Re: Who ends up with IBM's business?
they looked at the costs and measured the quality and realised it was cheaper to do the work in house
It usually is. The outsourcers win the contract on the basis of promised savings, the customer signs after lots of vendor-to-vendor bid comparisons, yet carefully not asking any questions about the vendor's average labour rate, their overheads, contract set-up costs, current margin, balance sheet capital and return on capital, and investor commitments for margin growth. A few scant years later the customer is paying far more than they did in house, for a worse service.
That is the problem with professional procurement functions and finance teams. They're very good at comparing the price of apples with apples. Unfortunately they never consider that next year you might need a pear, or the consequences of agreeing a price for the apples is below the cost to the seller.
Outsourcing works when you want a pure commodity service that won't vary much, you're not fussed by the delivery standard, and you're equally not fussed by the price you end up paying.
Interestingly, this isn't just ITO and BPO. Big companies often can't procure commodity outsource deals competently. Big, multi-national companies I've worked for have been repeatedly screwed on catering and cleaning, because the same rules apply: The winner bids a price well below any sustainable margin, and then has to build up to that level, either by "variations" or by hiding other profit streams. In those commodity examples, the commonest way of squeezing out profit is by making sure that the contract is labour and materials - in which case the vendor either directly hikes up the prices, or more commonly negotiates bulk discounts with their own materials suppliers that are taken centrally, and don't appear on the site invoices that the client companies see.