over the last several months i tried an experiment in the city in variable cost living. the idea was simple - i adore the flexibility that 'cloud' computing gives me to run my business on 100% variable cost basis. the drop.io service can expand and contract on an hourly basis to match demand, which means that even if we pay slightly elevated costs per unit of storage or compute, we can match our cost structure exactly to current demand and don't need to worry about modeling service changes that might meaningfully change use patterns. we save a ton of time, money, modeling, and capex by running on a variable basis with zero friction to expansion/contraction... and i am quite certain that this is the way the world is going to move very quickly.
i figured, if it works for compute and virtual storage, shouldn't it work for physical space? after all, really what is the difference between a compute cloud and a human city (not much theoretically)? starting with the convenient factor that i own very little -- and the few things i do own i can easily deposit at the parent's home in nj -- this summer for several months i took my bicycle and a duffel bag and simply spun up/down housing on a monthly and even daily basis based on what i needed. when my girlfriend was in town we stayed at nice hotels. when i was out of town traveling i paid no rent anywhere. in the middle of the summer i got a place with a great roof deck, earlier in the summer i went without. this wasn't about saving money per-se -- it was about maximizing the utility/value of my housing experience for the same money by living with zero lock-in. it was about the concept that flexibility was under-priced in the market.... and, i figured that with an unappealing but very real backstop of a bed in nj, i felt that i had limited downside.
the reason i thought this would work in nyc is that the 'cloud' is quite big (plenty of inventory to create a competitive market), and i thought the friction of finding, vetting, and spinning up/down housing would be low and managable (if not truly zero). for a while, i thought i had it right, and and when it was working i was exceedingly excited.
sadly, however, i have had to concede defeat and enter a one year lease in the east village as of a few days ago. as they say, epic fail.
there was really only one reason that it didn't work -- but it is the big one -- the fiction around transacting temporary real-estate is just too high... it shouldn't be, but it really is -- first, my 'trusted' marketplace, sublet.com, turned out to not be so trustworthy (despite the fact that it had fees on both ends of the service to keep people honest). great things were listed, but often not available. this created a lot of friction in figuring out what the market actually was to begin with.
second, my 'social' sources were also harder to work than expected. sure enough, facebook, twitter, etc. made it easy to get to people who said 'sure, i have a place you will like, you would be actually doing me a favor to take it' -- but then for totally legitimate reasons plans change by a day or two here or there. this is totally fine and expected obviously -- but it doesn't work when you are counting on that availability.
third, transactions with people and space are difficult... it isn't like trading one dollar bill for another -- the goods and contracts are not standardized enough to be efficient for variable cost structures. take, for instance, the last place that i stayed at for an extended period -- we used my simple and strong set of letters of understanding which structured things to my liking. as is fair, i said i would pick up the utility bills which would be deducted from my deposit the net of which would be returned no later than september 5th. i said upfront that i wouldn't pay for the cable -- because i never watch tv...
today on october 1st i finally after several emails got the accounting and a promise that my deposit was in the mail. i will get it back ultimately, but not without several extra emails... and apparently net of a $75/month fixed phone bill (who even has a land line, let alone pays that much for base service) $60/month internet service (again, someone hasn't repriced recently) and outrageous electric and cleaning bills. could i have structured less exposure to these things upfront, sure -- but that creates more friction still.
upshot, i am glad i did it and i learned a lot -- but perpetual variable cost living doesn't work in our metropolis as a cloud. some people have suggested that i buck up -- telling me that it will all be better in a few years, but i don't think so. i sadly fear that without a significant number of variable-cost-livers it will be near impossible to evolve the types of standardization and transparency necessary to start thinking about a square foot as an ec2 instance.
lessin 0, world 1