CLUSTERING – What It Is, What It Isn’t, and How to Do It 

This is a walk-through of the basic manual clustering metod, followed by other clustering options you can explore. 

At its heart, Genetic Genealogy only has three steps:

1. Clustering – grouping your matches into logical piles
2. Tree Building – quick & dirty trees, research trees, hypothetical trees, floating branches – (always use a private & unsearchable tree)
3. Working through your hypotheses – the real detective work

Clustering is simply step 1. It’s the filing cabinet. The genealogy that comes afterwards is what solves cases.

THE CLUSTERING METHOD (simple and effective)

Start with your first match that isn’t close family – roughly 400 cM or below. (These people usually share one grandparent with you, so they’re excellent “anchor” matches.)

If you’re working on a specific distant line, start around 100 cM (likely someone who only shares one great-grandparent with you)

1. Choose your first match and create a new group
   • Give them a coloured dot/square
   • Note: Ancestry, MyHeritage, and LivingDNA all have these tools. For others, use a spreadsheet and colour the cell
2. Open their Shared Matches
   • Add all shared matches down to about 30 cM into the same group.
   • Below ~30 cM you start picking up people who match each other for reasons unrelated to how they match you.
   • That’s your Cluster A.
3. Return to your match list
   • Scroll down to the next person without a colour.
   • Give them a new colour and repeat the process with their shared matches.
   • This becomes Cluster B.
4. Repeat for Cluster C, D, E…
   • Work your way down to about 30 cM.
   • You can go to 25 cM, but be cautious — those matches often represent much older connections.
5. Stop. You’re done.
   • Clustering should rarely takes more than an hour – sometimes much less.

Don’t overthink the clusters — clustering is an organisation tool!

THE REAL WORK STARTS NOW: GENEALOGY

Once you have your clusters:

1. Build trees for Cluster A (use a private & unsearchable tree – these trees at Ancestry are not indexed in their database, so essentially cannot be seen by anyone).
2. Research their trees, build quick trees – often referred to as 'Quick & Dirty' trees – for those without them, and look for the shared ancestors.
3. Eventually you’ll notice that Cluster A converges on the same ancestral couple.
   • Those are your ancestors — even if you don’t yet know how you connect.

Tools like WATO at DNAPainter can help you weigh up the possibilities. https://dnapainter.com/help/user-guide/wato-plus

Clustering doesn’t solve mysteries. Genealogy does.

WHY WE CLUSTER

Clustering has two simple purposes:

1. Organisation – turning hundreds of matches into tidy, logical groups
2. Genetic networks – each cluster is likely descended from the same ancestral couple, so you can research them together

That’s it. No magic. No shortcuts. Just structure.

TOOLS YOU CAN USE

Manual clustering:

• Coloured dots on Ancestry / MyHeritage / LivingDNA
• Spreadsheets (colour-coded rows/cells) – like the Leeds Method (which is a method to produce grandparent groups from a small sub-section of matches 400cM -> 90cM – more suited to US testers); extend this to 30cM for the traditional clustering.
• Pen & paper with highlighters (still perfectly valid!)

Automatic clustering (optional):

• Ancestry auto-clusters (standard & custom, this requires a Pro-Tools subscription)
• MyHeritage AutoClusters
• GEDmatch AutoClustering (requires Tier 1 level access)
• GeneticAffairs.com (various cluster styles)

These automated methods can highlight patterns you may have missed manually, but manual clustering is still the best starting point for most people.