The Statistics of Forever

I recently presented a talk titled “Preservation Is Forever.” In this talk, I reviewed the Terms of Service for many online services offering photo products and digital storage. Many of these services have terms indicating that they can delete photos after a period of time (typically one year without ordering) or otherwise discontinue or change their service. These statements are correct, but they should not be interpreted as indicating that they will do this, only that their is a risk of something bad happening.

Here’s where we need to discuss statistics. If there is a 10% chance of something bad happening in a given year, after seven years your photos are more likely to be lost than preserved.  With a 20% chance of of loss, this time decreases to four years. The same problem happens if we assume lower probabilities, although it takes longer. With a 5% chance of loss, your data is more likely to be lost than preserved after 14 years, and with a 2% chance of loss your data is likely lost after 35 years.

What is the actual chance of loss? I don’t know, but my estimate is that in any given year, the probability for loss is between 5 and 10%, given the many things that can change with online businesses. If this is in fact the case, then Forever, which states that it will preserve your data for your lifetime plus 100 years, is really the only option for long-term preservation. The reason Forever meets this need is because the Forever Guarantee Fund provides the resources to ensure long-term preservation, eliminating risks associated with other online businesses.

Note: To calculate years lifetime, use the formula Y=Log(0.5)/Log(1-p), where Y represents the expected lifetime in years, 0.5 represents 50% chance of data preservation, and p is the annual probability of data loss.

Advertisement