Why offshoring may or may not work for you?

Vinnie started it off with a link to a recent CIO magazine article. Sadagopan joins the debate and disagrees with the captive option being touted. In my view, I think the CIO magazine article is flawed not in the sense of the data being presented but rather the conclusions being drawn. Billions of dollars worth of software are being created using the Global Delivery Model and instances of successful relationships running for over 10 years are legion. I don’t doubt that it may have happened in the case study being quoted, but to say that after 4 years ROI goes down the tube because of a few cases is a hugely misleading conclusion. Let’s look at the 7-step solution being offered by CIO Magazine. Excerpt:

1. Don’t assume the hard work is over during the ramp-up period. Healthy offshore relationships require continuous improvement throughout their life cycles and extensive re-examination every three years.
2. Hold quarterly meetings with senior-level representation from both customer and vendor.
3. Rotate people out of the relationship management position after a few years to prevent burnout. Or consider having some permanently onsite at the offshore location. The travel required to manage relationships offshore can take a toll on employees.
4. Keep close tabs on vendor staff turnover. Pay attention to all levels, from the executive ranks to middle managers to line workers; defections at any level can have a negative effect.
5. Work with your offshore provider to cushion against the sudden loss of key employees. For example, on an important project place an extra resource who can be ready to jump in if necessary.
6. Make sure performance metrics mesh with reality. If they don’t, come up with new metrics that do.
7. Survey business users and internal IT staff about their experiences with offshored support or projects on a regular basis. They can be your best measure of how well or poorly the engagement is delivering.


The question I have is, why are these 7 steps any different for projects done with on-shore consulting companies or better still, in-house IT shops? To illustrate the point, let me rewrite the same 7 steps with some modifications.
1. Don’t assume the hard work is over during the ramp-up period. Healthy Business-IT relationships require continuous improvement throughout their life cycles and extensive re-examination every three
years. 2. Hold quarterly meetings with senior-level representation from both Business and IT.
3. Rotate people out of the relationship management position after a few years to prevent burnout. Or consider having some permanently onsite at the IT location. The travel required to manage relationships to the IT location can take a toll on employees. [Most companies have business users and IT people in separate buildings or in completely different locations. Agree international travel is a lot harder. But if the only way your projects will be done properly is by physically being at the location, its time to look change your IT group] 4. Keep close tabs on IT staff turnover. Pay attention to all levels, from the executive ranks to middle managers to line workers; defections at any level can have a negative effect. 5. Work with your IT to cushion against the sudden loss of key employees. For example, on an important project place an extra resource who can be ready to jump in if necessary. [Its standard project management practice to keep additional resources for contingencies] 6. Make sure performance metrics mesh with reality. If they don’t, come up with new metrics that do. [There is an old saying in IT, you can’t manage what you can’t measure meaningfully. Again an universal Software Engineering principle] 7. Survey business users about their experiences with IT support or projects on a regular basis. They can be your best measure of how well or poorly the engagement is delivering. I am not claiming that managing projects with an offshore component is easy. I am merely pointing out that its quite similar to managing IT projects in general and the techniques used come from traditional program management. In general IT projects are hard. In fact large projects are so hard that nearly 50% of them are cancelled. This statistic is from the renowned Software Metrics researcher Capers Jones.

Excerpt from his article “Software Project Management in the 21st Century
Table 1: Software Project Outcomes By Size of Project
Size Early On-Time Delayed Canceled Sum
1 FP 14.68% 83.16% 1.92% 0.25% 100.00%
10 FP 11.08% 81.25% 5.67% 2.00% 100.00%
100 FP 6.06% 74.77% 11.83% 7.33% 100.00%
1000 FP 1.24% 60.76% 17.67% 20.33% 100.00%
10000 FP 0.14% 28.03% 23.83% 48.00% 100.00%
100000 FP 0.00% 13.67% 21.33% 65.00% 100.00%
Average 5.53% 56.94% 13.71% 23.82% 100.00%

As can easily be seen from table 1, small software projects are successful in the majority of instances. The risks and hazards of cancellation or major delays rise quite rapidly as the overall application size goes up. Indeed, the development of large applications in excess of 10,000 function points is one of the most hazardous and risky business undertakings of the modern world.


The global delivery model is not for the faint of heart but it gives plenty of advantages to ones that know how to use it effectively. If you pick the correct companies and more importantly cultivate an open mindset within your company, it will beat your expectations and then some. If its not working for you, first look at your general IT project management practices and you will find the answer. One of my favorite clients was asked by someone “Does offshore work” and she responded “It works exceeding well for me and it will for you if you have the commitment to make it work”. I rest my case.
Technorati Tags: , ,


Comments

  1. Anonymous said April 21, 2006, 7:29 am:

    agree with you, the ramp and the downturn can happen in any project – GDM, internal etc. – I mentioned that in my original note. Howver GDM has not had as much scrunity especially in upfront transition investment. With deal sizes increasing and GDM rates now escalating much higher than other IT labot mdoels, expect more scrutiny. On the year 3-4, I do think given higher staff turnover, GDM is increasingly more susceptible to that…

  2. Anonymous said April 21, 2006, 1:34 pm:

    Thanks for the comment Vinnie. You have a lot of experience and I am sure you are seeing problems in the 3-4 year timeframe. Only explanation I can give, which I said already, is that it probably depends on the company you are working with. Maybe I have been fortunate to have not come across major issues in the 3-4 year timeframe.

  3. Anonymous said April 23, 2006, 5:39 pm:

    As the author of a recent book on Offshoring IT Services, I would like to buy into the views in your blog. While I agree with some of the points being made, the statement “Billions of dollars worth of software are being created using the Global Delivery Model and instances of successful relationships running for over 10 years are legion.” sounds a bit too presumptuous.

    Of the few billion dollars that offshoring companies are generating in revenue, a large percentage comes from services like maintenance, system integration, testing and the like. ‘creating billions of dollars in software’ may stretching the statistics.

  4. Anonymous said April 25, 2006, 5:11 am:

    Mohan Babu,

    Thank you very much for visiting and commenting on the post. Congratulations on your new book. Would love to read it. I have added it to my to do list.

    I think you are pulling my “billions of dollars of software being created” out of context and interpreting it literally. Anyway, I wrote those words and I guess you have your right to interpret. I had meant it as a way to say that this is a big industry producing a lot of value and a few cases of failures around the 3-4 year mark don’t make a pervasive trend.

    Also, “creating software” doesn’t always have to be creating from scratch. Maintenance involves a lot of development in the form of enhancements which are often treated like any other software development project. So I don’t agree with your narrow definition of “Creating” = “Development from scratch”. In fact, in today’s world, when you use so many component frameworks, one could argue that even software development is a maintenance of sort.

    Please go to the graphic in this page on the NASSCOM site. It shows 2004-2005 industry revenues to be 12.2 B. Of these just the top 3 (TCS, Infosys, Wipro) are each doing over $2B in revenue in 2005-2006 for a sum total of approx $7B. Anecdotally, these 3 companies do anywhere from 40-50% of revenues in Application Development Projects. You work for Infosys, so please comment on whether that figure of 40-50% is accurate from Infy’s experience. So being very conservative, if we say 30% are development projects (“creating software in your definition”) the total comes to $2.1B of revenue. If you apply the same ratio across the industry you get $4B (30% of 12.2B). If this is not “billions”, what is? And the industry is growing 30% atleast, so that $4B in 2004-2005 numbers is already $5B in 2005-2006.

    Also you say “of the few billion dollars that offshore..” I would love to hear your definition of few. When does “few billion dollars” become just “billlion dollars”? Or in other words is <5B few, or is it <10B that is few or is it <15B?