Sri Lanka’s budget BPO business model

Posted on July 6, 2011  /  5 Comments

We’re generally in favor of budget business models, but export industries that do not earn enough per employee to pay a decent wage have to be an exception. Here is some analysis I did on the Sri Lanka software export and offshoring BPO industries, based on official figures:

The total software earnings of USD 294 million are produced by 27,000 people. That is LKR 99,825 per employee per month. Lower than I expected.

On the BPO side, 13,000 people produce USD 98 million. That boils down to LKR 69,103 per employee per month.

Now that does not look too pretty, does it? Taking overheads and marketing into account, the firms in the BPO industry would be lucky to be able to spend LKR 40,000 on compensation per employee per month. That’s an average.

The full column is here. The preceding related column, which pointed out that Sri Lanka seems to short of people with skills is here. The two issues are related. Why we do not have enough people is because we do not pay them enough. Why we do not pay them enough is that the revenue per employee is too low.

One reason we cannot increase revenue per employee is because no one can get 5 nines (99.999) reliability in Sri Lanka. Good KPOs can work around this, but not the standard BPOs. And increasingly people are beginning to assume that video conferencing is possible. But how reliable is that in Sri Lanka?


  1. You just point out a problem. No solution. How typical Sri Lankan. Why don’t you tell us your solution? Bring RW back to power?

  2. The very nature of a column that is read by people is that it is not a comprehensive consultancy report. It focuses on one aspect of a phenomenon.

    Other aspects, including solutions, were discussed during the election period, which was the most appropriate time for that discussion:

    I have also stated that the purpose of my writing is not to give solutions per se, but to make people think about the choices that have to be made, the importance of tradeoffs, that one cannot both have the cake and eat it:

    I am working on pulling all this together into a book. But until then the columns are what exist, with their natural advantage of brevity and natural disadvantage of not encompassing all aspects. That too is a tradeoff some Sri Lankans have trouble appreciating, it seems.

  3. I think the revenue per employee argument is flawed as companies typically consider Sri Lanka as a cost center. Therefore the revenue reported for the LK entities are simply a cost+ figure rather than direct billing to end customers.

    In reality most software companies at least have much higher revenue per employee than 100K LKR a month. Revenue per employees will only go up in LK if the competition for the employees increase. And it will not increase unless we have a larger skill base to attract larger companies to set up here.

  4. This is an interesting and useful response. How do captive BPOs keep their books? They would, of course, have no marketing costs etc.

    But anyone who runs an organization will know there is a cost to an employee. I would suggest that this could be in the range of USD 500 per month in a modern organization that provides computers, air-conditioned work environments , health insurance, parking, training, welfare facilities and so on. If it is a very large organization, the per-employee could be lower.

    So even in captive BPOs, I suggest that the minimum per employee should be in the range of USD 1000, higher than the average as calculated.

  5. I could be wrong, but my impression is that most software firms are captive (so the cost center argument makes sense) while the BPO sector is a mix of captive vs. non-captive (HSBC, Aviva type organizations running captive operations, while many other local BPO operators doing contract based work). For the non-captives, the revenue per employee makes more sense. Do we have more people working in the captive or non-captive and what’s relevant revenue figure for each? Hopefully data from Nat’l ICT Workforce Survey will answer this.
    The workforce survey included IT training organizations as well, according to the new releases. These I assume are serving a predominantly local market?
    Also, looking averages are misleading – should look at mean (at least), I think, to be able to say anything useful. Specially if the the different sub-sectors are at the extremes of the cost (or earnings) per employee spectrum.
    Anyway, having the Survey data would be very useful…