What do you like best?
I've been using Coursera for more than 3 years. In that time I've watched it grow from a basic platform that provides classes between specific periods, including courseware, grading and assignments, to one that actively seeks to build community around the topics and courses, encourages learners to go at their own pace (by offering repeat sessions of courses and automatically advancing learners who have fallen behind to the next course offering), and exposes learners to a wide variety of institutions and topics.
What do you dislike?
There is no way to rate a course within coursera. Some of them are very mature, very well-developed, and others need a substantial amount of work. At the very least it would be helpful to have an idea of the level of challenge - rated by other Coursera users. For example, there are five or six machine learning courses currently being offered. These differ from one another based on platform (Matlab, Python, R), approach (start with theory or application), and skill level. It's the skill level that's difficult to quantify - a professor in computer science who is accustomed to getting upper-level undergrads may label a course as beginner, while another from sociology who is accustomed to getting sophomores with no programming background may label a course as advanced, but these could really provide an identical experience.
The sales-side is fairly heavy. If you want to sign up for a course for free (Audit it), sometimes you have to dig to find that as an option.
Lastly, there are no links within the sequence pages to find individual courses. Truth be told, I'm not going to shell out $495 for a 6-course sequence until I've completed at least one course and have a taste for the teaching style, methods and subject areas. There are series I've begun in earnest only to find I didn't enjoy the approach, or found that it was a language I was interested in but the focus for application was not what I was looking for.
Recommendations to others considering the product
Take your time to shop out courses and allow yourself to watch a few lectures before making a decision as to whether you'd want to delve in and take the course, for credit or for audit. There are multiple courses for some subject areas, and there are courses offered multiple times per year.
Also, give yourself the time to make an actual committment. I often see folks get frustrated when they commit an hour to a course that asks for 5 and probably took 9 or 10 because a video needed to be re-watched or an assignment didn't make sense the first time through. If you try to rush through it you're really just cheating yourself.
Before deciding whether to pay for a class, ask:
- Will this help my career?
- Is there sufficient content?
- Do i want to support this professor, school, course or platform?
The honest truth is, I don't know how much of a $79 course fee the prof actually sees. I hope it's a good chunk. I'd actually not mind going to a subscription-based method where I pay $200/year and get unlimited courses. I would pay that. I don't think I'd pay $100 per course.
What business problems are you solving with the product? What benefits have you realized?
The business problems I've sought answers for on Coursera have ranged from QA and program evaluation to process mapping to database interaction. I could not have had the start I did in my most recent job without the (now-discontinued) course on Database Systems taught by Jennifer Widom, nor would I be as far along as I am in learning R or Python. I also found UCSD's course on Big Data very helpful as we try to set up a localized hadoop cluster.
On a personal level I'm exploring where I belong in this Big Profitable Amorphous Thing Called Data Science. I think that's the technical name :-) The variety of offerings on Coursera for Data Science seem to drive a LOT of the site's traffic and participation.