In my opinion you first have figure out what the scenarios are for a user that visits the site, and then pick the user up depending on their individual intent. I would suppose, though I don't know about your audience, the following:
Group A: Most users will get to the site out of general interest, and maybe want to try the software
Some (but significantly less) users will already know that they want to buy the software
Group B: some of whom will be users of the free version who decided to make the conversion
Group C: some because they already saw the full version or got a personal recommendation strong enough to encourage immediate buying of the product
Beware that these numbers are solely based on my gut feeling, so you might want to do some research amongst your specific target audience to make sure you're steering in the right direction:
- Group A ~80%
- Group B ~15%
- Group C ~5%
Assuming that group B will to a high degree be directed to a specific landing page out of the free product version, that would leave you with about a 94% with download intent and 6% with buying intent on the home page. Pure guessing in numbers just to illustrate the thought process, this needs to be verified!
Now, depending on the real numbers you get out of your research it will be pretty clear what action should be more prominent. Since you're really not offering "various plans", I don't think the analogy to feature tables is applicable in your case. Instead, look at products that do the exact same thing as your client's: Offer a free, restricted version.
- MAMP offers download free & buy pro at the exact same weight
- FontExplorer hides both, download & buy, away in the depths of the site
- Balsamiq also doesn't promote either option, yet they have the special advantage to offer a live-demo launch, which – in this case – is similar to a free version
- Things refers to the AppStore and offers a free download at about the same priority
- OmniFocus goes with a slightly stronger download-button, yet both – download & buy – are surprisingly small in the header. Further down the download-button is clearly more important than the AppStore-button.
Now, we don't know how many, if any at all, of these have been tested against other options. My personal feeling – again, beware – is that if someone wants to buy the product they will invest the 3 seconds to look for a buying option. One of the most elegant solutions for this is, in my opinion, the way apple is doing it, e.g. here: iMac Page with buying option. The buy button is very easy to find, yet doesn't try to draw attention away from the content.
So the question is: What does the user want? Assuming from what I learned during my own projects, the answers will most likely be along those lines:
- The user wants to know that the trial version comes free
- The user wants to know what limitations there are before downloading it
- The user wants to know what the full version will cost
- The user wants to be assured that he can use the application to a relevant degree to form an opinion
Long story short, my feeling is that a solution like this would work great for you:
If this were my project, I would use a version like this and a version like your original for some qualitative user testing and then run A/B tests in the live version. As has been stated already, it is inevitable to do those tests on the live system, because only then will you get data relevant to your specific scenario.
Furthermore the goal of the site, in your case, has te be made clear. It can be confusing if downloads go up and, thus, direct buying-conversions go down in relation, while really – through the conversions through the trial version – absolute conversions numbers are going up.