You raise some really good (and difficult) questions and I hope the following (rather long) response is of use.
From my own experiences (and trying not too replicate Marcus' post too much) I found that moving towards blended learning created two issues:
When courses were being developed from scratch I generally found that there was more effort needed to develop a blended course than a trainer-led course. This was however offset by the fact that the running costs of blended solutions were lower (mainly due to reduction or removal of travelling and hotel costs) AND that trainers could deal with a higher volume of “delegates”.
I appreciate that these are very broad indications and to be honest the cost of development and the subsequent savings in deployment depended enormously on the approach adopted and the control measures put in place. For example, video elements can be developed at costs ranging from the fairly low to the painfully high and take from a few hours to a few months to produce. The manner in which you develop and combine the blended elements can also make a massive difference to the impact on your trainers’ time. Blended elements can have high, medium or low trainer contact e.g. a workbook, video or podcast could be low (or non-existent) contact whereas a virtual classroom would be high contact.
What I did find was this:
Overall the really big positives for me were that my trainers began to spend more time helping people rather than telling them stuff. This meant that the “delegates” got a far more personalised and faster service and the trainers got a much higher level of job satisfaction. This was underpinned by employee engagement data and shouldn’t be overlooked as a real positive for any organisation.
I hope this is of help and please let me know if there’s any other information I can help you with.