You spit in a tube, wait a few weeks, and open your results: 42% British, 28% German, 18% Scandinavian, 12% broadly Northwestern European. Your sister takes the same test and gets… different numbers. Same parents. Same grandparents. Same family tree going back centuries. So what's going on?

The short answer is that ancestry tests aren't reading a history book written in your DNA. They're making a statistical estimate based on how your particular slice of genetic code compares to living people around the world today. And that estimate depends on which slice you inherited, which populations the company uses as references, and which algorithm does the math.

When an ancestry company says you're 35% Italian, they don't mean they found an "Italian gene" hiding in your chromosomes. What they mean is that 35% of your DNA closely resembles the DNA of people currently living in Italy who participated in their reference panel. These reference panels are collections of DNA from modern individuals who have deep roots in a particular region—people whose grandparents all came from the same area.

Here's the catch: modern populations are the result of thousands of years of migration, mixing, and genetic drift. The people living in Italy today aren't genetically identical to the people who lived there 500 years ago when your ancestors may have left. So the company is comparing your DNA to a moving target. The reference populations are proxies for historical groups, not perfect mirrors of them.

This also means that if a company has a larger or more detailed reference panel for one region—say, they've collected more samples from East Africa than West Africa—their estimates for those regions will be more precise. Your results reflect the company's database as much as they reflect your actual ancestry. As companies collect more samples and refine their panels, your results can literally change without your DNA changing at all.

Takeaway

Ancestry percentages don't describe what your ancestors were—they describe how much your DNA resembles people living in certain places today. The map is not the territory.

You might assume you carry exactly 25% of each grandparent's DNA. After all, each parent gives you 50%, and each of them got 50% from their parents. The math seems clean. But biology doesn't do clean math. When your parents make eggs and sperm, their chromosomes go through a process called recombination—segments of DNA from their mother's and father's chromosomes get shuffled and swapped before being passed on. The result is that each egg or sperm carries a unique, randomized mix.

This means you might carry 32% of your paternal grandmother's DNA and only 18% of your paternal grandfather's. Your sister, built from a different egg and a different sperm, might carry 21% and 29% respectively. You both have the same four grandparents, but you inherited different random samples of their genomes. Go back another generation to great-grandparents, and the variation gets even wider. It's entirely possible to carry almost none of a particular great-grandparent's DNA.

This is why siblings can get noticeably different ancestry estimates. One sibling might show a stronger Scandinavian signal because they happened to inherit more of the chromosomal segments that resemble the Scandinavian reference panel. It's not that one sibling is "more Scandinavian" in any historical sense—they just got a different hand from the same genetic deck.

Takeaway

You don't inherit a neat, equal fraction from each ancestor. Every conception is a genetic lottery, which is why siblings—despite identical family trees—carry meaningfully different slices of their shared heritage.

Take the same tube of spit and send it to two different ancestry companies. You'll likely get two different breakdowns. One might say you're 40% French, while the other calls that same DNA 36% broadly Western European with 8% Italian. Neither is lying. They're using different reference panels, different statistical models, and different ways of drawing boundaries between populations that, genetically, overlap enormously.

Think of it this way: French DNA and German DNA aren't separated by a bright genetic border. People have moved across that region for millennia. One company's algorithm might lump those similar signals together under "Western European," while another tries to split them into finer categories. The finer the split, the more confident the algorithm has to be—and the more room there is for disagreement. Some companies are conservative, preferring broad labels. Others are aggressive, offering hyper-specific regional estimates that come with higher uncertainty.

This is why geneticists often recommend treating your ancestry results as a rough sketch rather than a photograph. The broad strokes—continental-level ancestry—are usually reliable. But the difference between "38% British" and "42% British" is well within the noise of the algorithm. When your results update after a company revises its methods, it's not that your ancestry changed overnight. It's that someone adjusted the lens.

Takeaway

Different ancestry companies are answering slightly different statistical questions with different tools. Treat the broad patterns as meaningful and the precise percentages as educated guesses.

Ancestry tests are genuinely powerful tools. They can connect you to distant relatives, confirm family stories, and reveal migration patterns you never knew about. But they're estimates built on statistics, not a direct readout of your past.

The most useful way to read your results is with appreciation for what they can tell you—broad patterns of heritage and genetic connection—and a healthy understanding of why the precise numbers will always wobble. Your DNA is real. The percentages are just one way of describing it.