. Though no big consolidation

moves have happened just as yet, there are enough indications of challenging

times ahead for smaller companies. "Smaller companies will remain stagnant

or go out of business because they will lose employees who would want to work

with companies that grow. A lot of people come to us and say they are doing

boring work. You can focus on a niche and survive, but you will stay small and

uninteresting," says Samir Bodas, CEO of Disha, now an Aztec company.

When Disha was thinking what to do next, sometime early 2004, it

realized it should either be a consolidation or a dash to the finish. It had

several options on the table and one of them was a term sheet from a VC stating

that it would put money into acquiring some testing companies. "We

evaluated a lot of testing companies and we didn't find the right DNA, the

right quality of company that can be build into a 100 mn dollar test business.

The second option was to get acquired," says Bodas.

Discussions with Aztec started in April last year and the

announcement came in September. "The reason why we did not entertain

discussions with large players was because it would be very difficult for them

to focus on a $50 mn to $100 mn business, when they are gunning for $2-3 bn

dollar revenue. It will always be an interesting niche for them but never a

focus area," he adds.

Aztec is adding 100 people every quarter and more than 50% of

this are testing people. Its total test strength is now close to 1,100 and the

company services some of the largest ISVs in the world-Microsoft, HP, Yahoo,

and AOL-among others.

Indeed, getting the right people is a challenge that few

companies can ill afford to ignore, more so the smaller ones. The capital

required to raise a huge army just might not be there. And without people, there

would be no business. But getting people with good knowledge of testing is a

broader and more important issue for the industry right now, handicapped with

'zero college curriculum' on the subject. "You need a large investment

internally or work with universities and colleges so that you get the right

pipeline," says Bodas. Aztec has made a very large internal training

investment and its R and D groups are working with universities, sponsoring

internships.

An even bigger problem, perhaps, for smaller companies is

attracting the mid-level talent. "You need to have a good pyramid so that

good talent sign up. That is where a lot of smaller companies have failed at.

The limiting factor here is not necessarily people. It is how well you want to

build the organization," he tells.

LogicaCMG is a good example. It has plotted career paths that

can potentially take a consultant over time from basic test execution to

managing director, with stops on the way. For example, an employee there can

explore specialist skills development, consultancy, service management and

sales, besides making testing a genuine career path with a huge dose of variety.

But even with training, maintaining the right competency levels

of staff is quite a challenge. Testing expert VK Kripanand says the testing

processes in most organizations are not empowered. The top management has to be

more sensitive to quality and one measure that helps in it is the testing

process. "Support by the top management to empower testing teams to put

critical stops at relevant times will be crucial for delivering quality

products. Market demand, competition and revenue considerations force management

teams to ship out low quality products that are not certified by test

teams," he notes.

Testing houses in the past have employed newcomers in the

testing domain and thus have not been able to handle situations as well as it

should have. "Testing is a critical activity and demands creativity and

experience. The industry is catching on and a lot of domain experts are now

looking at testing as a career option," he adds. It is important that the

industry recognizes this fact and encourages a transition path to help itself

evolve.

Problems of scaling up and getting the right people might be

addressed if the market opens up further. But there is also a need to get

well-tuned sales personnel as of now. While survival could be tough, smaller

companies do not agree there is no future for independent testing. "In all

the accounts we deal with, we have not encountered any of the large players till

today. Second, the market is opening so much, if I had another 400 people today,

I would have business for them," says Dr Prakash Mutalik, a group president

at RelQ. Most of his clients, he says, prefer dealing with independent companies

for a simple reason: most accounts RelQ deals in have multi-million dollar

software development involved and handing over the job to the same people who

have developed the software wouldn't ensure an unbiased job. This viewpoint is

backed by some companies like Lionbridge, which says that customers who look for

independent verification and validation services usually turn away from the

vendor providing the coding services for the application. "Say, if Infosys

is providing the application development services for an application, for the

testing part customers prefer to look for a vendor who is not the developer of

the application. It is here that small independent companies gain

respectability," says a Lionbridge spokesperson.

This contention will not hold water always, as testing teams at

big companies like Infosys is known to report directly to the Quality Assurance

management of its clients and have no reporting line into the development side

of projects, thereby providing clients an impartial assessment of production

readiness of the solution.

Give Me More



Meanwhile, a different challenge for testing players is the rising customer

expectations. They expect SLAs, and more value-added work out of Indian vendors.

Things like performance engineering. However, performance engineering is a very

specialized task, needed in the last three to six months of the product

lifecycle. And it is difficult to find these specialists. "What ISVs want

from us is helping them architect for performance, test for performance and then

do performance engineering. They want security testing, penetration testing,

which is to see if the product is susceptible to the top ten vulnerabilities

people are aware of. Sustenance engineering is another area people are

interested in," says Bodas of Disha.

To keep on getting business, companies perhaps need to think

about testing from an engineering point-of-view, rather than service, as Disha

or Stag is doing. "The people in large companies are maintenance people;

they take huge batch of test cases and run them again and again. But customers

will get stickier and give you higher value-based pricing if you do

value-adds," he adds.

Stag's Ashok T says his primary interest is to see if his

organization can come up with techniques and methods supported by tools that can

guarantee the quality of a software. "The intent is for a more scientific

way to test software. Over the last five years, we have applied various methods

for customers in many domains. We have distilled the whole thing and called it

Stag Test Engineering Method. This method will take us close to promising that a

software will not have any major issues," he says.

Another role that customers now take for granted from the Indian

vendors: advisory services. This is more true in cases of automation where CTOs

and CIOs need to be educated at times. Customers and some development teams

often have a misconception-automating is easy and inexpensive and it is

inexpensive to run, but you can't automate till you perfect the manual

testing. Initial effort in automation is very high and outfits like RelQ have

100 people who work solely on automation. Since licenses are expensive, one of

the first things a vendor has to tell his customer is what not to automate.

Though it is true that automation improves productivity, it can be misleading to

say that it is the only strategy to release products early. Depending on the

nature of the application, the amount of reusability and the architecture of the

application, a testing automation strategy could be worked out. "Companies

that are less experienced in this domain are the ones who probably get

misinformed and make plans that under deliver or over promise," says

testing expert Kripanand. The contribution by testing automation tool vendors in

this area is now happening, thus leading to a better situation than in the past,

he adds.

Testing houses are always challenged with the varied

requirements of end users. The nature of applications and the technologies used

is the biggest issue for automation. "Consider an application developed for

a BFSI segment. The technological considerations here are quite high: some may

be on Microsoft technologies, others on Java. Even some may be based on Internet

technologies with multi-tier set up. The testing automation needs for these

applications from a domain perspective are the same, but the automation

processes vary on the technical front," explains Kripanand.

Organizations are thus faced with the challenge of capitalizing

on this knowledge gained and the putting the same to use in future projects.

"A solution area that some companies are working on is on building a

testing framework for various technologies/domains with a lot of reusable

components put together," he informs. This when matured will help

organizations put together a testing project more rapidly in future.

But problems with automation for smaller players go beyond

technological considerations. Lionbridge says monolith test automation tool

vendors always target the end-user market and tend to be unfavorable towards the

middle-tier vendors who offer testing services for end users. This results in

small and medium-sized organizations not being able to offer the tool plus

services as a package to end-users, which results in high-priced solutions being

offered to end-user customers. For example, some tools are site centric and

project centric, as per the terms of purchase. So, a license purchased for a

particular building of an organization for some project cannot be used for

another project in the same physical building.

Managing director of Mercury India T Srinivasan, however, says

as long as the user understands that there is an RoI involved and he gets back

returns, the challenge disappears. "The challenge is understanding the

applicability of these tools, how they play in the application deployment and

change lifecycle, and the ability to get the expectations right in terms of what

can be automated," he says.

When one talks of process delivery, for example, automation

lends itself well because it gives a set of skills to hold good in terms of the

process of quality itself. "It is a combination of the implementer and the

customer's understanding of the need to get automation done. It's a

challenge on education, challenge on the implementer, challenge on the

environment, and challenge on the customer's mindset to get it done. You

mitigate the chances of failure by automating," he adds.

Indian testing houses, he says, shouldn't fall prey to

throwing more bodies of problems. The ability to raise the volume of automation

is critical, he says, because customers are looking at enabling the quality

processes rather than just getting low-cost services. "In the long term,

just as some European IT companies have become vulnerable to offshore companies,

Indian test houses are likely to become exposed to new entrants if they sell

purely on the cost angle. These companies will have to continue to move up the

value chain so that price is not an issue-and that requires investment,"

explains Alex Garrick, general manager of managed testing at LogicaCMG.

There are subtle indications of people in testing preparing to

do exactly that. The industry is witnessing a sea change in its commercial model

and methodologies. The hope is, it would be an exciting place to work soon. With

more people flocking this area, it would effectively ward off geographical

competition from Eastern Europe and China. And independent companies can hope

for more glorious ramp walks.

Goutam Das

The T-Plans

Who is going after what

Disha-Aztec: wireless, mobile and telecom; embedded

device driver testing; storage, networking

RelQ: corporate application, web applications etc; BFSI,

wireless, real time and embedded systems; software games

Stag: enterprise business, which comprises e-learning,

banking, healthcare and ERP; embedded and telecom

Lionbridge: performance benchmarking services,

certification services, localization testing services, and automation testing

LogicaCMG: financial services, telecommunications

Infosys: banking, retail and healthcare, energy, telecom,

transportation

Wipro: banking, embedded system testing in telecom,

manufacturing, IT and BPO, healthcare

Applabs: telecom, pharma, e-learning, software, hardware,

financial, media, airlines

On Test Automation

Lack of clear objectives on what test automation can achieve, ad

-hoc test automation efforts and poor testability of the application under test

are the challenges facing wide adoption of test automation. Test automation can

be built for many reasons (reduce lead time, increase coverage, ensure quality);

clear objectives will help the test team to choose the right strategy.

Test automation projects need to be managed very similar to a

software development effort (Conceptualize, Design, Build, Test and Deploy). The

degree of interaction (testability) between the test tool and the application

under test plays a crucial role in building robust test scripts which will

achieve its outlined objectives.  In order to build a testable application,

the test team needs to work in conjunction with the development team to build

testability as part of the product requirement.

Madhumurthy R, vice president, Technical Services, Applabs

Technologies

The Threat Perception

India is facing some competition in testing from Eastern

Europe, and China might be a rival some years from now. Where do these

geographies stand?

Eastern Europe

Some European companies do consider Eastern Europe-the old

Soviet Union, Romania, Hungary, Poland, and Czech Republic among others-for

some cost arbitrage factors.

The advantage: Strong education/ right engineers

The challenge: No critical mass of engineers

There are also sovereign risks with respect to IP and rule of

law. Also, companies from Eastern Europe have no great track record. Enterprises

often depend on a good and known individual for a project as opposed to a

company or country.

Have Indian independent validation companies lived up to customer expectations?

Ashok T, Stag Software: The fact that large companies now

have their own independent verification and validation groups, tells us that

there good business growth. If the servicing of business has been done well,

then these groups should grow. Some of the standalone-small and large-companies

have equally grown. The number of companies in independent testing has also

increased.

VK Kripanand, testing expert: Testing companies usually

are faced with the challenges of identifying the right expectations. All these

years, testing was considered an-end-stage activity. Some of the best practices

of software testing are to start early and validate all the time, when

development is happening. A heartbeat connection by the testing teams to the

requirements analysis, design and implementation teams are critical.

Many organizations test a final product against a set of

specifications. Areas like functional testing and performance testing are the

ones that are being outsourced majorly, as the infrastructure and resource needs

are quite high. The concept of requirement-driven development takes the level of

testing to a higher level and this is a concept that has to be embraced by

testing companies as well as by the companies outsourcing work to testing

companies. This, we feel, will improve customer expectations than its current

level.

Alex Garrick, GM of managed testing, LogicaCMG: The

global outsourced testing business is undergoing a revolution, moving from a

pure time and materials model to a fully managed approach. Indian independent

validation companies have come from nowhere to achieve a significant presence in

the domestic market. The question now is whether they can compete

internationally like their larger development counterparts.