There is no absolute standard. If your are doing Bayesian estimation, your posterior is a weighted average of the sample likelihood and your prior. In this case, you don’t need any observations. The posterior would simply be your prior. Adding observations will help you in updating your prior, i.e. the data becomes more informative, until asymptotically, the prior vanishes and the posterior becomes the likelihood.
The question for you is with which amount of updating the prior you feel comfortable. 52 observations does not sound too bad, but you should expect some parameters to be weakly identified, i.e. the prior you assign will be very important.