Start With the Shape of Your Match List

Your match list is not a leaderboard, and it is not a family tree. It is a set of clues, and the first thing to do is stop scrolling and look at its overall shape. How many matches do you have above roughly 200 centimorgans (cM)? Are they clustered in one branch of your tree or spread evenly across all four grandparent lines? A British tester with deep local roots will often see thousands of distant matches and very few close ones, which is normal and not a sign that something has gone wrong.

Before you analyse anything, sort your matches into three rough groups: close (generally above 200 cM), mid-range (roughly 40–200 cM) and distant (below 40 cM). Give each match a short note in the platform's notes field so you remember what you concluded six months ago. Genealogy moves slowly, and a tidy list beats a brilliant memory every time.

Shared Segments and Centimorgans: What the Numbers Mean

Centimorgans measure the length of DNA you share with another person. A shared segment is a continuous run of matching DNA on a chromosome; the cM figure is that run's length. Longer segments are far more significant than numerous tiny ones, because very short segments are frequently identical-by-chance rather than identical-by-descent.

• Parent–child: roughly 3,400–3,700 cM
• Full siblings: roughly 2,300–2,900 cM
• Grandparent, aunt or uncle: roughly 1,150–1,750 cM
• First cousin: roughly 550–1,150 cM
• Second cousin: roughly 200–400 cM
• Third cousin: roughly 50–150 cM
• Fourth cousin: roughly 20–50 cM

Treat anything below about 7 cM with caution, and be sceptical of segments under 10 cM unless they are supported by documentary evidence. Populations with endogamy — including many Irish, Jewish, Acadian and island communities — will show inflated totals, because shared ancestry many generations back produces extra matching segments that mimic a closer relationship than actually exists.

Why Predicted Relationships Are Only a Starting Point

Every testing platform offers a predicted relationship, and every one of those predictions is a range dressed up as a single answer. A match labelled "first cousin" might in truth be a half-first cousin, a first cousin once removed, or a great-grandparent's sibling's descendant. The cM total narrows the field; the family tree you build closes it.

A practical technique is to work in both directions at once. Note the match's cM figure, then ask what relationships in your tree plausibly produce that figure and fit the age gap. A shared total of 850 cM between people thirty years apart in age points somewhere very different from the same total between contemporaries. Age, geography and naming patterns are evidence too.

Ethnicity Estimates, Chromosome Browsers and Triangulation

Ethnicity estimates are the most entertaining part of a DNA test and the least reliable for genealogy. They are calculated by comparing your DNA with reference populations, and those references change as more people test. A 4% region can vanish from your report overnight. Use ethnicity results as a loose hint about migration history, never as proof of a specific ancestor.

The chromosome browser is where the genuine work happens. If you can see where you match a person, you can group matches into segments and begin assigning segments to ancestral couples. The gold standard is triangulation: three or more matches who all share the same overlapping segment on the same chromosome, and who all descend from a known common ancestor. Two matches sharing a segment is suggestive; three with a documented ancestor is compelling.

Contacting Matches With Confidence

Most matches never reply. That is not personal — many people test once out of curiosity and never log in again. Still, a good first message dramatically improves your odds.

• Keep it under 150 words and put the specific question in the first two lines.
• State your shared cM figure and predicted relationship, so they can verify you are genuine.
• Give two or three surnames and places from the relevant branch, not your entire tree.
• Ask one clear question — for example, whether they recognise a particular couple in a particular parish.
• Offer something back: a photograph, a certificate transcription or a summary of what you have already found.
• Never open with "you must be related to me" or a demand for their parents' details. Be warm, be brief, and give them an easy way to say no.

If a close match does not reply, look for them elsewhere — public trees, obituaries, electoral registers — and consider writing a letter by post. It feels old-fashioned and it works surprisingly often.

Building the Paper Trail Behind Every Match

DNA tells you that two people are related. Only documents tell you how. For every match you place in your tree, record the shared cM, the chromosome and segment start and end points if available, and the documentary source that confirms the connection — a birth certificate, a census return, a parish register entry.

Keep a simple research log, even a spreadsheet, listing the match's name or identifier, the date you checked, the cM total and your current hypothesis. Revisit it after each new match arrives. Patterns emerge slowly: a cluster of matches all pointing at one Welsh parish, or a stubborn segment that refuses to sit on any known line. That stubborn segment is not a failure. It is the next question, and honouring the questions is how family history rewards patience.