A positive "track 'em down" attitude is important. Avoid discouragement or thinking that genealogical research in a "burned county" is hopeless. The first and most important step to finding ancestors who lived in a burned county is creating a relentless "track 'em down" mindset no matter how hard it is, and no matter how long it takes.A more insidious issue occurs when a researcher encounters gaps in any particular record. The most common gap encountered is in the U.S. Federal Census records for 1890. Contrary to popular belief, the 1890 Census was not entirely destroyed by fire. Parts of this census were destroyed by fire but the remaining documents were maintained for many years and then finally destroyed by an act of the United States Congress when the head of the Library of Congress failed to recommend that the records be preserved. If you would like to read the real story about the destruction of the 1890 Census see "First in the Path of the Firemen"The Fate of the 1890 Population Census, Part 1" By Kellee Blake.
Of course, as your research extends further into the past, you will undoubtedly find more and more records missing of any particular series. For example, if you examine Catholic parish registers in Mexico you will find that some extend back into the 1500s while others only extend into the 1800s. Without doing a substantial amount of historical research, you might expect that the records were missing. However, the real reason was likely that particular parish you're looking at was not active prior to the earliest records. Any pertinent records were likely kept in some other parish. Sometimes the records are not missing due to any catastrophic occurrence merely through neglect.
Following the suggestions of the Research Wiki, the next suggestion concerns research logs:
Research logs are a must. Burned county research is not easy. The more difficult or complex a research problem becomes, the more valuable good research logs are. They are a cornerstone to good research and filing systems. Research logs show what has, and (just as importantly) has not worked. Also, use logs to explain in written comments your thinking about search strategies and what you are trying to accomplish—why you turned to that set of records.Although this paragraph refers to "burned counties" it is also applicable to the situation where any records are missing for any reason. In some instances, such as my example for Mexico, there are civil registration records that were kept contemporaneously with the parish registers. In a place like Mexico, you can run out of records rather rapidly. Carrying on with the example from Mexico, if the records were lost or missing, you might become aware that copies of all of the parish records, before independence, were sent to Spain and are in the Spanish Archives. So lost can be relative. In the United States it would also be rare not to have alternative records to search even when certain types of county records have been destroyed.
Focusing on one type of record to the exclusion of others is usually at the root of the problem. Turning to the burned counties problem, certain types of records must be reconstructed for the county to operate. Counties cannot exist without land records, tax records, marriage records, and many other types of records. In addition, in many places throughout the United States, copies of land records are kept by title companies and abstractors. If you realize how land sales work in the United States, you would immediately realize that a courthouse fire would merely be a temporary impediment. Land records would have to be reconstituted in order for property to be bought and sold in the county.
Lost records or missing records are really an invitation to extend your understanding of the type of records that could have been maintained and likely were maintained that cover the gaps in your favorite records. Follow the rest of the extensive suggestions on the Research Wiki page for Burned Counties.