Recently Forrester Research released a report that says that offshore captives, which have been much in vogue for the last three years, are “imploding”. A blog post here makes for interesting reading, especially the comments.
I am not surprised. Captives were gathering steam in my last year at Infosys. There was this buzz around captives that created a feeding frenzy for management consultants and research analyst firms. We knew these were bad decisions but were helpless in the face of advice from firms that had the ears of the CEO and CFOs. Plus we had a clear conflict of interest. Captives were bad for our business and so we couldn’t be relied upon for good advice.
There are two sets of reasons why making captives for enterprise IT is harder than it looks (product development has different drivers and I don’t cover it in this post). One set are structural reasons – economic, employment market etc. Another set are around how to manage distributed teams.
The market for Indian tech professionals who work on corporate IT systems is very unlike the ones you will find in developed markets. I have written about it here and here in the past. The average wages will be low but the annual growth in wages could be well over 20%. IT Services companies offer the best deal for employees – career growth, leadership opportunities, international travel and a brand name employer (yup, Infosys’ brand beats any US corporate’s in India hands down). Bottomline – the captive will have to pay much much more to keep an employee compared to an IT Services firm – and will still have to deal with higher attrition.
Most IT Services firms in India are run very efficiently. Overheads if measured by $/employee are probably among the lowest anywhere. India based costs, economies of scale and use of technology in automating business processes are like a triple whammy on costs. On top of that, real estate costs are low because facilities are typically owned and leases are negotiated with state governments at favourable terms.
If you do the math realistically, a captive’s costs are well above an IT Services company’s, even including their sales & marketing costs and margins. So there really shouldn’t be any economic reason to go build a captive.
But the reason why captives fail is actually not economic. Its not as if these captives are being closed/sold/outsourced because of a small difference in costs versus the cost of outsourcing. The real reason why captives fail is because IT organizations don’t know how to make distributed teams work. Here are a few things you will find in how IT organizations work with captives:
1. They don’t have and don’t realize the need for stronger life cycle processes for distributed development.
2. Organization models put all decision making in headquarters and none in the captive.
3. Dull, repetitive tasks like testing or support that nobody in headquarters wants to do are the first to be shipped out.
4. Average experience levels in captives will be lower. Just the nature of the Indian market. Companies won’t invest in adequate training and then arrive at incorrect conclusions that employees in the captive don’t “get” the business side of things.
5. Small things matter. Like when conference calls are scheduled. If they are always during work hours at headquarters but outside of work hours in Bangalore, there’s a sign of dysfunction right there.
6. Headcount at the captive keeps growing, but the organization model is such that headcount at headquarters never comes down. The business case goes phut.
At the end of it, you’re left with a bunch of demotivated employees looking at an endless stream of dull, routine work and a cost reduction plan that never gets off the ground.
That doesn’t mean that distributed development doesn’t exist. You have but to look at any open source project to realize how well they can work. But unfortunately, enterprise IT still, by and large, hasn’t figured out how to hack it.