Selection of very first datapoint of annual cyclicality time series for forecasting

I am new to time series, sorry if my question is stupid. I have an annual cyclicality (cycle starts from January), monthly frequency time series data. I would like to forecast the next few months on this data.

My question is that the very first datapoint (starting date) must be January 开发者_JAVA技巧(start of the cycle) or it can be any months and the models (I am using auto arima and prophet) do not care about the very first datapoint?

The alternative is to drop the first few months datapoint till I get to the very first January datapoint.

Which option should I use?

Thank you in advance!