It is conservative by design, and sometimes too conservative: a strategy that could genuinely trade the closing auction is penalised by being pushed to the next open, which on gap-prone instruments adds meaningful noise.
The point is not to be precisely right but to be wrong in a direction that does not flatter you. If a result depends on capturing the close rather than the next open, that dependence should be visible, and next-bar execution makes it visible immediately.
Worked example: a daily mean-reversion rule buying weak closes shows +0.42% per trade filled at the signal close, and +0.06% filled at the next open. The 0.36% difference was the overnight bounce, and it was never available to you.
Related: signal-lag, look-ahead-bias, fill-assumption, vectorised-backtest