One answer is that it is certainly not always a waste of time to apply for a job even if you don't think that you fit the ad.  It's easy to just keep copying your application, even trivial in MathJobs.  Moreover, you should take the details of ads with a grain of salt.  The semantics of research areas are often based on internal department politics.  Who knows what they really mean if they say that they want to hire in "mathematical physics" or "geometry", etc.  Even the advertised level of seniority is sometimes negotiable.

Another answer is that, for the sake of your sanity, you should have a "Plan B" if you want a tenure-track research job in mathematics.  The set of available positions from math majors, to graduate school, to postdocs, to tenure-track research positions is a severe pyramid at every stage.  Then, two-body problems can make things even more difficult.  The two basic choices for a Plan B are industry and teaching, and you have to decide which one is more livable for you.

In my case, it would have been industry.  It is true that it takes new skills to have a career in industry, some combination of training especially in computer programming, and sometimes also statistics and applied math.  But it is not necessarily all that formidable.  If you can get past the personnel office, many managers realize that mathematicians are fairly adaptable to new skills.

Likewise, teaching is not necessarily all that limiting; many people with teaching positions find ways to have good access to the research community.  Some go on to get good research positions at a later stage.  But if "Dr. H" is not at peace with this type of "Plan B", then maybe he should consider the other type.

For both answers, my experience is in the United States.  That said, an important answer for all countries is to expand your geographical range as much as possible.  Just as a matter of basic statistics, it can make a huge difference.  When my wife and I were on the job market, we applied all over the US.  Although we're happy where we are, we could have given more consideration to the rest of the English-speaking world, certainly Canada but also even places like Singapore.