Should I Use Filtered or Bottled Water for Tea? | Taste Test

The last time I visited my favorite tea house, I asked my tea dealer a question.

"Why is it that every time I buy tea here and make it at home, it doesn't taste as good as when you do it?"

"Simple," she answered. "It's the water."

"But couldn't it also be your technique or the equipment you use? Maybe you're doing something right that I'm doing wrong."

That's when she told me about another customer with a similar question. "I made some tea at his home with the same equipment that I have here. I did everything the same. But he didn't have my water, so the tea was different."

These differences are subtle, the kind you might not notice if you're brewing flavored tea, making a pitcher of iced tea, or planning to add lots of milk and sugar. But in fine tea, the kind that's meant to be drunk straight, subtlety is everything. Brew a tea one way and it might taste nice. Brew it with some extra care and it can floor you with its complexity, aroma, and finish that lingers for minutes after each sip.

Ask tea experts about brewing and they'll tell you how your source of water is critical, how better water turns a good tea great and bad water can make even pricey tea taste sour and acrid. Anecdotally I've seen this claim in action, but I've never put numbers to it—until now. Can we quantify the impact of water source on tea?

The Claim

While advice varies from tea person to tea person, the general argument all comes down to what's in your water besides water.

That includes minerals endemic to a water supply, additives like fluorine, and even dissolved solids from old pipes. Too many minerals and your water is "hard," which can add funky metallic flavors to tea. Absolutely pure water like distilled is no good either, tea people claim—it's so bland it can make tea taste dull. Water for tea should have a neutral pH so it doesn't turn the tea sour or bitter. And it should be freshly boiled, as water that's been boiled again and again can absorb off flavors from the air around it.

That's a lot to think about before you even pick out a tea to brew. And it's a lot of variables to test. But a good starting point is the three sources of water most people have at their disposal for brewing tea: tap, filtered (I'm talking about the basic charcoal filtrations like Brita), and bottled.

Most tea experts I know vouch for some kind of bottled spring water or home filtration system. Spring water has a relatively neutral pH and a Goldilocks balance of minerals—not too many, not too few. Filters are especially valuable in regions where the tap water is hard or otherwise funky-tasting. If you wouldn't drink your tap water as water, you probably wouldn't want to drink it as tea.

My tea shop uses an especially pricey Japanese iWater system that filters out heavy metals, chlorine, and other chemicals. "We've been using this filtration system from beginning," my tea dealer tells me. "I tried other filtered and bottled spring water but haven't seen anything work better than this."

But for now let's keep to the basics: how do tap, filtered, and bottled water stack up for your average cup of tea?

The Test

Our tasting had three rounds, each for its own kind of tea: a black English Breakfast, a green tea, and an oolong. Both the black and green teas came in upmarket packaging and were sold at a bougie supermarket—they clearly advertised themselves as something "fancy," though high-end tea drinkers may not consider them so. As for the oolong, it was inarguably fancy—a 30-year aged bao zhong from a small farm in Taiwan, one of the teas Leandra and I keep around the office for when we want a taste of the good stuff.

We incidentally wound up with three quality levels of tea: one decent, one not-so-decent, and one great. In a pre-test tasting I found the black tea perfectly drinkable but ordinary. The green tea was overly grassy and easily developed bitter flavors if the tea bag was left in a cup for too long, a common problem for low-quality green teas. The oolong, as usual, was great. I was curious to see how these differences would play out in our results.

I brewed each tea in four pots of water: one tap, one filtered, and two kinds of bottled. Each sample was heated to the same temperature, had the same concentration of tea, and steeped for equal times. The green tea brewed around 170°F; the black and oolong just off boiling.

Tasters rated their preferences on a ten-point scale and left comments about the teas' relative bitterness and complexity of flavor. They tasted blind, i.e. they were unaware of the different water sources in the four samples.

The Results

Each round of tea told its own story.

With the decent black tea, the bottled waters scored almost equally well while both tap and filtered received equally low scores—over a three point difference on a 10-point scale. Tasters found the tap and filtered samples "exceedingly bitter and astringent," "harsh and sour," even "horrible." Meanwhile the bottled water samples received remarks like "toasty in a good way," "faint floral taste," and "hints of chocolate and licorice."

The bitter-lawn-clippings green tea scored low marks across the board. But the comments were far kinder on the bottled samples, with notes of "chrysanthemum," "floral," and "clover," and descriptions of a more balanced bitterness. The tap and filtered water samples got comments like "tastes like old tea in a bad Chinese restaurant."

And what about the fancy-pants aged oolong? Unsurprisingly it scored generally well. Ratings for three samples—both bottled waters and the tap—fell within a point of each other. (Curiously the filtered water sample scored much lower.) Tasters praised the bao zhong's "fruity," "Christmas" flavor and its "delicious creaminess." But the tea brewed with filtered water "tastes like it was filtered through an old sock."

So Does Water Matter?

Yes, but it's not the only thing that does.

These results suggest an interaction effect between the quality of a tea and the quality of water used to brew it. Overall, tasters quite liked the bottled water samples, but when a tea is truly unpalatable, no source of water can save it. On the other hand, if you have a fine tea on your hands, you might like it no matter the water source, but better water may help.

My personal experience of the tasting is biased—I knew all about the nature of the test and the identity of the samples. But I was shocked at the consistency of my scores. In every round I rated one particular brand of bottled water—Eternal—as my favorite, and I was astonished by the complexity it brought to each tea. A kinda dull black tea suddenly tasted rounded, malty, and sweet; an otherwise bad green had a new-found balance of bitter and creamy flavors; and the oolong I thought I knew took on a new dimension of...is that fresh biscuits I smell?

In two of three rounds, tasters couldn't detect a difference between filtered water and tap, and when they did, filtered water scored significantly worse. But that result may not be true everywhere. New York City tap water is some of the best in the country: renowned for its clean, neutral flavor and balance of minerals. The difference between filtered and tap here may be small compared to other regions with harder tap water, hence the rule I mentioned above: if you wouldn't drink your tap water as water, don't brew it as tea.

What'll I do in the future? For one, I'll stick to buying good tea, as even the best water on the planet can't make bad tea taste good. And I'll keep a few bottles of that fancy Eternal spring water for when I want a good tea to really shine. But I'll hold off on the Brita filter, at least in New York.

Our Tasting Methodology: All taste tests are conducted completely blind and without discussion. Tasters taste samples in random order. For example, taster A may taste sample 1 first, while taster B will taste sample 6 first. This is to prevent palate fatigue from unfairly giving any one sample an advantage. Tasters are asked to fill out tasting sheets ranking the samples for various criteria that vary from sample to sample. All data is tabulated and results are calculated with no editorial input in order to give us the most impartial representation of actual results possible.