Estimating Twitter's burn rate - $7.5m do ya?

From the "Strategic Wild Assed Guess" Dept...........

Paul Boutin of The Industry Standard asks "what is Twitter's burn rate". This is the sort of challenge we consultancy types cannot ignore, but sadly the detailed data required - number of active users, messages per user, average price Twitter pays for ingress and egress bandwidth etc, volume of mobile usage -is not available in the public domain.

This does not stop us from estimating the likely burn rate from "educated" (ie stuff we've seen elsewhere that may be similar) assumptions. For what its worth, that "educated assumption" tells us its about $1 per annum per Twitter user. This figure is widely variable depending on assumed usage, usage of mobile etc, so we would go between $0.5 and $2 per annum to get a ballpark. As you get bigger, the actual bandwidth costs go down with volume, but the increased complexity of the social graph adds to the total transaction messaging required per user twt, so we'll assume that it all comes out in the wash and stick at that $1.

Twitter has, we believe, c 3m users or thereabouts - so thats a burn of c $1.5 to $6m per annum - lets say, at our $1 per user that its about $3m

Add to that the staff - $100k fully loaded cost is the number we usually use, so over 27 people that's $2.7m per annum - add 10% for contractors, consultants etc and round up to c$3m.

So, thats about $6m (well, between c $4.5m and $9m) to keep the shop running - but you have to add extraordinary items like system development, acquisitions, other CAPEX etc etc. Lets add in 25% of burn as a rule of thumb, about another $1.5m

Total Burn then is c $7.5m (give or take c 2 $m either way).

And of course, if it continues to grow at breakneck pace and even mainstreams later this year, it could have 6m users or more by year end - a mean burn rate for 2009 (at the $1 per user estimate over 4.5 mean users per annum) of nearer $9m+...give or take $3m either way.

Which of course would tally with Twitter needing to tap the markets about now (given the lead time of getting funding) after taking $15m in May 2008.

Caveat - the above data is on the digital equivalent of a beermat, so should only be taken as gospel after said beer is drunk