Organising Your Genealogy Research: Files, Names and Citations That Scale

A simple system for organising genealogy research: file naming, folder structure, source citations and a research log that still works when your tree has 2,000 people.

A family tree with fifty people fits in your head. A tree with two thousand people does not, and the difference between a hobby and a headache is the system you use to organise what you find. This is a minimal system that scales.

One tree, one master file

Keep a single master tree in one program and treat every other copy (online trees, shared files, the chart on the wall) as an export. When you find something new, it goes into the master first. Export a GEDCOM every time you finish a research session and keep the last few versions.

A folder for every surname, a file name for every document

Create one folder per surname and, inside it, one folder per couple. Name scanned documents so that they sort themselves:

• 1901-03-12_birth_Kowalska-Anna_Krakow.jpg
• 1925-06-20_marriage_Kowalski-Jan_Nowak-Maria_Warsaw.pdf
• 1930_census_Kowalski-household_Lodz.jpg

Date first, then the event, then the people, then the place. You will never wonder what a file contains again.

Cite sources as you go

A citation answers three questions: what is the record, where is it kept, and how did you see it. A workable format for a parish register:

Baptisms 1895–1910, St Mary's parish, Kraków, entry 42 of 1901; digital image, Polish State Archives (szukajwarchiwach.gov.pl), accessed 3 March 2026.

Put the citation on the fact in your program, not in a separate document. When you export the tree, the citations travel with it and appear as footnotes in your book.

Keep a research log

A research log is a table with four columns: date, source searched, what you were looking for, and result. Include failures. Searched marriages 1920–1930 in Łódź parish, no Kowalski saves you from repeating the same search a year later.

Mark certainty honestly

Most programs let you flag the quality of a source or add a note such as probable or unverified. Use those flags. A tree in which every fact looks equally certain misleads everyone who reads it, including you.

Separate living and deceased

Keep birth dates of living relatives in your master file, but export a privatised version for anything you publish. Our article on privacy in family trees explains what to hide and why.

Review once a year

Every January, pick one branch and re-read it from the sources. You will find a merged duplicate, a misread date or a link you would no longer accept. Cleaning as you go keeps the tree trustworthy.

Print to check

Nothing exposes errors faster than a printed chart. A person born before their parent, a marriage at the age of eight, a child appearing in two families: on a diagram they jump out at you. Generating a quick chart after every big session is the cheapest proofreading you will ever get.