In the current parlance, silos are actually a relatively recent concept, at least as a derogatory term, for lack of communication throughout an organization or for restrictions in the flow of information from one business unit to another. After all, in the 1990s, people would occasionally discuss silos in terms of vertical integration. In such cases, organizations attempted more control over their supply chains so as to be less dependent on external suppliers or distributors, and the term silo accurately describes a cylindrical structure that keeps in the organization's money and keeps out those other fingers grasping for a share of corporate revenues.
(Of course, vertical integration does not always work well unless the organization is prepared to offer raw materials and/or production/distribution services as good as those offered by businesses in the supply chain with which it competes. For example, many businesses that twenty years ago attempted forward integration found to their dismay that shipping actually involves something more than buying a few trucks and hiring a few drivers. Ronin Ro—in his book Tales to Astonish—discusses just such a scenario in the publishing industry when he quotes former Editor-In-Chief Tom DeFalco's commentary on the Marvel publishing company's decision, in 1994, to purchase its distributor, a decision the company made "because they felt that they could cut out the middleman and make more money -- which was insane" [page 279]. Ro notes that Marvel subsequently declared bankruptcy on December 27, 1996 [page 282].)
Silos that developed as part of efforts in vertical integration have the effect of containing multiple business units and, in theory, of enabling interactions across departments. Indeed, if the marketing and retailing divisions are not working cooperatively, then the business will likely soon feel the disastrous effects of the disconnect. But the cooperation is in such cases aided by the integration. And many businesses continue to have in place vertical integration for at least a partial managerial structure. For example, as Daniel Lyons notes in the 2010 article "Going Vertical," Apple's iPad developed out of just such a model.
But these businesses employ silos to enhance the flow of goods, services, and—perhaps most importantly—information to all business units within the organization. The term silos became derogatory when it began to refer to a system that inhibited communication within the organization, a system in which—in this example—marketing and retailing do not communicate with the result that potential customers appear at a store to buy snow shoes but find instead only water skis. ("We're sorry. We just don't have any of those in stock.") The point is that silos contain things within a metaphorical shell that needs to be at least—but often isn't—semi permeable. In the case of information silos, what is contained is knowledge about and of use to many or all of the divisions and/or departments within the business. If that information does not flow, the business does not work, at least optimally.
Obviously, businesses plan—rightly or wrongly—to implement vertically integrated systems. They do not plan to implement information silos. So how do the silos develop?
In the article "A matter of trust and respect," Marcel Côté provides as concise an answer to the question as I have seen: "Generally, silos are an offshoot of decentralized management. Ambitious managers, responding to the objectives asked of them, pull those reporting to them along in their quest. As a result, their department's interest takes precedence of the well-being of the organization."
The decentralized business model has developed as predominant over the past twenty or thirty years. Prior to 1980, a more centralized model, no doubt strongly influenced by a military hierarchy, held sway. Since that time, a number of factors have combined to prompt organizations to adopt a more decentralized structure. First, organizations have grown so large and their segments so multi various that their implementing a single, command and control, authoritative hierarchy for all corporate segments is simply impossible. Instead, individual divisions have become industries in and of themselves, and those divisions demand specialized management by industry professionals. In addition, with the expansion of global markets, opportunities as well as competition have intensified. This competition requires constant scrutiny that a single centralized hierarchy cannot provide.
These large, separate divisions can and do function autonomously. Furthermore, globalization, curiously enough, appears to have inspired a trend toward localization in which these autonomous units are expected to respond to site-specific preferences. In terms of authority, finances, organizational structures, and much, much more, decentralization is an extremely powerful model. But, as Peter Cappelli notes in an excellent analysis from his book Talent on Demand: Managing Talent in an Age of Uncertainty, in terms of talent management, decentralization limits the ability of large organizations to locate, monitor, track, and develop from under the organizational umbrella clever and accomplished personnel who might easily, in a different era, have risen to a C-level manager position but who are, in the decentralized model, destined to remain with a much smaller division, if they remain with the company at all. More than likely, Cappelli asserts, such personnel will seek promotion by leaving the organization (page 232). This model has spawned a recurring scenario in which businesses emphasize recruitment and hiring more than they emphasize internal development, relying on the options to buy (through hiring) or rent (through outsourcing) personnel rather than on the option to make (or develop internally) personnel.
(In The Big Skills Gap and Talent Management, I discussed business' talent management options, as here, in terms of buy, make, or rent. Professor Cappelli discusses these talent management options in terms of only buy or make, combining hired and outsourced talent within the buy option. Obviously, I consider my divisions more reflective of the current business environment, though Professor Cappelli and I probably conflate some obvious categories, such as contingent labor. But that is a story for another blog.)
Although Professor Cappelli does not dwell on the point, it seems likely that organizations' current dependence on the buy option, as opposed to the make and/or rent options, stems to a certain extent from the structure of a decentralized business model. Larger organizations are simply not aware of the talent at their disposal because they are no longer unified entities but rather a grouping of departments, divisions, and other business segments often joined through mergers and acquisitions by parent corporations with no goal other than to add yet another profitable business segment to their operations.
Information silos and organizational reliance on hiring as its chief method of talent management have formed as a by product of this decentralization. Herein lies the problem: both the silos and the exclusion of internal personnel from the talent development process are unintended and, thus, unmonitored results of decentralization. It is difficult to imagine that a business would choose information silos as a method of advancing organizational effectiveness, but the same business might well choose the buy option as its method of talent management simply out of ignorance of possible alternatives. The point is that the business should make that decision based on evidence analyzed within a process rather than have it the unintended result from other, almost unrelated decisions.
The fact is that most large organizations already undoubtedly collect and have at their disposal large amounts of information related to their workforce. Employed strategically, this information could enable them to employ that talent more effectively. Employed restrictively—as within an information silo—this information serves little good in a wider organizational context. As organizations become aware of this structural problem and as qualified talent becomes difficult to buy—a scenario that appears increasingly likely—then businesses can take steps to remedy their current situations and to utilize their existing talent pool more effectively.