Management consultants don’t have the most stellar reputation for offering value for money – although, truth be told, the infamy of investment bankers has long since left them in the shade – so I was impressed when, early this year, rumours reached me of a fascinating new study in which a consulting firm was implicitly agreeing to subject its advice to a randomised trial.

Not to put too fine a point on it, that’s a ballsy decision. Almost the only people who ever agree to test their stuff in a randomised trial are the pharmaceutical companies, and I am fairly sure they’re underwhelmed by the experience.

Perhaps Accenture (for it was they) didn’t realise what they were letting themselves in for. After all, the study – conducted by a team of economists at the World Bank, Berkeley and Stanford – was ostensibly about whether modern management techniques would improve the productivity of large textile firms in India. (A draft, “Does Management Matter? Evidence from India”, is available from the website of co-author Nick Bloom at Stanford.) But how do you improve management techniques? You send a management consulting firm in.

The researchers hired Accenture to provide management consulting services to 14 factories in Mumbai, chosen at random from a group of 20. The other six factories served as a control group and received a diagnostic performance audit but little serious advice. The World Bank and Stanford paid Accenture at a heavily discounted rate; the factory owners paid nothing.

The results were undeniably impressive. The effect of a few months of consulting advice was that profits rose by almost a fifth, to the tune of several hundred thousand dollars a year. (The latest draft of the study puts the figure at $230,000; an earlier calculation reckoned it was even higher.) Output was up, inventory was tighter and defect rates were halved.

Accenture’s fees for the five-month consulting gig, at commercial rates, would have been roughly the same amount as the increase in profits – so the arrangement would have paid for itself by the end of the year. If any of the consulting advice stuck, this would have been a fantastic investment. The indications are that the advice more than sticks: the new procedures generate more information, more ideas for running a tight ship, and a spiral of continuous improvement.

But before we all rush out to hire a management consultant, two notes of caution. The first is that as the experiment was designed to evaluate management techniques rather than consultants, Accenture was left in charge of collecting its own performance data. (The data was subject to some independent checking, however, and there is no evidence that anything was amiss.)

The more substantial caveat is that textile companies in India have their own distinct problems: tools and machinery were left lying around, and stock control was frequently non-existent. If a worker needed to find a particular item, the technique of choice was to forage around in the storage bins until something useful emerged. Modern inventory-management techniques made a big difference in this particular sector of Mumbai’s economy, but that does not mean that Vodafone or Barclays or the civil service have as much to gain from bringing the consultants in.

Nick Bloom and his colleagues have also studied management practices across the world, and Brazil, China and India rank poorly. Just imagine what these emerging giants will do once they get the consultants in.

As for whether management consultants can earn their keep in London or New York – I don’t know the answer. But I know a few economists who would happily supervise a randomised trial.