While I agree with Godin's general premise, the obvious (at least to me) implication of this to hackers is to make "enterprise" software. This is rather unfortunate. The sales cycle for such systems tends to be very drawn out, and large companies decide to purchase from vendors in a manner that seems rather arbitrary using criteria that is often non-obvious e.g. the long term survivability of the vendor is frequently at least as important as the quality of the product they provide.
Why is that the obvious implication, especially whith the two home-run examples he gives?
Starbucks and Microsoft are built on smaller transactions than one would expect from enterprise software. Starbucks does it $3 at a time. Microsoft established itself selling desktop software for for a few hundred a pop, and for quite a while, it was selling a couple of copies at a time until they gained enough momentum to sell their desktop software to whole departments and later even whole enterprises.
Interesting to consider that the average annual revenue Starbucks gets from a regular customer, is probably similar, or even a bit more than the annual revenue Microsoft gets per-seat from its best enterprise customers. A reminder that one should balance the difficulty of acquiring a customer against the lifetime value of that customer.
Getting $1 from a million people might be a tough business, but getting $20/year from 50,000 people one dollar at a time is conceivable.
Flickr might be a good example (though a little on the big side). When they were acquired by Yahoo, just a little over a year from their debut, they already had ~400,000 registered users. Within a few more months, they had a million, and a few months after that a million more. I'd postulate that nearly all of that post-acquisition growth (and much more) came from existing momentum rather than their association with Yahoo.
I don't know if the same ratios held back then, but when I looked a few months back, about 15% of flickr accounts were flickr pro accounts bringing in $24.95 in annual revenue. Thats 300,000 x $24.95, or $7M in annual revenue, just from pro accounts.
Margins were low, or negative, because they were going for growth, but they had various knobs they could twiddle. First off, a little after the Yahoo acquisition they cut the pro subscription price (by half, I believe). In addition, they controlled the amount of storage and upload bandwidth available to both free and pro account holders. Taken together these variables gave them latitude in picking margins. Even a relatively paltry 13% margin would have generated $1M in profit.
