Sunk costs is money spent that cannot be recovered (as is often the case in business…).
Sunk costs are irrelevant to your business plan and should not impact your decision to continue or stop. Including sunk cost in the business plan might lead to the wrong decision. In your investment decisions, you make mistakes every now and then. Count on it! So don’t through more ‘good’ money at ‘bad’ projects.
Continuing to invest in a project to recover lost resources (money, time etc) does not make sense. Here a couple of examples to reflect upon:
- You have spent 8 billion Euros acquiring a UMTS licence in Germany in 2001. But any UMTS business plan presented to you, including licence money that you have already spent, has a negative NPV. Should you launch UMTS services in 2003?
- You have spent 2000 Euros acquiring 100 shares of Deutsche Telekom at Euro 20 per share. The shares are now traded at Euro 10 a share. Should you keep the shares or sell them?
- You have spent years acquiring skills in the aerospace industry and the industry is in crisis. Should you work in the aerospace industry in the future?
- You have spent 1000 manhours developing a new product and demand is disappointing. Should you continue developing this product or ‘kill it’?
- You have sold a project for 100 manhours but have spent 1000 manhours already and the software is still not finished. There is no way you can get more money than the agreed 100 hours. Should you invest another 500 hours to complete the software?
There is no right or wrong answer to the questions raised above. But the correct answer should only depend on your belief of what will happen in the future, not on past decisions – some of them ‘wrong’, with hindsight.
In practice not including sunk costs can be really hard – we are human beings. Decision to stop and walk away can be very very tough – for example after you have spent billions of Euro on a UMTS licence and nobody wants to go to the board to explain why as a new entrant, your perspective of success are slim in a crowded market.