The average of the most recent N periods, recalculated each period, used to smooth out noise and show the trend.
A moving average smooths a series by replacing each point with the average of the last N periods. A three-month moving average of sales dampens one-off spikes; a twelve-month one removes seasonality entirely. It is also the simplest forecast: next period equals the latest moving average.
Simple moving average at period t = (value at t + value at t-1 + ... + value at t-N+1) / N. A weighted moving average gives more recent periods larger weights that sum to 1.
Monthly sales of $100,000, $120,000, $110,000 and $130,000 give a three-month moving average of $110,000 at month three and ($120,000 + $110,000 + $130,000) / 3 = $120,000 at month four.
It lags: in a rising trend it always reads low, and a turning point shows up N/2 periods late. A window that does not match the season length leaves seasonality in. The full guide is moving average in Excel.