As I understand it, unless you hold a commercial pilot's license, it would be a violation of FAA regulations for you to charge your passengers any more than the pro-rata cost of the trip--fuel, airplane depreciation, etc.
Services a better business than software?
on
Andreesen "Grows Up"
·
· Score: 2, Informative
While much is made of the inevitable haggling that takes place at the quarter's end (when customers know they have software companies in a tough spot), this article glosses over the difficulties inherent in running a services business, particularly one like Loudcloud.
A software company has terrific gross margins: as we know, software costs almost nothing to reproduce and distribute once it's developed. Sure, the costs of developing it can be high (in a commerical setting), but each additional product costs virtually nothing to stamp onto a CD (or make available on a server for download). The basic point is that the inherently high margins in software provide a great cushion of potential profit, and one can grow a business to vast proportions with little additional development effort (but with some sales effort). Selling a product to one person or 1x10^6 people takes little additional development effort, in theory.
In a services business, however, each additional customer requires expensive infrastructure and personnel to develop. Whereas software has high fixed costs and low variable costs, services businesses (like IT consulting, lawyers, etc.) tend to have high variable costs (mostly labor) and low fixed costs (rent, etc.). The problem comes when you try to grow these businesses to tremendous scale. It is extraordinarily expensive to hire, train, and retain talented people, especially in IT.
For a place like Accenture (or a tiny 4-person consulting firm), the people are the assets--growing the business means growing the employee base, period. This is a fine model, and tends to have low fixed costs, but the profit margins tend to be lower than in packaged software. That said, people-based services businesses don't depend on expensive equipment with which to perform the service.
In Loudcloud's case, you have the worst of both worlds. On the one hand, you need racks and racks of expensive servers, storage, etc. with which to provide the service. On the other hand, you need extremely expensive people to manage all of these complicated setups. You have the high fixed costs of a product-style business, and the high variable costs of a "bodyshop" type business. In addition, you typically don't have proprietary offerings--one can buy managed hosting from a many companies, large and small. Loudcloud tries to address this in two ways, namely by developing "Opsware" (to lower the number of people required to run the business and to give it a pseudo-proprietary edge) and by (presumably) purchasing new equipment as it adds new customers.
From a profitability perspective, though, Andreesen is wrong--packaged software is, as an industry, one of the highest-margin businesses going (both gross and net), while services (esp computer services) tend to have lower margins and lower profitability.
Besides, there are lots of ways to deal with the "end-of-quarter" haggling. You think Nike and Ford don't haggle on the price of services that are provided to them?
To be perfectly honest, I think that while kids playing with technology is cool, it truly suspends disbelief to argue that it will result in tremendous advances or new ideas. Frankly, taking that same money and educating poor children around the world will pay back far greater returns than a two-day conference.
I work for a venture capital firm, and in the height of the internet boom we saw many, many ASP-type projects. It seemed, for a while anyway, that all new software projects would be delivered as ASPs.
Interestingly, though, this can make life difficult for the vendor. While the subscription model has its advantages (dependable revenue stream even after 100% customer adoption of your product, low incremental start-up costs for your customers), it can also be deadly to a startup.
The basic problem is that costs grow at the same rate for a company selling shrink-wrapped software or subscription software. However, where shrink-wrapped vendors get paid in large chunks (at least for serious enterprise software, $50K++), subscription vendors get paid in increments. As a result, revenues can severely lag costs--you've borne the expense of developing, selling and supporting the product, but you haven't recouped much money until you sign up lots of customers (leading to rising sales and support costs).
In a profitable company, this isn't deadly. But for a startup, every dollar in losses must be funded through selling stock. Booking revenues is the best way to increase the value of your company and to give up less stock in each sale, thereby preserving your ownership. In times like these, when funding is hard to come by, slow revenue progress can equal death.
The other issue, which applies to all companies, is that you're putting off a stream of payments (revenues) into the future, instead of collecting it all at once. Based on the time value of money, the sum of the subcription payments must exceed the up-front purchase price (as there is a discount factor for dollars to be collected way out into the future). The basic idea is that if you had the full lump sum today, you could invest it over the same time period of the contract. To your customers, it's a wash economically. But in a competitive market, it might be trouble, as it might mean you have to charge incrementally higher prices, leaving you vulnerable to competitors.
Finally, figuring out the pricing of an ASP subscription is really tough. How many months of payments should be required to have equalled the up-front license cost? Set it too low, and your customers get sticker shock. Set it too high, and you just put your own revenues further and further out into the future. Software pricing can be pretty arbitrary to begin with, and ASP models can add a layer of complexity. It's not that they can't succeed--plenty of businesses bill per-month. It's just a question of whether it's the right model for enterprise software companies.
Slashdot Mirror
It's still up at MIT. MIT just uses http://web.mit.edu for their main site--the egg is at http://www.mit.edu/404.html.
As I understand it, unless you hold a commercial pilot's license, it would be a violation of FAA regulations for you to charge your passengers any more than the pro-rata cost of the trip--fuel, airplane depreciation, etc.
While much is made of the inevitable haggling that takes place at the quarter's end (when customers know they have software companies in a tough spot), this article glosses over the difficulties inherent in running a services business, particularly one like Loudcloud.
A software company has terrific gross margins: as we know, software costs almost nothing to reproduce and distribute once it's developed. Sure, the costs of developing it can be high (in a commerical setting), but each additional product costs virtually nothing to stamp onto a CD (or make available on a server for download). The basic point is that the inherently high margins in software provide a great cushion of potential profit, and one can grow a business to vast proportions with little additional development effort (but with some sales effort). Selling a product to one person or 1x10^6 people takes little additional development effort, in theory.
In a services business, however, each additional customer requires expensive infrastructure and personnel to develop. Whereas software has high fixed costs and low variable costs, services businesses (like IT consulting, lawyers, etc.) tend to have high variable costs (mostly labor) and low fixed costs (rent, etc.). The problem comes when you try to grow these businesses to tremendous scale. It is extraordinarily expensive to hire, train, and retain talented people, especially in IT.
For a place like Accenture (or a tiny 4-person consulting firm), the people are the assets--growing the business means growing the employee base, period. This is a fine model, and tends to have low fixed costs, but the profit margins tend to be lower than in packaged software. That said, people-based services businesses don't depend on expensive equipment with which to perform the service.
In Loudcloud's case, you have the worst of both worlds. On the one hand, you need racks and racks of expensive servers, storage, etc. with which to provide the service. On the other hand, you need extremely expensive people to manage all of these complicated setups. You have the high fixed costs of a product-style business, and the high variable costs of a "bodyshop" type business. In addition, you typically don't have proprietary offerings--one can buy managed hosting from a many companies, large and small. Loudcloud tries to address this in two ways, namely by developing "Opsware" (to lower the number of people required to run the business and to give it a pseudo-proprietary edge) and by (presumably) purchasing new equipment as it adds new customers.
From a profitability perspective, though, Andreesen is wrong--packaged software is, as an industry, one of the highest-margin businesses going (both gross and net), while services (esp computer services) tend to have lower margins and lower profitability.
Besides, there are lots of ways to deal with the "end-of-quarter" haggling. You think Nike and Ford don't haggle on the price of services that are provided to them?
This sounds like a big conference put on by the MIT Media Lab a few years back:
www.jrsummit.net
To be perfectly honest, I think that while kids playing with technology is cool, it truly suspends disbelief to argue that it will result in tremendous advances or new ideas. Frankly, taking that same money and educating poor children around the world will pay back far greater returns than a two-day conference.
Interestingly, though, this can make life difficult for the vendor. While the subscription model has its advantages (dependable revenue stream even after 100% customer adoption of your product, low incremental start-up costs for your customers), it can also be deadly to a startup.
The basic problem is that costs grow at the same rate for a company selling shrink-wrapped software or subscription software. However, where shrink-wrapped vendors get paid in large chunks (at least for serious enterprise software, $50K++), subscription vendors get paid in increments. As a result, revenues can severely lag costs--you've borne the expense of developing, selling and supporting the product, but you haven't recouped much money until you sign up lots of customers (leading to rising sales and support costs).
In a profitable company, this isn't deadly. But for a startup, every dollar in losses must be funded through selling stock. Booking revenues is the best way to increase the value of your company and to give up less stock in each sale, thereby preserving your ownership. In times like these, when funding is hard to come by, slow revenue progress can equal death.
The other issue, which applies to all companies, is that you're putting off a stream of payments (revenues) into the future, instead of collecting it all at once. Based on the time value of money, the sum of the subcription payments must exceed the up-front purchase price (as there is a discount factor for dollars to be collected way out into the future). The basic idea is that if you had the full lump sum today, you could invest it over the same time period of the contract. To your customers, it's a wash economically. But in a competitive market, it might be trouble, as it might mean you have to charge incrementally higher prices, leaving you vulnerable to competitors.
Finally, figuring out the pricing of an ASP subscription is really tough. How many months of payments should be required to have equalled the up-front license cost? Set it too low, and your customers get sticker shock. Set it too high, and you just put your own revenues further and further out into the future. Software pricing can be pretty arbitrary to begin with, and ASP models can add a layer of complexity. It's not that they can't succeed--plenty of businesses bill per-month. It's just a question of whether it's the right model for enterprise software companies.