How to Confirm a DNA Match: The Full Step-by-Step Workflow
Confirming a DNA match means proving how you're related, not just that you are. The shared centimorgan (cM) total narrows the possibilities to a range — it never names a single relationship on its own. To confirm a match you stack four more layers on top of the cM figure: shared matches, ages and generations, which side of your family the match falls on, and documentary records. Where a chromosome browser is available, triangulation adds the final piece.
That's the whole method in a paragraph, and the rest of this guide walks through each layer with real numbers and realistic timelines. If you want the shortcut first, run your shared cM figure through the free DNA match calculator — it turns a number into the list of relationships it could actually represent, which is where every confirmation starts.
The five-step confirmation ladder. Tap each step to see exactly what to do:
Tap a step above.
Why a Centimorgan Total Alone Can't Confirm Anything
A centimorgan measures the length of shared DNA, and the amount you share with a relative is partly random. Beyond parent and child — where the figure is essentially fixed — every relationship produces a range, and those ranges overlap heavily.
Share 1,750 cM with someone and you could be looking at a half-sibling, a grandparent, a grandchild, an aunt, an uncle, a niece or a nephew. All of those sit near that average. The DNA cannot tell them apart, because the DNA doesn't record direction, generation or which parent it came through. It only records quantity.
This is the single most common mistake new testers make: they see a label like "Close family — 1st cousin" on their results page and treat it as a verdict. It's a guess generated from the cM total, and the testing company is transparent that it's a prediction. Our guide to what a centimorgan actually is explains why the randomness exists in the first place.
So the honest starting position is: the number gives you a shortlist. Everything after that is genealogy.
Step 1: Turn the cM Number Into a Shortlist
Start by writing down the exact shared cM total and the number of shared segments. Ignore the company's relationship label for now — you'll come back to it.
Then look the figure up against the Shared cM Project, the crowd-sourced dataset behind the tool most researchers use at DNA Painter. It collects tens of thousands of real, documented relationships and reports the observed range for each one, not just a theoretical average.
What each cM band typically means
| Shared cM (approx.) | Average relationship strength | Relationships that commonly fall here |
|---|---|---|
| 3,400+ | Identical | Identical twin, or the same person tested twice |
| 2,300–3,500 | Immediate family | Parent/child (~3,475, near-fixed), full sibling (avg ~2,600) |
| 1,600–2,400 | Very close | Full sibling (low end), half-sibling, grandparent/grandchild, aunt/uncle, niece/nephew |
| 1,150–1,600 | Close | Half-sibling, grandparent, aunt/uncle, niece/nephew (all avg ~1,750) |
| 575–1,330 | Cousin-level close | First cousin (avg ~866), great-aunt/uncle, half aunt/uncle, first cousin once removed (high end) |
| 200–620 | Mid-range | First cousin once removed (avg ~433), half first cousin, first cousin twice removed |
| 90–360 | Distant-ish | Second cousin (avg ~229), half 1C1R, 2C1R (high end) |
| 40–200 | Distant | Second cousin once removed (avg ~122), third cousin (avg ~73) |
| Under 40 | Very distant / uncertain | Third to fifth cousin, half versions of the above, or a coincidental match |
Two honest caveats about this table. First, the boundaries are soft — a full sibling can come in at 1,613 cM and an aunt can come in at 2,400, and both are documented. Second, under about 40 cM you're in territory where a chunk of matches are "identical by chance" rather than by descent, which means no genealogical relationship exists at all in a recent, findable timeframe.
The shared cM chart guide goes deeper on reading those ranges. For a fast answer, plug your number into the DNA match calculator and it'll list the candidates for you.
Step 2: Use Shared Matches to Pick a Side
This is the step that does the heaviest lifting, and it's free on every major platform. Every site offers some version of it — AncestryDNA calls it "Shared Matches," MyHeritage and FamilyTreeDNA call it "in common with," 23andMe shows "Relatives in Common."
Open the shared-match list for your mystery match and look for people you've already identified.
- If your known maternal cousins appear, the match is on your mother's side.
- If your known paternal cousins appear, it's your father's side.
- If nobody you recognize appears, the match sits on a branch you haven't mapped yet — which is information too.
Going one level deeper
Once you know the side, repeat the trick within that side. Do the shared matches cluster around your maternal grandfather's surname group, or your maternal grandmother's? Each round of clustering cuts the shortlist roughly in half.
When you have dozens of unassigned matches, this is worth doing systematically rather than one at a time. The Leeds Method sorts matches into four color groups that usually correspond to your four grandparents, using nothing more than a spreadsheet and the shared-match lists you already have. It's the single best afternoon of work you can do on a messy match list.
One limitation to know: shared-match tools have thresholds. AncestryDNA only shows shared matches at 20 cM and above, so genuinely distant relatives can share matches you'll never see listed. Absence of shared matches isn't proof of anything.
Step 3: Let Ages and Generations Narrow It Further
Birth years quietly eliminate more possibilities than most people realize, and they cost nothing to check. Look for the match's approximate age on their profile, in their tree, or in an obituary.
Say you share 900 cM with someone. The candidates include first cousin, great-aunt, half-aunt, and first cousin once removed. Now add ages:
- Match is 30 years older than you — half-aunt, great-aunt or 1C1R (the older direction) become the front-runners; a first cousin becomes unlikely.
- Match is within a few years of your age — first cousin or half-first-cousin-ish relationships move up.
- Match is 30 years younger — 1C1R in the younger direction, or a half-niece.
The logic is simple. A generation is roughly 25 to 30 years on average, so each generation of difference between two people shifts the likely relationship by one step. It's not a rule — people have children at 19 and at 45 — but as a tiebreaker between two relationships that share the same cM range, age is excellent.
If the "removed" language is what's tripping you up, our plain-English explainer on first cousins once removed untangles it in a couple of minutes.
Step 4: Build Quick Trees and Check the Records
Now you leave DNA behind and do actual genealogy. This is the step people skip, and it's the step that produces proof.
Build a "quick and dirty" tree for the match
You don't need a complete, sourced tree. You need enough of one to spot where their line and your line meet. Take the match's known ancestors — from their posted tree, their surname, their location, an obituary — and work backwards three or four generations as fast as you reasonably can.
Useful free sources for this: FamilySearch, FindAGrave, published obituaries, and census records. Our roundups of free genealogy websites and the FamilySearch guide cover where to look without paying for a second subscription.
Look for convergence
When you've built quick trees for two or three of the match's closest relatives, look for a couple who appears in all of them. That convergence point is almost always your common ancestral couple. Then check whether that couple connects to your own documented tree.
Prove it with paper
A hypothesis becomes a confirmation when independent records agree:
- Census records placing the right people in the right household at the right time
- Birth, marriage and death certificates naming parents
- Obituaries listing surviving relatives (these are gold — they name siblings, children and in-laws in one paragraph)
- Wills and probate files naming heirs and their relationships
- Church records for baptisms and marriages, especially pre-census
When the DNA range, the shared matches, the ages and the paper trail all point at the same couple, you've confirmed the match. That's the standard. Not one of them alone.
Step 5: Triangulate on a Chromosome Browser (Where You Can)
Triangulation is the strongest technical confirmation available to hobbyist researchers. The idea: if you, your match, and a third known relative all share the same segment on the same chromosome at the same start and end points, that segment almost certainly came from one shared ancestor.
Two people overlapping proves very little — you could have inherited that stretch from opposite sides of the family. Three people overlapping on a meaningful segment is a different story.
Where triangulation is possible
| Platform | Chromosome browser | Triangulation possible? |
|---|---|---|
| AncestryDNA | No | No — use shared matches and trees instead |
| 23andMe | Yes | Yes, with relatives who've opted into sharing |
| MyHeritage DNA | Yes | Yes, and it flags triangulated groups directly |
| FamilyTreeDNA | Yes | Yes, including a dedicated matrix view |
| GEDmatch | Yes | Yes — the most flexible tools, free tier included |
AncestryDNA has no chromosome browser. That's the most important practical fact in this section, because Ancestry has the largest database and is where most people test. On Ancestry you confirm matches with shared matches, ThruLines hints and documentary research — and it works fine, it just takes more tree-building. If you want segment data, download your raw file from Ancestry and upload it free to MyHeritage, FamilyTreeDNA and GEDmatch.
A sensible segment threshold
Small segments are unreliable. Many experienced researchers won't triangulate below about 7 cM, and treat anything under 10–15 cM with caution, because false segments are common at that size. Bigger segments, fewer headaches. Our chromosome browser explainer walks through reading the display itself.
What a Confirmed Match Actually Looks Like
Here's a worked example, kept deliberately ordinary.
You share 442 cM with a match named R. The calculator says the likely candidates are first cousin once removed, half first cousin, first cousin twice removed, or half aunt/uncle. Four real possibilities.
Shared matches show three cousins you've already confirmed on your father's mother's line. Side settled: paternal, grandmother's branch. Half the candidates drop away immediately.
R's obituary-derived birth year is 1955; yours is 1981. That 26-year gap points at one generation of difference, which favors first cousin once removed over half first cousin.
You build R's tree back three generations and find R's grandparents are the same couple as your great-grandparents — your paternal grandmother's parents. The census confirms the household. A marriage record confirms the parents' names.
Conclusion: R is your first cousin once removed, on your paternal grandmother's side. The cM total, the shared matches, the age gap and the records all agree. That's a confirmed match, and it took maybe two evenings.
How Long Should Confirming a Match Take?
Being realistic about this saves a lot of frustration.
- A close match with a public tree — often under an hour. The tree tells you the answer and you're just verifying it.
- A close match with no tree but an identifiable name — an evening or two, mostly obituary and census work.
- A mid-range match (200–500 cM) with no tree — several days to a few weeks of on-and-off work, building quick trees for their relatives until something converges.
- A distant match (under 100 cM) — potentially months, and sometimes the honest answer is "unresolved." Plenty of distant matches never get confirmed, and that's normal.
Two things speed everything up: more testers in your family (each confirmed relative becomes a reference point for every future match), and testing on more than one platform.
When You Can't Confirm It
Sometimes the evidence runs out. That's a legitimate outcome, and recording it honestly is better research than forcing a conclusion.
Say so in your notes: "Shares 96 cM. Maternal side per shared matches. No tree, surname unknown, no convergence found. Revisit if new close matches appear." Then move on. New people test every week, and a case that's stuck today can crack open the moment one new relative uploads their results.
A word on unexpected results
Confirming matches sometimes turns up things nobody was expecting — a parent who isn't the parent on the birth certificate, a sibling nobody mentioned, an adoption that was never discussed. This happens more often than people assume, and it can be genuinely destabilising for everyone involved.
If you find yourself in that position, slow down. There's no deadline. Confirm your findings carefully before drawing conclusions, think about who else is affected, and consider whether the information is yours alone to share. Kindness costs nothing and unrings no bells.
FAQ
How many centimorgans confirms a relationship?
No specific number confirms a relationship on its own. Above about 2,300 cM you can be confident it's a parent, child or full sibling, but every other range covers several possible relationships. Confirmation comes from combining the cM total with shared matches, ages, family side and documentary records.
Can I confirm a DNA match on AncestryDNA without a chromosome browser?
Yes. Ancestry has no chromosome browser, so you confirm matches using shared matches, ThruLines hints, tree-building and records instead. It's a well-established approach and it works — it simply requires more genealogy and less segment analysis. If you want segment data as well, upload your raw Ancestry file free to MyHeritage, FamilyTreeDNA or GEDmatch.
What does triangulation mean in DNA testing?
Triangulation means three or more people all sharing the same DNA segment at the same chromosome location, which indicates the segment came from one common ancestor. Two people overlapping isn't enough, because the overlap could come from different sides of your family. Most researchers only triangulate segments of roughly 7 cM or larger.
Why do my DNA match's predicted relationships keep changing?
Predicted relationships are calculated from the shared cM total, and companies update their algorithms and reference panels periodically. The underlying DNA hasn't changed — only the label attached to it. Treat every prediction as a shortlist, not a conclusion.
Should I message a DNA match to confirm the connection?
Often yes, and many matches are delighted to help. Keep the first message short, friendly and low-pressure: how much you share, what you think the connection might be, and one specific question. Never open with life-changing news — if you suspect a close, sensitive relationship, take time to be certain first and consider how the message will land on their end.
What if a match's tree contradicts the DNA?
Trust the DNA over the tree. Online trees are frequently copied without sources and errors propagate fast. When shared cM and shared matches point one way and a posted tree points another, the tree is usually the thing that's wrong — go back to primary records and rebuild that branch yourself.
Confirm It, Don't Guess It
The workflow is always the same, whether the match shares 2,000 cM or 60. Get the range, pick a side with shared matches, use ages to break ties, build quick trees until something converges, prove it with records, and triangulate if your platform lets you. Five layers, each one narrowing what came before.
Start where every confirmation starts: run your shared cM figure through the free DNA match calculator to see exactly which relationships are on the table. Then use the common ancestor chart to sketch where the connection has to sit — and go find the paper that proves it.




