Molecular Biology protocols, utilities and oddities
Green Web Hosting
Transformation efficiency is determined by transforming an aliquot of competent cells with a supercoiled plasmid of choice, counting the number of colonies obtained, and finally expressing this number as CFU/ug, that is, the total number of colonies obtainable by (theorethically) transforming cells with one microgram of supercoiled plasmid.
The formula you should apply is:
Transformation Efficiency (CFU/ug) = (colonies per plate, CFU) * (total volume, ul) * (dilution applied if any) * (1,000,000 pg/g)  /  (volume plated, ul) * (pg DNA used) 
I have written a small online calculator that might help you:

Online transformation efficiency calculator
How many picograms of supercoiled plasmid DNA did you use? pg
Final transformation volume (including SOC/LB added at recovery): ul 
How many microliters did you plate? ul
Did you dilute prior to plating? If so, how many times? x
How many colonies did you count? cfu
1) Smaller plasmids will give a higher efficiency value, for two reasons: there are more molecules of a smaller plasmid in one microgram of DNA, and smaller plasmids enter bacterial cells much better than longer ones. So, it's a good idea to choose a well-known, high-copy, small plasmid such as pUC18 or pUC19, and keep using that one over time, so you'll be able to easily compare the transformation efficiency of different lots of cells.
2) Overloading the transformation with plasmid DNA will tend to subestimate efficiency, so the less DNA you use, the "prettier" your number should be. 
So! How much DNA should you use, and how should you plate the cells, you ask?
Your goal is to choose the right amount of DNA and dilutions to plate, that will give you at least one readable plate (i.e. a plate with a trustable number of colonies). This means that you'll want a plate with more than 20-25 colonies (for the number to be statistically more relevant) and less than 80-100 (i.e. to avoid crowded plates). So, you need to "tune" DNA and platings to your target range of CFU/ug. I have prepared a transformation efficiency "assistant" spreadsheet that's able to analyze your specific needs, and you're heartly invited to check it, but here are some short guidelines.
If you're aiming at the range 1*10^6 to 1*10^8 and your final transformation volume is 1 ml, then transform with 1 ng of supercoiled plasmid DNA. Plate 20 ul of undiluted transformation, of 1:5 and of 1:25 dilutions. Then use this simplified formula: CFU/ug = colonies * dilution * 50,000
If you're aiming at the range 1*10^8 to 1*10^10, then transform with 10 pg DNA and plate as above. Your simplified formula is: CFU/ug = colonies * dilution * 5,000,000
If you're only interested in knowing if the new batch passes a certain efficiency limit, i.e. above 10^7 or above 10^9, you might again use my spreadsheet to setup your experiment. Anyways, here are a couple of simple suggestions. For 1*10^7, with 1ml of transformation volume, transform with 125 pg DNA and plate 20 ul. You're aiming at >=25 colonies in the plate. For 1*10^9, use 12,5 pg DNA and plate a 1:10 dilution. Aim again at >= 25 colonies per plate.
Additional link: you might like to check a couple of posts in BioTechniques Forums where I've contributed too: here and here