When management is awful, fragmentation like centralising is no solution.
Few management issues are clear-cut. But one fundamental matter shouldn't be at all ambiguous. Either scale has economies, or it hasn't. Believers in the bigger-is-better formula could allow that, beyond a certain point, the disadvantages of size outweigh its advantages. The small-is-beautiful brigade, however, would always pitch that point far lower - thinking in hundreds where the rival school thinks in thousands, or hundreds of them.
Fragmentation has no fears for these advocates - nor, apparently, for the Major privatisers. British Rail is already being carved into two dozen passenger franchises, operating their services on railroads run by Railtrack, which will itself be split into 10 geographical zones. Their managers, to quote the Financial Times, 'will be given a high degree of devolved power to run their businesses', aiming at quarterly targets set by HQ, 'with financial rewards geared to the degree of success'.
Devolution and incentives are the nub of the fragmentation theory - one even older than Management Today, which was probably the first to call for splitting up to achieve clear visibility. Fragmentation theorists, though, go much further. They have the recipe for universal success. Break down the organisation, any organisation, into discrete components, place each component in the care of a dedicated single manager, hold said manager entirely responsible for the success or failure of this devolved enterprise, and reward or remove the leader accordingly. In the 1960s this was called giving a man 'a business like his own'. Success was scant.
Lou Gerstner, the new man wrestling with the formidable difficulties of IBM, explained why in an interview with Fortune: 'I don't think getting the economics right in a business is necessarily the same thing as decentralisation. There is a misconception that small is always more beautiful than big. Just fragmenting an organisation does not create conditions sufficient for success.' If it did, of course, small to medium-sized firms would always excel the large: but they don't.
In fact, any solution attempted at IBM will tend towards fragmenting the monolith - simply because its markets have fragmented, thus greatly altering the economics of the business. That is why one of the few remaining mainframe competitors, Unisys, is making two huge businesses (consulting services and open systems) more autonomous, 'trying to achieve the advantages specialised companies have', according to chairman James Unruh, 'without losing the advantages of economy of scale and the breadth of market of large companies'.
Never sneer at scale advantages. The planned building society merger between Leeds Permanent and National and Provincial will eliminate 149 branches and 1,600 staff. No theoretical or practical gains in managerial zest, speed of response, faster communication, or any other supposed benefit of smallness, could match those economies. Equally, any efficiency gains made through devolution on the rails look certain to be swamped by the diseconomies of fragmenting services that would otherwise be central. In fact, while the theorists have increasingly favoured fragmentation, the practitioners have found amalgamation much more attractive. Unisys, after all, put itself together (with the mainframe merger of Burroughs and Sperry) before starting to take itself apart. Though many mergers have disappointed - Unisys made huge losses for four years - that doesn't disprove the principle: other things being equal, the larger the turnover, the greater the economic leverage.
What other things must be equal? First, if overheads are allowed to rise in step with turnover, that defeats the object of the exercise - which is to push more revenue through the same pipeline. Second, companions in misery are no happier for being larger: the Unisys components were about to pass from decades of domination by IBM to years of agony adjusting to a post-mainframe world. Third, turnover must be reasonably non-diverse so that costs can genuinely be shared.
Diversity versus concentration is another fundamental issue that looks clear-cut - but isn't. The totally homogeneous, one-product company looks vulnerable: strategic diversification is the answer. But Coca-Cola rightly sold off its only major by-blow, Columbia Pictures, and has enjoyed an unprecedented run of mega-brand success. Royal Dutch-Shell has managed its homogeneous oil business with conspicuous, consistent excellence: but has recently fallen, yet again, into the chemical soup. Diversified, conglomerate groups offer no economies and no managerial certainties: yet their basic principle is fragmentation.
Each unit is supposed to have the 'high degree of devolved power to run their businesses', reported above, that Robert Horton (whose previous company, BP, was equally flat-footed outside oil) plans to achieve at Railtrack: quarterly targets set by HQ are the method that Harold Geneen made part of the exploded conglomerate legend of ITT: 'financial rewards geared to the degree of success' are meant to make head office supervision a breeze, almost superfluous.
The reality is that divisions are just as prone to strategic and tactical error, and to inertia, as separate companies. Central managements that like to back-seat drive and second-guess will always find excuses to do so - and often necessity will have the same centralising result. Gerstner is right. Fragmentation, like centralising, offers no management solution - other than visibility. The viability or otherwise (mostly otherwise) of each BR fragment will be much clearer. But that will do nothing per se to improve awful economics - or awful management.