This is the most rigorously studied chromatographic parameter, but, unfortunately, it is the least important one.
Simple example. If your components have selectivity (a) equal to 1, does not matter how high the efficiency of your column you won't be able to separate your components.
Reminder. Selectivity is the ratio of the capacity factors.
If the selectivity of your system for target components is 3 and more, then even the column with only 100 theoretical plates will separate them.
In HPLC you can not significantly adjust the efficiency by changing the flow rate. Typical Van Deemter plot for HPLC column is shown on the left. As one can see, there is no significant change in the applicable flow rate region.
Most of the modern HPLC columns are showing efficiency of 10000 theoretical plates and more (being measured on the well-optimized chromatographic system). If you see the efficiency of your new column of 5000 or lower, most probably you have some extra-column broadening effect.
Here are some recommendations on how to avoid unnecessary loss of the system efficiency.
The only important parts (in terms of efficiency) of your chromatographic system are the injector, column and detector flow cell.
Try to avoid any unnecessary extra interconnections between your injector and column, and between your column and detector flow cell. It is strongly recommended to have just one piece of capillary tubing between injector and column, column and detector.
Many detectors have special unions (marked "IN") for column connection. If you open your detector and check what is behind this union, you will be very surprised to see several additional connections and mile-long tubing either wrapped around the cell or going into the strange aluminum block (supposedly acts like thermal exchanger). These detectors designed by spectroscopists, who have no idea what HPLC is. They try to make extremely stable baseline, but the baseline noise due to the flow and eluent composition fluctuations is much bigger then from temperature fluctuations. If you remove all these parts and pull single tubing right from your flow cell into your column you will gain several thousands theoretical plates and will not notice any noise increase.
Somebody may argue that this makes it inconvenient to change columns. If you will be using single piece finger-tight nut, your column change will be fast and very convenient.