Herding cats: a new model for management


Unilever is up against it. In 2003, it needed 234,000 employees to net $3.5 billion. P&G needed half that number of people to net nearly twice as much ($6.5 billion).

One strategy is to rationalize the famously sprawling structure of the company. This is a good thing. Far flung companies are not just expensive to run. They can be ponderous and unresponsive.

Another strategy is to insist on global approaches. Country managers for Dove are no longer able to modify packaging, formulation or advertising. Knorr’s soup and bouillon cubes, a $2.7 billion annual business, is no longer produced and designed as locally as it once was.

Economies of scale are a good thing too. But you can’t read today’s Wall Street Journal without thinking that Unilever is playing out a traditional view of management.

Simon Clift is head of marketing for Unilever’s home and personal care division. In the WSJ piece, he says, that there used to be so many local players in the new product development process, it was “like herding cats. There were no strategic priorities at all.” Clift has imposed a new regime: “Recently a major region wanted to launch a new packaging for Ponds’ face cream. We said, ‘No.’ Gone are the days when you can decide packaging locally.”

The trouble is that some consumer taste and preference comes increasingly from its local context. (We may define “local” variously: national, regional, urban, ethnic, linguistic, cultural, subcultural, neighborhood, class, age, lifestyle, etc.) In this various marketplace, giving the local player some control over the formula, the package and the advertising is probably a good thing. Ironically, Unilever appears to move away from this approach just as the rest of the world is thinking hard about whether “herding cats” might actually be a new model for management.

There is a “paradigm problem” at the root of this. The old management model was an Enlightenment project. The manager sought rationality for a corporation that, left to its own devices, devolved into superstition, localism, the unsystematic, and the increasingly variable. Efficiency escaped this organization like heat from a New England barn. There were many enemies of great strategy. Local knowledge and practice was one of the most pernicious.

The new model says that the corporation must forsake its Enlightenment pursuit of single rationalities and find a way to respond to the variability of the marketplace. How does the corporation read and capture local variation and build this back into the products and services that go to market? This will be messy and difficult. It will take a thorough rethinking of some Enlightenment impulses and a reinvention of the management handbook. But it will happen. Local variation is growing, and it is now one of the great competitive opportunities in industries, sectors and product categories where just about everything else has been tapped.

The “herding cats” model of managements discourages some aspects of ‘the vision thing,” and the idea that corporations must be run with a great thundering idea that come from on high, obliging every subordinate to demonstrate fidelity to Rome, and make themselves the vessels through which the idea passes—in one direction only. (Actually, the Catholic church was pretty good in some cases in absorbing local practice and there may be something we can learn here.)

We are now looking for something a good deal more dialogic, where the locality streams intelligence back from the periphery and accommodation out into the world. No, this won’t be that awful California “sharing and caring” regime that says that everyone must be heard and every idea is precious. The dialogic is sometimes mistaken for the idiotic, but let’s not make that mistake here.

The solution has to be some center periphery model that says, first, that the center is constantly fed intelligence and inspiration from the periphery, and, second, that the centre responds with something like “strategic vistas” that establish parameters within which the locals may make choices. There has to be a fail safe provision here as there is in the Delphi system. Locals must be given the option of insisting on their own course of action but these exceptions must be very carefully justified and they must pay out if the exception is ever to happen again.

The corporation is the most nimble, adaptive, transformative thing in our worlds. But it is only beginning to master the full range of its responsiveness.


Ball, Deborah. 2005. Despite Revamp, Unwieldy Unilever Falls Behind Rivals. WSJ. January 3, 2005, pp. A1, A5.

3 thoughts on “Herding cats: a new model for management

  1. steve

    I am going to say this in a much stronger way than I actually believe, just for clarity’s sake: Herding cats is the job of the market. The corporation’s job is to be an effective cat.

    Absent the “Enlightenment” idea of a coherent and focused organization, the corporation’s management structure is an overhead that cannot justify its cost. I always get itchy when management pundits start ragging on the Enlightenment, or “Newtonian thinking” or “Taylorism.” Beyond intellectual prejudice, I believe this itch comes from recognizing that the value added by management is largely the improvement of complementarities across activities.

    Dealing with multifarious, changing, fragmented customer needs and opportunities requires lots of trial and error, with resources diverted quickly and ruthlessly out of losers and into winners. Competitive markets are very good at this. No management structure is ever likely to be able to match, let alone surpass, the speed and precision of the market in this role.

    Where management justifies giving up some of that cat-herding efficiency is where it can exploit scale economies, shared resources, mutual adjustment of marketing and design and manufacturing, or the like. Markets often (not always) have trouble exploiting these kinds of complementarities, especially when they must first be established.

    Real-world organizations such as Unilever of course face more ambiguites and gray areas than this brief argument accommodates, but the basic logic applies: Can the supervisory layer of the corporation perform useful enough coordination activities to justify its costs? Without some “enlightened” concept of how restrictions on the freedom of local units can add value, the answer will surely be “No.”

  2. Grant

    Steve, great response, and I like this idea very much, make the unit of intelligence and action maximally mobile and let the play of the market decide the matter. The trouble is the world is so changeable and so multiassumptional. No, that isn’t a word, what I’m trying to say is that at any given moment many markets turn on a dime, advantage goes not just to the party with the tightest fit to demand, but the one that put its money on the right assumptions. This means you can have a great product but you bet on the right assumptions. It also means that the party who wins one day, can lose big time the next. Surely, this gives advantage to firms that cultivate within themselves a variety of ideas, assumptions and approaches. I mean, consider the alternative. If you are Sony and you discover that Beta is not going to go, you dont want to have to start from zero in the development of something else. You want that up and running.

    This reminds me of consulting for advertising firms. Once I got inside these firms I was surprised to discover how little like the academic world they were. The executives of the firms were not pressing for better and clearer and fewer ideas of what advertising was. They did not want to tie there future to any single star. If a client walked in the door with a Freudian notion of what advertising was, the ad firm wanted someone who could walk this walk and talk this talk. Jungian? Sure we’ve got one of those too. There was adaptive advantage to variety. (I mean this example mostly metaphorically.) Thanks, Grant

Comments are closed.