A lagged value in a time series is a value of a variable corresponding to an earlier time. For example, in a monthly time series, the first lagged value will be the value for the previous month and so on.
Consider a time series $x_t$ where the value of variable $x$ is recorded at different time periods/points $t=\dots,-2,-1,0,1,2,\dots$. If the current time period/point is $t$, then lag $1$ of $x_t$ is $x_{t-1}$. E.g. if $x$ are sales and we are at the end of February 2021, $x_t$ is the February sales, $x_{t-1}$ is the January 2021 sales, $x_{t-2}$ is the December 2020 sales etc. More generally, lag $h$ of $x_t$ is $x_{t-h}$ for $h=1,2,\dots$. Lags can be used in autoregressive modelling and forecasting, the autocorrelation and partial autocorrelation functions (ACF and PACF) etc.
