Oracle and the Dunbar Number
A little while ago I got a nice comment from buddy and guru Mark Tamis. He wrote, “I was thinking, it may be a good thing for Oracle to use its cash and buy up Salesforce, and then stick Benioff at the helm. What do you think?”
I have to say it was and is an interesting idea because Larry is 70 and just stepped up to Executive Chairman, leaving the CEO duties to Mark Hurd and Safra Catz. This idea has been surfaced before too. About a year ago it came up and died a gentle death when few outlets picked up on it.
Also, it might be human nature to think like this — to look for a single strong leader to take us to the promised land (whatever that is) — but real life experience seems to run in the opposite direction. Rather than seeking the uber boss, societies, at least the successful ones, instead split the organization whenever it gets too big.
There is a lot of historical precedence for the split over the big boss approach and it has changed over time in a predictable way. To understand the predictability you need a little math in the form of the Dunbar Number.
The Dunbar Number is actually a very elastic thing and could more easily be described as a concept or even a law of sociology (I dunno, I ain’t a sociologist). It says that the average human can keep track of about 150 or maybe 220 people and after that not so good results. This seems like a big number to me but I am an introvert, not shy, just not into maintaining lots of relationships and I am quite comfortable with ideas.
If you look at human organizations, pre-social media, the Dunbar concept applies remarkably well. Of course, CEOs are always trying to limit their direct reports but go up or down a level and you see interesting things. For example, in the Middle Ages, monasteries were working communities of about the Dunbar number. When one got too big, the abbot split off a unit and told one of the members to go elsewhere, build a new monastery, and keep up the tradition. That’s actually how the monastic tradition spread and if you read “How the Irish Saved Civilization” by Thomas Cahill you’ll see the story played out.
Now, social media has turned the Dunbar Number on its head but even with that assist there is a practical limit to the number of friends you can have even online. Perhaps that number varies by individual but the point is that it’s finite and probably not as big as the number of followers many of us have.
There are other examples too such as the military company, the atomic unit of military effectiveness. Corps, divisions, regiments, and battalions are all different aggregations of companies. But what’s this have to do with Oracle? Well it’s indirect. A story in today’s New York Times announces that HP’s CEO, Meg Whitman, wants to split the company into a PC unit and an Enterprise one.
My analysis is that HP is too big to be effective at pursuing a strategy even with all of the computers and communication infrastructure the company has so it’s a natural to seek a way to group similar products and skills into separate companies. Refer to the monastery idea and consider Dunbar and ask yourself if this makes sense.
So, all this is to say that my response to Mark Tamis and his idea is to think small-er. Rather than trying to find the uber boss, if that person even exists, it might be time for Oracle to consider cleaving itself into logical units that have more autonomy than they currently do but that still tree up to a single entity. This might be a worthwhile trend for a lot of first generation Silicon Valley companies.
Oracle is currently made up of a Byzantine assortment of in-house developed technologies and bought companies. It is also a player in almost every part of the tech sector from hardware to apps. The company’s current tag line, “hardware and software engineered to work together” is well chosen to give an impression that is no longer needed. The goods might be engineered together but that’s not the same as being designed and built for the purpose from whole cloth. In fact, Oracle reflects the marketplace it tries to serve which is sophisticated, complex, and aggressively heterogeneous, whatever the marketing lingo says.
Certainly Meg Whitman is rolling the dice with this split but from my perspective it is a logical and appropriate thing to do, and one that has historically delivered results. Would Oracle ever consider splitting into a hardware enterprise, a legacy software company, and a third dedicated to more modern web/social/mobile technologies? Never say never out in the valley, except maybe to the idea of Benioff taking over Oracle.