This is a pretty good post. I've also done pretty well consulting. Some quick notes.
First: $210k/yr is not necessarily really good money for a US software consultant. When consulting full time, please try to keep two things in mind:
(1) Your cost basis is higher than it was when you were a W2 FTE. If you're making $100k/yr in salary, your employer is paying substantially more than $100k/yr to keep you on staff; you have a "fully loaded" cost that includes not only infrastructure stuff like computers and office space and Google accounts and training and vacation and sick days, but also your benefits and a pretty substantial chunk of taxes, and a bunch of tax planning stuff that your W2 hides from you. You're now on the hook for all of that.
(2) More importantly: employers are on the hook for the fully loaded costs of their employees indefinitely. Well-run companies hire developers with the expectation of keeping them on staff with no fixed end date (run don't walk from any that don't). Which means that the decision to hire a freelancer versus a full-time employee is not simply based on rate; it's also based on the fact that the freelancer comes with a guarantee that the relationship can be severed the moment it's no longer valuable. That guarantee has a lot of value; "double your FTE rate" isn't even stretching it. If you're giving that up for free, by working at a rate comparable to what you'd be making in a good job, you're doing it wrong.
I don't think it ever makes sense to work hourly. I've written a ton of posts here about why that is; here's a link to the one people seem to like the most: https://news.ycombinator.com/item?id=4103417
What I can say with almost 10 years remove from that post is that I was if anything underselling my position on this. When I was beating the drum on not doing hourly work, I was at Matasano, and we had a day rate (you couldn't buy work for us at increments less than a day). After that, we started another consultancy, where our minimum billable increment went up... uh... substantially from that. You can do week rates, and not sell in less than 1-week increments; you can do month rates; you can do more than that.
The classic dumb argument about hourly versus not tends to devolve to debates about the pitfalls of fixed-rate work. I don't advocate for project rates (I'd do a project rate, I guess, if it made sense; I'm not religious about them). Rather: I think you should provide your customers with a proposal for a total cost for the project based on an estimate of the number of days (weeks, months) you think it'll take, and a SOW for a T&M project with available prorated overages if it takes longer. Then do your best to deliver according to your estimate; if you blow the estimate because you screwed up, eat the overage; if you blow the estimate because your customer didn't get you access to the systems you needed to work on until 3 weeks after the kickoff, they eat the overage. Nobody has ever pushed back on me for this.
When I spelled this out on HN back in like 2010, people responded as if it was black magic. I think what's really happening is that people who run serious consulting firms just don't write a lot of HN comments, because I know of lots of big firms that work this way.