Some responses:

* I won’t defend your experience at l’Arpege as clearly it wasn’t good for you. That said, next time we’re in Paris, I think you and I should try again.

* Keller mentions his investors. But what were the returns for his investors. Were they 2x? 5x? 10x? I’m speculating of course, but I doubt they were in line with what a typical (not every) venture fund would require. And I would imagine that if Keller had showed up at his VC’s office and offered them 2x, they would have sent him back with his check and told to scale scale scale.

* I would claim that you actually support my point with Keller. When Keller scales horizontally — with Per Se — the result is as good as the original (FL). But when Keller goes for scale and tries to cash in — with Bouchon — the result is, well, chain-like.

I’m not opposed to investment and I’m not opposed to scale. While we haven’t taken any investment here at JFM, we have shipped four of our own apps in our first 18 months of full operations.

Horizontal scaling to me is where you slowly and carefully extend your operation while always maintaining (or improving) the standards of the original. And when the choice is between quality and scale/speed, you always choose quality, because otherwise the brand gets diluted (like Bouchon dilutes Keller in my opinion). Keller spent months at Per Se getting it up to speed. He’s got live videoconference between the FL and PS kitchens. He had his hand-picked personally trained guy running the kitchen in New York. Believe me, he doesn’t invest the same amount of time in the Bouchons of the world. And it shows.

And while I have no problem with investors, my point is that in our industry, the VC model is often incompatible with horizontal scaling. When push comes to shove the VC wants scale even if it dilutes quality “just a bit”. And I get this, I’m just pointing out the incompatibility.

And to make sure that people don’t read into this too much… I do believe you can make something of quality with VC investment. And I do believe that things at scale can be of high quality. I just think the money and the expected returns make it much much harder.

I thought of your post when reading this LA Times article by Thomas Keller:

http://www.latimes.com/features/food/la-fo-side10-2008sep10,0,7679327,full.story

He starts off the article with, “When I was learning to cook, I never dreamed I’d wind up designing a line of porcelain.” Further in he mentions, “When I opened the French Laundry, my goals were simple: Pay back my investors, improve the restaurant, and offer good benefits to my staff.” Note the investors. He’s definitely scaled way up — slowly, to be sure, but on the back of investors and to a level of quality people generally admire far more than Wolfgang Puck’s.

It’s easy to pick on Wolfgang, and I’ll happily play along with that end of the analogy. You lost me a bit with L’Arpege, where I had an extremely unpleasant meal — good food, certainly, but in my mind crassly commercial service (my then-girlfriend, now-wife was given a menu with no prices by a waiter who then proceeded to almost urgently push her to order the appetizer and main that were vastly more expensive — on my menu with prices — than any other choice) — but, I get the point, and in general I definitely see where you’re coming from. But Keller?

Having eaten at Bouchon in Las Vegas, I can definitely see some support for your analogy. That place is not at all what I think of as Keller-quality. But: Keller running French Laundry and Per Se on opposite coasts, with comparable, very high quality, and soon to be four other restaurants besides, seems not to support your argument.

I’m not sure how far I’m willing to push my argument (I’m picky about restaurants), but someone could definitely go further with the “quality at scale” counterarguments: Boulud? Vongerichten? I seem to remember someone saying recently that a chef with a single restaurant in the US would go broke.

So, I don’t know. Maybe I just got distracted by bad memories of L’Arpege and you lost me there, but it seems to me that reaching for scale might not be the VC’s worst sin. IBM and Microsoft and now Google have all managed to attract very large populations of extremely talented people, with, in the latter case, their VCs (apparently) urging them towards extremely high quality, not just decent quality.

The TapTapTap approaches by VCs (like those to 37signals and many others) is still totally wrong, the Kleiner iPhone Fund is nuts, and there is definitely something broken about applying the VC model to these situations. I’m just not convinced that the bug is in reaching for scale alone. I think it probably has more to do with VCs being too trend-focused, instead. Chasing around the latest developer interest (Java, Rails, iPhone, etc) might very well pay off from time to time, but I bet the failures cost more than all the successes put together.