business diversity, diversification

Diversifying makes perfect sense - except when it makes perfect nonsense. On one hand, wisdom dictates spreading the risks, so that decline in one market or product, instead of sinking the whole company, can be offset by booming sales elsewhere. On the other hand, managers well versed in one industry easily get out of their depth in strange waters - and the company can drown in the consequent losses.

Boards never believe this can happen to them. So Dixons, say, goes boldly into launching Freeserve, despite the awkward fit between an Internet service provider and a High Street retailer, and achieving rapid success, while no doubt forgetting sad precedents. For example, EMI, a successful music business, produced a world-beating medical innovation in the body-scanner. The profits appeared to pour in at first. But the golden flood was illusory.

As quality and cost problems caught up with the company (and heavy competition arrived), huge losses forced EMI to lose its independence. This was an organic diversification, generated wholly within the company, rather than an acquisition. Buying a business generally increases the risks. True, you acquire a management team that presumably knows what it's doing. But you risk falling foul of a simple proposition: good companies are expensive, cheap companies bad.

Either way, you create problems - difficulties so serious that at least half of all significant mergers and acquisitions appear to have failed. To justify its costly price, the acquired diversification has to perform much better than before. That increases the chances that the new owners will interfere, and in a business strange to them, at that. They are very possibly damned if they do, and damned if they don't. Interfere, and the acquired management may walk out or, worse, stay on disgruntled. Don't interfere, and you may wake up one morning to find that the golden eggs are addled.

As for bad companies, they require new management. By definition, the diversifying business has nobody with expertise in the new line of country. The acquirers must either put their trust in an expensively imported wonderman (or wonderwoman) or rely on their own people to learn on the job. Either way, the acquiring board can't hope to exercise adequate supervision. That is barred by the very lack of knowledge that made the bad acquisition problematic in the first place.

There is one easy solution: fiddle the books. But whatever games people play with restructuring charges, goodwill treatment, etc., the economic realities of unsuccessful diversification will persist. Whatever the accounts say, part of the shareholders' capital has been destroyed - a paradoxical result, given that diversification is intended to enhance their wealth. That can happen: but the conditions are often freakish. For instance, the late Sir John Davis of the Rank Organisation used to boast of the hugely successful diversification which created Rank Xerox, a multi-billionaire baby, for an initial outlay of £300,000.

That was all a little New Jersey company named Haloid could extract for the non-American rights to a bulky document copier that embodied the first use of xerography - and to that clumsy machine's nimble successors. As the new technology surged forward, the Rank Organisation became in effect a one-business company, with its other diversifications mostly eating into the Xerox gold. Far from proving Davis's point, the Rank saga shows that a concentrated company, diverting all its energies to one highly profitable area of business, probably offers better rewards for shareholders.

Coca-Cola, though going through a sticky patch of late, is a clear proof. So is Microsoft - even though, at first sight, it seems highly diversified, with 186 products, ranging from encyclopaedias to browsers. Look closely, however, and you see that the diversifications all revolve around the core of the PC operating systems with which Bill Gates began. All businesses - even Coke - grow by adding new products, markets and lines. The crucial issue is whether or not the additions are extensions of the company's existing expertise.

Supplements are by no means the same as extensions. When British Aerospace bought Rover, wondrous guff was babbled about the transfer of aerospace technology to car manufacture. If any such transfer took place, it was totally insignificant. Similar false claims were made by General Motors when it diversified into computer services and electronic defence systems by buying, respectively, Electronic Data Services and Hughes Aircraft.

Both buys, however, proved to be enormous successes, earning GM large profits both from operations and from the billions eventually made by spinning off the buys. This doesn't disprove the argument against diversification. It proves something else: that GM is very good at this game (much better, in fact, than at making and selling cars). There are businesses which specialise in diversifying, including one - Warren Buffett's Berkshire Hathaway - which is among the most spectacular investments of all time; $10,000 invested in the shares in 1965 was worth $51 million in 1999.

Buffett has bought and kept businesses in furniture retailing, ice cream parlours, shoe manufacture, newspapers, insurance, executive jets, etc., etc. They now generate extraordinary returns on equity from $80 billion of assets, with 47,566 employees. Like GM's board, only more so, Buffett fully understands the principles of astute diversification. First, the buy must sell at or preferably below its intrinsic business value. Second, it must be a business you can understand. Third, it must have excellent returns on capital, terrific past results and predictably good future prospects.

Equally important, it must have honest and competent managers who are happy to get on with running the business without interference. Diversifiers may object that such conditions are rarely met. That is all the more reason for diversifying rarely. Buffett waits until something that meets his criteria comes along. Otherwise, he has better uses for his shareholders' money. Note that neither he nor GM proceeded by trial and error. They expected, and got, a near 100% record.

Note also that GM realised its jackpot, pocketing hard cash. The apparently unstoppable trend towards mega-mergers and massive acquisitions has been offset somewhat by de-diversifying, as large companies sell their past buys, good or bad. Sometimes the sales have been profitable (like BAe's offloading of Rover), sometimes not. They have very seldom been planned, however. It's simply that one strategy (or bunch of boardroom strategists) has been replaced by another, or that disposal is an obvious way of appeasing the investors baying for 'shareholder value'.

That can sometimes result in apparent lunacy, when a company like ICI spins off the jewel in its diversified crown - in this case, its pharmaceutical business. Warren Buffett, in contrast, has said, rather oddly, that he will never sell some of his businesses, even if offered far more than they are worth. The problem faced by ICI was that the stock market valued what is now the Zeneca drugs business at far less than its real worth. That is the paradoxical fate of diversifications: if they fail, they are a drag on the company: if they succeed, you cannot extract their full value from the stock market.

This awkward paradox is another reason for 'concentric diversification' a la Microsoft. This means extending the business into areas which all relate to a common core and existing strengths. Buying Hotmail, for instance, added a valuable e-mail service to the other internet activities on which Bill Gates had bet Microsoft's future. To take an organic case, diversifying into cars built on the engineering, marketing and distribution know-how which had made Honda a world force in motorbikes - and powerfully extended the brand.

Brand extension can be used for unrelated diversification as well. The spectacular example here is the Virgin group. Opinion is divided on the value of using the same brand for disparate businesses. It obviously helps with the initial recognition: if you know about Virgin airlines, you take Virgin cola more seriously than a no-name product. But Branson-style diversifications are only as good as the business, not the brand - and it's evident that great marketers often use new brands for genuinely new businesses. Forrest Mars would have been foolish to use his candy-coated surname for his stunning successful Petfoods venture.

That was another example of using transferrable skills and know-how from one activity (confectionery) to something quite different. Finding such opportunities, though, is as rare as discovering one of Buffett's beautiful buys. Making them pay off, however, has a Golden Rule. Don't try and run the diversification, organic or acquired, within the existing business. Establish the new activity on its own, with a self-contained management that has full autonomy and the resources required to create success. Otherwise the existing managers will very probably cripple the baby enterprise.

Theory is one thing, however, practice another. All diversifying managements believe (like Dixons with Freeserve) that their venture will be the exception that proves the rule, and that the elusive pot of gold will be theirs at the end of the rainbow. Dixons may, of course, triumph. But all too often, such beliefs prove to be triumphs only of hope over experience.