If this kit includes the synthesis of cDNA beforehand (by SSIII): Do not use more than 7.5 Units SSIII/uL during RT (e.g., do not use the 10 U/uL as suggested). Otherwise, excess denatured RT enzyme carried over into the qPCR phase [95C doesn't stop all RT enzyme's cDNA-binding capability; even when denatured] can still bind to cDNA during the qPCR, causing sporadic qPCR amplifications (e.g. as noted by the company-stated case concerning 'Multiscribe RT' [an AB enzyme]). Other than that, try to use one specific concentration of pure total RNA/uL during RT synthesis of your cDNA. This specific concentration can range anywhere from 1 ng total RNA/uL RT rxn to 50 ng total/uL RT rxn -- BUT, whatever you choose, make sure all RT rxns contain that exact same amount of total RNA/uL. Too much RNA, and you 'scare' the RT enzyme into submission (inhibit it); the SSIII RT enzyme can indeed be inhibitied by too much total RNA added. Generally, any more than 50 ng total RNA/uL during RT and you place yourself in danger of introducing RNA sample-related contaminants that can inhibit the SSIII enzyme (depending on the actual purity of the RNA); e.g., its lack or abundance of RT enzyme-inhibitory material going into the RT rxn(s) in the first place. This highlights the reason to use pure (inhibitor and other-crap-free), high-quality RNA.
On the other hand (since there are many SSIII qPCR/RT-qPCR mastermixes currently available, and since you have not specified product #), if you are talking about an SSIII RT-qPCR kit that is of the One-step variety (where both RT and Taq are present in the mastermix), boost the final [Mg++] to 5 mM rather than using the default 3 mM, and use your primers at 775 nM each, and the accompanying hydrolysis probe(s) at 150 nM. But, if you are using the SYBR Green variety of this mastermix, use 500 nM final concentration of each primer in the final reactions, and leave the default 3 mM [Mg++] in place (so as to help encourage Taq specificity and discourage primer dimer formation).
Be certain the ROX content (and/or FAM content for MJ and BR machines) or lack of thereof is appropriate for your particular machine platform. Also recall that the original scientists who invented SSII and SSIII still reside at Quanta BioSciences - and thus perhaps check them out as an alternative source of mastermix if pricing and/or other unexpected corporate what-nots place you in the precarious state of a financial conundrum. But, never blame the dog for its bite, nor the corporation for its pricing, I guess.
Sometimes as low as 100 nM can give you good reactions. I should have said that there is a good range (between 200 and 600 nM), it depends on your quality of samples and initial care taken with primer design - but I have found, (with the SSIII mixes) 500 nM gives robust results. It's better to start with the volume up loud (higher primer concentrations) than using too little primer and not giving your samples a chance to show you what they contain. Optimization and a little trouble-shooting is always good. If 200 nM is working for you and it gives you clean results - stick with it. There's really no way to make blanket generalizations for all reactions -- but this mix (and its varieties) I am somewhat familiar with. With primer concentration, sometimes when you turn up the volume, you also bring up the noise floor, but at other times, there is very little noise to be concerned with - so, turn up the volume when you can!
Thank you for being here and helping to make these techniques better. Be sure to check out droplet digital PCR if/when you can. I am not in a situation where it is affordable - but, wow, is it a great idea. Absolute copy number capability - based on Poisson modeling of single copy compartmentalized nano-scale amplification phenomena. It appears to be the next big wave... but the price has to come down significantly first. In the mean-time, qPCR/RT-qPCR has the chance to improve its troubled reputation. There are still things that remain 'relatively' cheap yet sublime in the kitchen: the spoon, the fork, and qPCR/RT-qPCR.