Why DNA reaches where paper records cannot

Brick walls usually form for dull, human reasons: a baptism register with no father named, a family who moved three counties in a decade, a surname spelled four different ways by four different clerks. DNA does not read handwriting and it does not care about spelling. It measures inheritance, and it can point you towards a branch of a family you would never have thought to search. Think of your match list as a new kind of index — not a list of names, but a set of cousins who all quietly point at the same couple.

The trick is to stop treating the match list as a curiosity and start treating it as evidence. That means organising it, measuring it, and then doing the patient work of turning genetic proximity into a place and a name.

Get organised before you get excited

Before you analyse anything, spend an evening making your matches sortable. A simple spreadsheet is enough. For each match worth keeping, note their display name and username, the shared centimorgans (cM), the number of segments, whether they have a tree, and the surnames and places you can see. Add a column for the date you last looked, because new matches arrive every week and a wall you cannot climb today may have a key next month.

Then colour-code. This is the single most useful thing you can do:

• Take your closest match and assign them a colour. Mark every match who shares with them in that same colour.
• Take the next uncoloured match and repeat with a new colour.
• Keep going until most matches have a colour. You will usually end up with four clusters, roughly your four grandparent lines.
• Once you recognise a shared ancestral couple inside a cluster, give that cluster a surname.

Ignore anything under about 15 cM for now. Those matches are noisy, and chasing them is how people lose a year.

Reading shared segments without the jargon

Shared cM gives you a rough relationship band. A parent and child share around 3,400 cM; an aunt or uncle around 1,700; a first cousin around 866; a second cousin around 229; a third cousin around 73; a fourth cousin around 35. These are averages, not rules, and ranges overlap — which is why you should never announce a relationship on cM alone.

Segment size matters as much as the total. One 60 cM segment is a far stronger signal than six segments of 10 cM, which may simply reflect a shared regional ancestry rather than a recent common ancestor. The gold standard is triangulation: three or more matches who share the same segment on the same chromosome, and who can all be traced on paper to one ancestral couple.

Two warnings. Endogamy — common in small parishes, island communities, and some religious or immigrant groups — inflates shared cM and creates false closeness. So do "pile-up" regions, where the same stretch of chromosome is shared by thousands of unrelated people. If a match looks too close for the paper trail, suspect endogamy before you suspect a scandal.

Build quick trees for your matches

Most matches have no tree, or three names and a question mark. Build a mirror tree anyway: their names, their places, worked backwards as far as the records allow. Label it clearly as a working hypothesis, never publish it as fact, and keep it separate from your own tree.

Then widen it sideways. The descendants of your brick-wall ancestor's siblings are the people you match — so when you find a candidate couple in the right parish, chase every child, not just your direct line. A marriage in the neighbouring village in 1841 is often the answer you have been hunting for a decade.

Contacting cousins: warm, brief, patient

A good first message is under 150 words. Say who you are, name the match or segment you share, mention one specific surname or place, and ask a single easy question. Offer something — a transcription, a photograph, a copy of your research — rather than requesting access to theirs.

• Do not open with a theory about their parentage, however confident you are. Bricks fall on people unexpectedly.
• Remember the account holder may not be the tester: a daughter may manage a late father's kit, or a gift may never have been activated.
• Follow up once after four to six weeks, then leave it. Silence is usually inattention, not refusal.
• Never publish living people's details, and do not share someone else's tree or data without asking.

When a match does reply, reciprocity works wonders. Share your findings, credit their help, and ask what they would like to know. The cousins who answer are often the ones who hold the photograph with a name written on the back.

When the DNA rewrites the story

Occasionally the science quietly overturns a documented line: a grandfather who was not the biological one, an informal adoption, a surname taken from a stepfather. Approach it as a research problem rather than a family bombshell. Look for the smallest explanation that fits the evidence, and test it against records before telling anyone. And when you do tell someone, lead with the discovery you can prove, offer support, and let them set the pace. DNA breaks walls — but it is patience, courtesy and a tidy spreadsheet that get you through them.