I was mostly reflecting on a pattern in the people I've met, so most of my examples won't be persuasive.
Musing on some less personal examples:
Religious missionaries are selected for atypical faith and a resistance to "leave me alone" social cues. For many people, talking to a more moderate believer would lead to a greater shift in opinion. (Not that the goals of missionary activity are to convince the average person).
People who explicitly advocate for utilitarianism tend to care enough about the system to "bite the bullet" on certain issues, scaring away newcomers. (Peter Singer on bestiality and infanticide, Eliezer on dust specks and torture). People who are vaguely utilitarian but not too concerned about consistency can almost certainly do a better job convincing a non-utilitarian that they should be a bit more utilitarian.
Kant had some actually useful ethical insights, but said some downright stupid things in the application of his ideas (like: You shouldn't lie to a murderer who comes knocking at your door looking for a victim, but you should "be silent" or something).
If you're a progressive with a progressive social circle and want to learn about critiques of progressivism or about conservative thought in general, neo-reaction is about the worst starting place ever. It's like a conservative in the deep south trying to learn about the political left by reading Marx and browsing Tumblr.
The "highest quality" non-commercial strength training materials (meaning they looked really shiny and had the most investment sunk into them) are often extreme and "purity" minded, and were produced by people who trigger the "crank" and "how did THIS become a central feature of your identity?" red flags.
Health innovations in general (some of which I've adopted, albeit in a less extreme form) tend to spread fastest via apocalyptic messengers ("fructose is literally poison"), new-age people, or the self-experiment crowd, (who don't upset me, but would strike many people as cranks).
You seem to be conflating two things. People who give logically bad arguments for their positions, and people who say things that trigger listeners absurdity heuristic.
The more radical positions tend to be more logically coherent, hence easier to logically defend. On the other hand they're also more likely to trigger people's absurdity heuristics.
More moderate positions are harder to defend since you wind up jumping through hoops to explain why your logic doesn't apply to certain cases. This means that in practice the more moderate position functions as a Trojan Horse for the more radical position.
Unfamiliar or unpopular ideas will tend to reach you via proponents who:
The basic idea: It's unpleasant to promote ideas that result in social sanction, and frustrating when your ideas are met with indifference. Both situations are more likely when talking to an ideological out-group. Given a range of positions on an in-group belief, who will decide to promote the belief to outsiders? On average, it will be those who believe the benefits of the idea are large relative to in-group opinion (extremists), those who view the social costs as small (disagreeable people), and those who are dispositionally drawn to promoting weird ideas (cranks).
I don't want to push this pattern too far. This isn't a refutation of any particular idea. There are reasonable people in the world, and some of them even express their opinions in public, (in spite of being reasonable). And sometimes the truth will be unavoidably unfamiliar and unpopular, etc. But there are also...
Some benefits that stem from recognizing these selection effects:
I think the first benefit listed is the most useful.
To sum up: An unpopular idea will tend to get poor representation for social reasons, which will makes it seem like a worse idea than it really is, even granting that many unpopular ideas are unpopular for good reason. So when you encounter a idea that seem unpopular, you're probably hearing about it from a sub-optimal source, and you should try to be charitable towards the idea before dismissing it.