Casadellibro.com (shouldn't it be Casadellibro.es?)

The case describes the situation Casa de Libro was facing when the internet bubble burst and with it their projection of sales. They had to choose whether to stick with the already acquired high tech system, or change to a new one, more inexpensive in both maintenance and development.

I will first start with how I would mathematically frame the problem regarding revenues to solve this, then name the most important factors I would need to take into account for the inputs of the problem (with assumptions about them) and finally make the choice.

The mathematical approach

I would frame the problem following the next steps:

• Build brand new selling projections of sales, for at least three scenarios (best, medium and worst), and estimate some probabilities.
  
  
• For each scenario, draft how your plan would evolve if you chose each of the options. Probably, in the worst case of sales, your sales would be low, and we wouldn't need to change to a new server until much time after that. In the best case of sales, the simpler system would run out of resources fast and I would need to change it.


• Draft, following the plan, what would be the monetary outcomes in each of the scenarios, and the risks incurred.
  


• If the risks incurred weren't dangerous (we will further develop on this), I would calculate the expected profits (average of the expected profits for each scenario, weighted by their probability) of starting either with each option , compare and choose the highest

Until here, it looks simple... but not all in life are maths!

The factors to take into account

The procedure above leads to different results, or may even be invalid depending on management's preferences and strategy for casadellibro.com:

• The first question is... which are the forecasts now?! In each scenario, when would it be necessary to change to a bigger system if the smaller one was adopted? Is it long enough to expect a shift in prices and technology regarding these kinds of systems? How much will have technology evolved?
  
  Most of the internet forecasts were far far far away from reality. Books was probably one of the greatest deception, having experts procastinating the disappearance of physical bookstores in the short term. I therefore assume their forecasts were possibly 10 to 100 times to high, thanks to all of the massive internet adoption assumptions and "winner takes all" theories.
  
  If this is so, probably the smaller system could do it's job for several years before needing an update, and it would be worth to follow this path to avoid the high maintainance costs and expect lower development and maintainance costs when the time to change will have arrived.
  
  
• How much does Casa del libro gives up its previous strategic objectives?
  
  Given Planeta has overall a conservative strategy, and, as said in the case, cost control and loss reduction was the priority, I would assume Planeta wouldn't want to invest heavily in big sunk costs for a smaller than forecasted market.
  
  

• What kind of risks does management prefer? Do they prefer the economical risk of overstating demand or the risks of a potential need to reescale and redesign?
  
  What happens when the system runs out of resources? Will it become a bit slower? Will it deny service to a few users? Will it deny service to many users? Is reescaling a safe operation or will it leave customers unatended for two weeks? Could the reescaling process damage the online version of Casa Del Libro reputation forever? Could it even affect the whole Casa del Libro brand?
  
  I assume that the problems of the Microsoft solution would not damage the Casa del Libro brand in any case. Standardized solutions often offer better possibilities for migration than tailormade expensive ones.
  
  
• What are the technical advantages of each solution? Is it just a question of capacity or also quality of service, stability and features? If there are differences in features, how much value do we assign to them given the environment?
  
  I assume there are features that will probably be better in the Unix system, such as faster search for users, but I would expect a standardized product to develop faster than an expensive tailor-made system, as R&D costs will be distributed among a higher number of clients.
  
  Therefore, when the time to upgrade has come (if ever), computers will be less expensive, have better performance and come with better systems. Starting with the big system incurrs in the risk of becoming obsolete in some years too.

My choice...

...as you probably have already guessed from above would be the Microsoft solution

Nowadays....

I would consider outsource the whole system development and maintainance:

- Less risk for better service (outsourcing companies can dinamically increment resources as my demand grows, with no idle resources being paid for)
- No core competence is lost - unless they pretend to be the next amazon. Competitive advantage online would be the same as offline: a great catalogue and supplier base, and the Casa del Libro brand
- Brand risks would be controllable by taking an active role in design and quality control

I assume in this case I can compete for a certain amount of time with amazon, thanks to brand awareness, proximity, and knowledge of the spanish book market and demand.