Saturday, June 30, 2007

Ski Rental and Value Propositions

Most of us (esp. in the US) would have rented items like Skis, Camping Gear, etc. from time to time. For the occasional skier or camper, this is an obvious choice. However, the more serious enthusiasts have an interesting problem - to rent or to buy? This is a well-researched topic and the math proves conclusively that the optimal strategy involves renting until the aggregate rental cost equals the cost of the ski equipment.

Now consider the following -

1. I bought a pair of roller-blades after taking my first lesson while attending Grad school. Over the course of a couple of years, I became a very proficient skater. Much like the researchers predict in their paper, I had a major accident that required 18 stitches over my left eye. Purchasing the rollerblades outright gave me the option to practice at short notice in my free time (an extremely valuable commodity in grad school) and is a decision that paid off well. Once you master rollerblading, many other sports like Skiing and Ice-skating are easy to learn and enjoy.

2. Taking a similar approach, I bought a cable modem for $200 in 2000 when I moved to the Bay Area. Comcast was renting these modems for approximately $10 a month and I reckoned that I'd my investment would pay for itself in less than two years. However, within a few months of my buying the modem, Comcast reduced the rental price to a ridiculous $2 a month, making me kick myself. A good 6 years after buying the modem, I still haven't broken even!

These examples illustrate why
the ski rental algorithm is great for the average case. Given the amount of skating that I did, I'd not have spent much more had I followed this approach while buying my rollerblades. With the Cable modem, I'd definitely have freed up precious resources for investment elsewhere.

Now, step back a bit and consider a startup (with scarce resources). Following the ski rental algorithm gives the startup great leverage in conserving its resources and investing in the most critical areas. At an early stage, leasing desktops from Dell/HP helps the startup free up capital to invest more on the product. Capital Expenditure can be done later after the product has stabilized and the revenue stream starts picking up.

This idea is also the cornerstone of many SaaS offerings. What do you think is the primary value proposition of I'd argue that the primary value proposition of is not CRM and its software, but the lower cost of ownership of such a solution. As I've noted in my earlier posts, comparable solutions from Siebel, SAP and Peoplesoft were very expensive to procure and had a very high TCO.'s innovation lie not in providing a hosted service, but in identifying that the Ski Rental algorithm is such a no-brainer.

Wednesday, June 20, 2007

Cars, Planes and the Airline Industry

One of the most popular keywords in the software industry today is to Software as a Service (SaaS), popularized by’s Marc Benioff. Marc has done a tremendous job positioning his company and its eponymous website as an inexpensive and indispensable tool for many small and medium businesses. However, will this trend catch on? Is there sufficient play in the market for more SaaS vendors and is that got lucky?

There is a close parallel between the various software segments – consumer software, enterprise software and SaaS with the transportation industry – car manufacturers, plane manufacturers and airlines. Drawing out the analogy on some aspects brings out the challenges faced by each of these segments and clearly shows that the SaaS vendors have a tough challenge ahead.

Customer Expectation - customers expect good quality, but also understand that there will be the occasional failure. Failure of parts in a car is relatively inexpensive to fix, even if the manufacturers have to do a mass recall. However, airlines expect a very high quality product when they buy a plane and will not buy even if a few parts have quality issues. An airline must ensure high reliability of their planes and also offer great in-flight service in order to stay in business. Even though the quality of software from enterprise vendors is often not as high as their customers would like, the expectation of quality in a SaaS tool is the highest – the software must provide Quality of Service guarantees and also be highly available.

Lifespan - The average maximum lifespan of cars is around 10 years, of planes 30+ years. Typically, we expect airlines to outlive various airplane models by adapting newer technologies. It is not that consumer or enterprise software companies have failed to build a sustainable business, but a SaaS vendor will be required to adapt more quickly to the massive technological shifts that occur so frequently in our industry.

Maintenance - car dealers and even unauthorized mechanics can do an oil-change for cars, whereas only well trained engineers with appropriate tools can do so for airplanes. Airline carriers have trained engineers who perform routine maintenance on the planes very frequently (a few thousand miles or so) in order to make sure that the planes are flight-worthy. A SaaS vendor should not only have the best architects and engineers building the product, but also an extremely capable bunch of administrators. Neither consumer nor enterprise software companies have to worry about training or recruiting administrators for their software.

Sales - In the plane market, most deals are big and a win/loss can alter the landscape of the market and profitability of the company. However, a few lost deals do not make a big difference to the pecking order of the car manufacturers provided they do not let it aggravate. The airline industry is the one with the highest volume of business and the lowest margins. The current rate for even sophisticated tools like is less than $100/month per seat and this is a huge stumbling block for SaaS vendors. Why would they build great software and settle for lower margins, if they could instead sell the software for a much better price?

Market – the car market is flooded by many manufacturers with models for every segment, whereas the airplane market is dominated by fewer players - Boeing, Airbus, Llockheed. Airline carriers are slowly specializing in the international, budget domestic and hi-end domestic markets. Not sure how the SaaS market will shake out and grow, but it is possible that we may have vendors offering custom services to larger organizations with strict SLAs and smaller players that target the mass market. This may already be happening with large IT vendors like EDS, IBM Global Services, Infosys, etc. developing and maintaining custom applications for large clients like GM, GE and others. I’m sure they are already thinking of the next big thing after services and this is an area with a lot of possibilities. The biggest challenge before smaller players like is to sell to the larger enterprises, as they cannot risk selling to only small and medium businesses.

The bottom-line is that very high customer expectation for quality and the general dynamics of the market make it very difficult for small startups to offer Software as a Service and build a sustainable business. Its probably for the best, because the few that do so, will have a clear and distinguished value proposition. As a corollary, I would not dream of starting a SaaS startup unless I have a clear and distinguished value proposition.