Monday, August 15, 2005

Bruce Eckel asks What is Consulting?

And there's a link to a classic Joel Spolsky story too which sounds suspiciously like it's about Ars Digita.

When I worked at an Ars Digita inspired start-up a couple of years ago, I had a suggested solution for the problem Joel raises : take a lesson from how smart individual cells scale up to multi-cell creatures : mitosis.

Particularly splitting into smaller fairly autonomous units which specialize in different things according to the tastes of those joining each of the separated units.

What happens next is that each subdivision either thrives or dies. If it thrives, it probably grows and sub-specializes again. If it dies, good members may get absorbed back into other cells. There's no reason to write a manual, because no two cells are meant to be doing exactly the same thing. Each will become more specialized, expert in it's field. And that specialization can lead to wider recognition and higher fees.

A reader writes :

I'm not sure I understand or believe this. Can I ask a couple of questions?

Sure. Fire away.

Isn't there a danger in becoming over specialized and going bust?

Sure. Isn't there always. But here's how it's different from just striking out with a completely independent company. Other cells in the company are at least partially aware of and watching what goes on in your cell. In the event that your specialization turns out to be not what your market needs, then a good member of a cell should have a high hope of being absorbed into a more succesful cell.

This could possibly be put on a slightly more formal basis. Eg. cells have some obligation to recruit from inside other cells before looking elsewhere.

Isn't this like an ordinary large company then?

No, because it's not heirarchical. Cells are fairly independent, with just a few co-ordinating principles such as a shared constitution, maybe some shared services (eg. office space, server hosting),

So how does this count as "scaling up" for the greedy founders? How do they get paid more?

Well, I'd say they could get paid more by selling services to the other cells. Let's suppose one cell becomes an extreme programming consultant, while the other promotes a variation on the unified process. Other cells might become venders of software or legal services to these consultancy cells. They can charge high prices for being extremely attentive to the needs of their siblings.

In which case, how constrained are cells to buy from other cells in the overall company. Isn't this just a restricted, limited internal market, which is likely to be less efficient than buying from outside? Why are these cells still part of the same company? Wouldn't they just be better off as entirely independent companies with maybe a shared history and a completely informal culture of co-operation and sharing of staff. Rather like Savile Row's tailoring culture in London or Silicon Valley?

Good point. The actual formality of the interdependence might be negotiable.

Which reminds me. Would enterprising independent start-ups be allowed to join this slime-mould?

Hmmm. Hadn't thought of that. Maybe ...