The Hindu Business Line : ‘Rate per hour metric doesn’t help’

What is your hobby, you ask, hoping you can catch your first-ever round of golf with him sometime, but all he says is, “Only reading! I read all kinds of books. But I try to read more than just one book when I am on a topic.”

The CEO and MD of Infosys Technology, S Gopalakrishnan’s current interest is in Brain Plasticity.

You peer at him in complete incomprehension and he explains patiently. “It refers to the changes occurring in the brain as a result of experience. The brain is a continuously evolving organ.”

He says that unlike earlier believed, the brain is able to learn new things at any age. “Keep feeding inputs to it on a theme and it will absorb all it can.”

Interestingly, it also optimises and creates ‘space’ for itself by discarding ‘old’ things that are not repeated to it, says Gopalakrishnan, before delving into the problem at hand: the economic slowdown in the West:

It has been a year of turmoil for economies across the globe, impacting the Indian IT services industry as well. Worse, the outlook for the next year isn’t clear yet for players in this indusry. Here’s what Gopalakrishnan had to say:

Pricing has been impacted. But don’t you get down to negotiate on pricing only after you have exhausted all other levers, such as offshoring, etc?

No. It depends on the deal. In some of the larger deals, clients tell us that they have 100 people on project with an idea of the cost and if they hand it over to us then they would expect between 20 and 30 per cent savings. So they have a minimum expectation that they want us to better. It starts off as a rate game.

We move the discussion away from people — so we don’t talk about the rate per hour but overall savings.

Rate per hour is not a good metric, in my opinion, because the productivity from one person to another varies dramatically. Ultimately what anyone wants is cost per function point, which is a better metric, along with quality of software delivered.

These metrics are relevant. The latter will determine the total cost of ownership because you develop software over six months or a year and use it for 5-10 years. If the quality is not good, then you keep on paying for the lifetime of the product.

Maintenance contracts are more complicated. There, we have to analyse the volatility. Some pieces of software are all right. Others have a high number of maintenance tickets. If software has to be modified — then more you do it, the worse it becomes. This is called spaghetti code. At some point it becomes unmaintainable — so you write it again.

So we look at history — in terms of number of tickets. The surrogate is the number of people required to maintain that software. It’s not a simple metric since the experience of people determines productivity.

Someone with a lot experience can identify where the problem lies by just listening to the problem.

For others, it may take hours. Productivity varies dramatically. That is why rate per hour is a terrible metric to use.

Unfortunately, that is the easiest metric for purchase departments and deal consultants to benchmark on. All else is complex and requires understanding of software.

We are trying to slowly educate deal consultants and clients and move them to industry specific measurements and metrics rather than generic stuff such as just rate per person. Rate per function point, defects delivered, how many function points can be maintained by one person, maintenance ratios, volatility of software (how many requests come?), are all good metrics.

As negotiations move to pricing, we try to push for fixed price, because ultimately, what the client wants is a certain price. What we want is, within that price, an opportunity to make some money.

Rate per hour does very little for us. Whatever we do, productivity improvement, tools, etc goes to customer. Here, if we are smarter, we can also make some money. That is why it is better to move to Fixed Price

But, for fixed price contracts, isn’t maturity of the client important. It’s not meant for all clients…?

If clients have data and people who understand these things, it’s easy. By and large, clients are now starting to look at fixed price and deal consultants are willing to talk about these kinds of measurements.

Some customers are becoming smarter — they are including productivity improvement agreements, in addition to this, in the contract. So if $ X is the rate, to continue the same rate next year, the customer wants a 5 per cent productivity improvement.

If that is the customer mindset, why is pricing affected?

Pricing is affected simply because of competition. Volume has come down, clients want to save on costs.

If volume has come down and pricing is also lower, then all the levers that the IT industry has traditionally used become less and less effective in bad times…?

Right now, the only thing that really, really matters is growth. When growth happens, all other metrics will follow suit.

Having said that, we have managed our expenses very well, business is tightly run and we have been able to meet the margin expectation.

It feels good that we have managed currency movements and the drop in growth without affecting margins.

Next year, we are projecting a 3 per cent reduction in margin because: one, the impact of the revenue per employee drop will be felt whole of next year; and two, we are increasing our sales cost and the reason is simply this — normally when there is growth, automatically, sales cost goes up, percentage remains same but absolute value goes up. But even now, we want the percentage and the absolute value to go up.

But how did you arrive at the 1 per cent growth estimate in S&M?

We want an increase of 100 people in our sales teams across the globe…

But research reports tell us that companies that increase sales and marketing costs by 3-4 per cent during a downturn, typically do well.

In most studies, they look at US or western nations as the benchmark. For Indian companies — all these numbers are different. Our S&M costs are lesser than peers in the industry, because, we don’t put out advertisements etc. So, marketing costs are lower.

You will be honouring all offers made to campus last year, for this year. What plans do you have for this workforce? Is there a move to a solution-centric offering using new manpower?

If you take next year’s projections, we are projecting 16,000 would join from campus. Add another 2,000 experienced hires. That makes a gross addition of 18,000. Attrition would be about 10 per cent.

That means a net addition of 8,000. This year it was about 13,000 or 14,000. Utilisation is optimally at 78-80 per cent (Currently it is at 74.3 per cent).

You have to put it in context — divide the bench we have by number of centres (10) we have and by the number of business units we have.

Second, many of these 16,000 people at entry level would undergo extended training (Training extended from 3.5 months to six months).

They would start joining from July onwards, some would start in January 2010 and would come in next fiscal. We have also launched several initiatives to keep them engaged on internal projects – some open source projects within the company. We have to manage this manpower — that is the biggest challenge. All of us would have excess capacity.

It is not simple to rotate people through projects — clients won’t like it. This is not the time to have challenges in the projects.

But would there be a substantial change in solution offerings?

Most IP development happens with very small core teams. About 15, 20 or 30 people is the average team size. It has to have select, very good people.

Large numbers of manpower required, because we are deploying them across 8,000 projects for the company. So growth is the only answer.

All these offers were made last April or May when the situation was better. In our second quarter (ended September 2008), we saw 6 per cent sequential growth. Very different times — those.. There are some signs of recovery now, though — if they are real and indeed and sustained then recovery would come earlier than we think. All the banking companies, JP Morgan, Goldman Sachs…they are getting good margins. Some say they would pay back government money and once they start making money they will start spending it. If the recovery happens then we can use these people.

Optimistically recovery is next fiscal — 2010. Pessimistically, four years!

So there is no question of campus offers this year.

Nasscom has requested member companies to go to campus only in the eighth semester. This is not a response to what is happening today but to feedback from campus saying that students don’t focus on studies once they get an offer. They then feel that marks are irrelevant — having already got a job. We will all go to campus only in January-February 2010 — at that point we will have a better picture of what is happening.

Any new thinking on variable pay for the coming year?

We have not given increases, but when we budget, we will budget for variable pay. That would not go away. Because growth has come down, variable pay will also slump, but there is no change in the way we look at variable pay.

This year, how have you dealt with the bottom 5 per cent?

When we rank people yearly, we know the bottom 5 per cent — always. Typically we put them through a performance improvement programme which gives them six months to improve. The person could improve and get out of that programme. Two, he/she could leave on her own — 30-40 per cent do that and being Infosys Alumni, they get good jobs. Three, no improvement happens and we would ask them to leave. The number there is small. We are lenient, generally. There’s always a shortage of resources — if they make a good effort, we let them continue.

This time around, bottom is still 5 per cent and quite a few people actually left — because they knew this time it would be tough — the only change we made was we could not allow all of them to go through the programme — we got feedback from their managers about the real bottom, the real bad ones. We asked them to leave immediately. About 2,100 made up the bottom 5 per cent. Out of that, less than half only have been asked to leave. It’s a small number compared to the overall number.

For the first time we see a net decline in active client status — it’s down by 4 clients.

It has happened in the past also, we do get transactional clients (clients coming in for one transaction or so) — for example in consulting, Finacle, other services get such clients, too. We have 37 client additions — we are happy with that, and happy with the repeat business. It’s just that growth of accounts has slowed down significantly. If you look at the top 10 clients they declined in growth by 3.7 per cent, and the rest by 3 per cent. So when projects are getting over, they are not getting replaced. New clients are also not ramping up. That is the challenge right now. We are not worried about client additions.

So there is no concern on the client-win ratio for you now?

No. The Pipeline is ok — we are winning deals.

You talked about three transformational deals.

They are primarily from Europe and North Amercia. At least one is from a new client.

Do you expect turbulence in the telecom sector to benefit you in the form of increased offshoring? (British Telecom, said to be Infosys’ client recently laid off 10,000 employees).

I cannot talk about specific clients. But turbulence in a client industry does not automatically mean more business for our industry. Our top client came down essentially because of currency volatility. It came down in size also, but most of the slump was due to the currency.

The last time we met, you said Infosys would go after opportunities in the systems integration business. Its contribution to revenues has gone up from 2.8 per cent to 3.6 per cent of revenues this fiscal from last fiscal.

SI is slowly starting to yield results. We are also finding opportunities around Finacle. Around Finacle, you will need a lot of change management, training, hardware and software upgrade, etc. There is a lot of ‘surround’ work, which we got only sometimes earlier. Now, we are trying to go after that consistently.

Any positives we can spot in the macroeconomic environment. Your own manufacturing vertical’s contribution to revenues is up, Intel recently said that the PC industry had touched the bottom (From this, one inference could be that the PC industry would begin to grow and hence software purchases would follow hardware purchases.)

It’s all too early. If you ask me, manufacturing for us has gone up because our footprint is broad. We do engineering services, embedded systems, research and development, BPO, Product development and IT work. No other industry (requires) such breadth of services.

In aerospace, for example, we have significant back orders so the slowdown has not affected them. And luckily for us, our clients in manufacturing, in general, are not affected by the slowdown. But if something happens to GM or Chrysler, then it would have a ripple effect. So manufacturing is not out of the woods yet.

If the financial services sector has seen the worst and starts to recover, we may see construction and the manufacturing sectors taking a hit for some time before recovering. So, it may come in waves.

(But that is not predictable). In our client survey, for the first time, many of our clients have said that for this downturn, cost cutting is across the board, including cutting offshoring.

Only when recovery starts that phenomenon of offshoring would revive. Right now, the mood is all about cost cutting. No one expected this to be as bad as it is. Now that reality has struck, they expect the worst. It’s always like this, right? Initially, there is denial, then reluctant acceptance, then ‘the world has fallen apart’ and then ‘it is not bad as we thought it was’.

Right now, they are expecting worse things to happen and are not yet willing to invest in projects.

You must be happy with your China operations reporting a profit.

A lot of that has to do with cost management — in this environment, smaller operations are challenged — they don’t have scale to spread overheads. So it is cost management.

Industry watchers are of the view that even after recovery, the IT industry would at best reach early double digit or even only single digit growth?

Average growth may come down. But individual companies would see better than average growth. Innovative, start-up companies would also do extremely well — new companies would emerge through innovation. That is because innovation will drive change in this industry for the next 20-30 years, at least I believe that. I am bullish about the medium and long-term prospects of this industry.

bharatk@thehindu.co.in

Related Stories:
‘Clients don’t have a firm number to spend’
IT budgets in BFSI space to shrink 6-10% this fiscal: Infosys
Infosys forecasts decline in earnings in the year ahead
‘Too early to talk of a recovery’
Infosys China breaks even in fourth quarter
Infosys freezes wage hikes, hiring

More Stories on : Interview | Software | Infosys Technologies Ltd

Article E-Mail :: Comment :: Syndication :: Printer Friendly Page

Stories in this Section
With thee by my side...

Can’t see the top

Towards ‘citizen relationship management’

Sold on the Web

Positive step against AIDS

‘Rate per hour metric doesn’t help’

SME market is the place to be…

In defence of the computer

Another Indian game for Western players

Quiz

Growing a bank through the punched card era

Cartoon

For the SMB

More for e-governance