Basically, skip tracing is the process of finding people who, for whatever reason, do not want to be found-- "skips", i.e. people who have "skipped out" on some social, marital, financial or other obligation.
There are a wide range of authors that write in this area, explaining techniques. Keep in mind that texts (paper or online) on skip tracing get dated very rapidly, as existing techniques are constantly refined, and new methods devised. A few names of authors are Edmund Pankau, Robert Scott, Dennis King (and many others).
Skip tracing has the same "cat-and-mouse-chase" quality as tax litigation, or criminal investigations. The more sophisticated "skips" are constantly ahead of the curve of the best "tracers", who are constantly refining their methods to catch them. This is exacerbated by the fact that virtually every skip tracing text can be "read" as both a handbook for locating "skips", and a handbook for avoiding "tracers".
As you are asking about "the collections area" in particular, cost-effectiveness is your main concern. I would guess that the lion's share of collections work is simply locating people who have decided that they no longer feel like paying their credit card bills. Some people are de facto untraceable, because the cost of tracing them would exceed the value of any potential financial reward of finding them. This problem is exacerbated by the fact that many "skipping" techniques are quite easy and inexpensive, while the tracing methods that currently exist to respond to them are expensive and onerous.
The reason for the recent up-surge in interest in "skip tracing" is due to the wide availability of inexpensive software such as "Net Detective". Prior to the advent of online databases and the Internet, basic skip tracing (and background investigations in general) were much more expensive (and thus less rewarding for collections purposes). So there are a large category of unsophisticated "skips" that are much easier to find. Outside this category, most of the industry still has to rely on the old techniques, part of the private detective's stock-in-trade for decades ("spoofing", "trash covers", and the like).
There are also several drawbacks to this software.
1. It is very "Americano-centric". The U.S. is the most computerized society (with the possible exception of Japan), and one in which information on individuals can be gathered and maintained with the least constraint. Thus, Net Detective may be effective in locating a skip within the U.S., but of little use elsewhere. Even U.K. citizens, for example, are considerably less tracked (except that in the U.K., surveillance cameras are much more ubiquitous).
2. It is cheap. This means any "skip" with the slightest sophistication has a copy, and uses it to search for traces of him or her-self, and takes steps to eliminate them.
3. Simple techniques to avoid it cost even less than Net Detective itself.
Someone who earns their living and pays their expenses exclusively in cash leaves few traces. If they move in with someone and do not have their names on leases, telephone service, electricity or gas, cable/satellite TV, Internet, cell phone etc., they can be very difficult to find, especially if they move often and work in industries where cash tips and short term contracts are common (construction, bars/nightclubs, etc.).
If they are sophisticated enough to craft counterfeit ID using computer programs and laser printers, or engage in Identity Theft, AND they lie to their flat-mate about their identity, they will often only be found by lucky accident. If they engage in international travel, they may locate assets in other jurisdictions, and access them through the ATM network.
People with track records in credit and identity fraud are often quite difficult to find, as are people "connected" to organized crime, especially of the international variety. That being said, anyone can be found, if one is willing to spend enough, and look for long enough.
I hope this was of some help.