We present a new Umkehr ozone profile retrieval algorithm (UMK04) that has been optimized for the study of monthly mean anomalies (MMA) to assess climate variability in multi-year time series. Although the Umkehr technique is too noisy to monitor short-term variability in atmospheric ozone, it is capable of monitoring long-term changes in MMA with less then 5% uncertainty in the stratosphere, and with no influence from a priori information. By examining the information content of UMK04 we conclude that Umkehr data contain useful information about long-term ozone trend down to the surface, provided the data are analyzed as column ozone amounts in 8-layers, consisting of two ~9.6 km layers in the lower atmosphere (253--1013, 63--253 hPa), five ~4.8 km layers (32--63, 16--32, 8--16, 4--8, 2--4 hPa) in the stratosphere, plus a broad top layer spanning from 0--4 hPa.