Sometimes you should follow your genealogy hunches. My definition of a genealogy hunch is when many of the facts don't fit your ancestor trail, although a few do, but you just have that funny feeling this might be your family tree.
I'm not saying you automatically claim the family as yours. The trick is to do more research on that specific person or family you found, and see if you can prove or disprove your genealogy hunch about them.
As an example, recently I took up a hunt for a branch of my King family that I had put on hold several years ago. The family in question consisted of father Thomas born circa 1835 in Upper Canada (present day Ontario), a wife born in England and seven children all born in Ontario between circa 1859 and 1869. The last sighting of this ancestral branch of my family tree was in 1871.
I had never been able to find them in the 1881 Canadian census or the 1880 American census, or any later genealogy records. I searched land records, vital statistics (births, marriages, deaths) and census records. The problem of finding them was made even more challenging due to the common surname KING and the common first names of the children - James, Elizabeth, Thomas and so on. Often there were several individuals who might or might not be the correct ancestor.
This weekend I pulled out the old genealogy notes and went back at it. This time I determined that I would find them in 1880/1881 census, that I would stick to that census until something turned up. Hours later and still nothing. I was becoming increasingly frustrated.
I used both FamilySearch.org and Ancestry.com to hunt in the USA 1880 census and only FamilySearch.org for the 1881 Canadian census. I used both because each website has a different type of search engine. I used wildcards. I used soundex. I broadened my search and tried various methods - no first name, just an approximate year of birth and birth location. I tried each child (no parents). I tried a child with a father's name. A child with only a mother's name. I used approximate dates of birth and locations of birth with no first or last name. No search trick that I know turned up the family.
But I had noticed a family living in Michigan which was very tantalizing. The names of husband and wife and all the children fit with my missing family. The years of birth were also correct. But the problem was the birth locations! The father Thomas King gave his place of birth as New York instead of Ontario. He also said his parents were both born in New York where my Thomas' parents were both born in England. His wife said she was born in Ohio not England and the children's births were given as Ohio and Michigan. I checked the image in case the transcript was in error. It clearly showed the locations transcribed. But I kept coming back to that family, and was unable to discard them.
Finally I took my own advice and began tracking that family, following them forward in the Michigan census for 1900, and having a hunt on the Pilot FamilySearch.org website in the Michigan Vital Statistics. Bingo!
The 1900 Michigan census showed Thomas as being born in "Canada Eng" (which is English speaking Canada, generally Ontario), father and mother born England but with a different wife. That threw me but I turned to the Michigan Vital Statistics on FamilySearch.org and quickly found the marriages of several of the children of Thomas and his wife. Since her maiden name (Daville) was given in those records, I knew I had the right family. The next piece of the puzzle was the find of Thomas marrying to a second wife just before that 1900 census was taken. In his marriage record his father's and mother's names were given and they were the correct parents.
I spent an enjoyable evening tracing Thomas and his family in Michigan, and was even able to find some obituaries of some of his grand-children. What a lucky break but it may not have happened had I not ignored the glaring inconsistencies in those birth locations in the 1880 census. Following my genealogy hunch, and tracing the family I thought might be the right one proved to be the key. Finding those primary source records that proved the family to be the family branch of my ancestor tree was what was required and with a few more hours of research it all fell into place.