Someone who is familiar with all 3 should put up a post describing the pros and cons of each...
Well, I do spend a significant proportion of my time doing just that, but talking to/emailing prospective users. The bare bones are relatively simple (and bearing in mind that some users, especially those with a few more years on the clock, don't have too much in the way of computing skills and are also often looking for simplicity rather than 1001 features):
WiFi Logger: Probably the simplest and most cost-effective way for users to upload to online weather networks or connect
via WiFi to local software;
Nano/MB Pro: More powerful but higher-priced and with a more complex interface than WFL. But able to function as a complete standalone local computing device, especially for those with greater computing skills. Inevitably, you have to use the built-in software implementation rather than having the flexibility to choose your own software that using a separate computer would allow, but if Nano does all you need then that's obviously not an issue;
Weatherlink Live: Also reasonably costly and, for most users, closely tied to using weatherlink.com and the WL smartphone app. But for those who are happy to use wl.com, especially with a Pro subscription, then it's an easy and appealing way of getting all your data on to the Internet and and with the responsibility of managing the data handled by someone else. There's the added bonus of full sensor flexibility (ie totally your choice of transmitters/sensors on all 8 wireless channels, which MB Pro Red - but not Nano - can also offer, but not other Davis-compatible loggers either, other than the legacy Envoy8X). WLL is for Davis
wireless stations only, ie not cabled.
But beyond this short summary, it's not easy to provide useful extra comparative detail without getting into the minutiae of exactly how various features work on the different devices, and with quite a lot of qualifications and footnotes. It's more down to trying to tease out exactly what the individual user is seeking to achieve (though often they can't articulate a clear idea of this, which isn't particularly a criticism - if you've never used a similar device then how would know what kind of features it can offer?), their budget, level of computer skills and so on.
And as per other posts, this summary is broadly true today but in eg 6 months time who knows how the data landscape may have evolved further.