Anonymous
To calculate the daily average bookings, using historical data, I would first add up all the bookings in the time-frame under consideration, then divide by the number of days in said time-frame. Of course, the key factor here is which time-frame to use. A shorter time frame maybe more sensitive to special events, like holidays, promotions or system outages, however it could also be a more timely indicator than say a longer time-frame, like the daily average bookings measured over the past 2 years. If the daily average over the past week is trending higher, this time-series maybe more informative than say the daily averages using 2 year time-windows. Now, depending on the event, you could simply ignore the days where the event happened, e.g. don't include the day of the system outage. Alternatively, you could the time-series of daily averages and smooth the time-series, either with a moving average or try something more complicated like loess. Lastly, there are seasonal factors too that may effect the time-series, which can be decomposed into components to better understand the trend.