Why a hojicha latte tastes different at a shop and at home

Powder, dissolved, cut with milk

Powder stirred into hot water in a small bowl turns to a thick brown liquid. What separates the café cup from the home cup is not the quality of the leaf or the skill of the hand, but ratio, temperature and foam.

Line up the published recipes and most, café and home alike, are built the same way. Powder is dissolved in hot water to make a concentrate, which is then cut with milk. What differs is a handful of numbers inside that method, and how they are handled.

The recipes show the shape of the method, and no more. After the shape comes the share of milk.

Mostly milk, by weight

Warmed milk poured into the cup gives off a faint sweetness. In the recipes built from powder, more than four fifths of the cup's weight is milk.

One coffee shop's recipe dissolves powder, sugar and water into a syrup and adds warmed milk. A tea maker's recipe for home kitchens likewise dissolves powder in a little water and tops it up with milk. In both, milk makes up almost the whole cup.

That share is calculated from the published weights; the pages do not print it. How powders are made and chosen comes up again under blending. After powder comes the version brewed from leaf.

Brewed from leaf, a little milk

Bōcha in the pot, boiling water over it, and toasty steam from the stems rises. In the version brewed from leaf, milk is only about a fifth of the cup.

A bōcha shop's published latte adds a little milk and sugar to strongly brewed tea. Set beside the powder recipes, the share of milk differs widely under the same name. It is another way to build the drink, with the tea in the lead.

Which is better is not something the recipe numbers decide. After the ratio come heating and foam.

Heat and foam in the pan

Whisk the milk in a small pan and fine bubbles drift in from the edge. The two remaining differences are how far the milk is heated, and whether it is stirred.

The home recipe heats milk in a pan to just short of boiling, stirring with a whisk. The coffee shop's recipe says only warmed milk, with no temperature. Under the same words, how it is heated and stirred changes the milk before it reaches the mouth.

Heating and foam are taken further in the piece on deciding the milk ratio first. After the pan comes what goes into a chain's cup.

Inside a chain's cup

Lift the lid on a chain's cup and a sweet, creamy smell comes up. A chain's hojicha latte is not built from tea leaf alone either.

One large coffee chain's standard item blends hojicha with black tea leaf. A white-chocolate-flavoured syrup and fresh cream are added. As the name says, it is hojicha and black tea in one cup.

Whether it is on the menu changes with the season. The figures and names are gathered just below. A little brown foam is left at the bottom of the paper cup.

What changes is the astringency

A soft, sweet steam rises from the cup once the milk is in. What roast depth changes most is not the aroma but the source of astringency.

In public trials, what fell as the roast deepened was the catechins. The compounds measured as roast aroma came out higher in the light roast than the dark. If a dark roast suits milk, the reason lies on the side of less astringency.

For choosing a dark roast in pursuit of aroma, current public measurements offer no support. After the answer come the sellers' words.

Roasted grain and nuts, in words

Pick up a packet in front of the shelf and a faint toasty smell escapes. In a merchant's account, the dark roast's aroma holds its own against milk.

One tea merchant recommends a dark roast because its roasted-grain, nutty aroma survives in milk. It is a seller's description, not a measured one. Even so, as a target when choosing, it is clearly put.

The sources go as far as this: that is how the seller describes it. After the words come the measured figures.

Roast deeper, less astringency

A sip leaves a faint astringency under the tongue. The deeper the roast, the lower the total of catechins, the source of that astringency.

As astringency leaves a ripening persimmon, total catechins fall as the roast deepens. In a prefectural tea station's figures, the dark roast held about a fifth of the unroasted level. Light, medium and dark were made by drum speed, not by temperature.

What falls, though, is the total. Taken type by type, some rise sharply at the light roast before falling. Roasting does not shave the astringency away evenly; it reshuffles it first, then reduces it. After astringency comes the toasty aroma.

More toastiness in the light roast

Lean over the cup and a toasty steam reaches the nose. The compounds measured as roast aroma are not greatest at the deepest roast.

In the same trial, the compounds used as a marker of toastiness are absent from unroasted tea and appear only with roasting. In first-flush tea and karibancha, the lighter the roast, the more there was. In second-flush tea, roast depth made no difference.

The sources support this direction only within that trial's conditions. Next comes another trial, one that moved the temperature.

Higher heat, the other way

Inside a roasting drum it is too hot to hold a hand near. In another trial that raised the temperature, the toasty compounds increased instead.

When a research institute in another prefecture roasted stems, every aroma compound singled out as important rose with the temperature. One trial moved the drum speed, the other the temperature. The conditions differ, so neither is simply wrong.

Deeper means toastier, and lighter means toastier, each hold only inside their own trial. After the trials comes which way a word on the shelf points.

The same word, facing away

Beside the tea, a bag of dark-roast coffee beans sits on the shelf. The same word, dark roast, points in opposite directions for coffee and for hojicha.

In a coffee body's account, roasting deeper moves coffee towards bitterness. Hojicha's dark roast, as just seen, moves towards less of what makes astringency. The same two characters sit on one shelf, facing away from each other.

The figures and trial conditions are gathered just below. The milk in the cup has quietly cooled a little.

The ratio comes first

Pour milk into a measuring jug and the white lines of the scale waver. What gets settled first is not the leaf or the roast depth, but the share of milk.

Decide what share of the cup is milk, then build a strength of tea to fit. In that order, each change can be followed one at a time.

Which share tastes best is not something the recipe numbers settle. After the ratio comes how it relates to strength.

Strength and ratio, separate dials

Leave tea standing in the pot and its colour slowly deepens. Strength of tea and share of milk are separate dials.

In the published recipes, as the piece comparing café and home cups showed, milk ranges from a splash to nearly the whole cup. A cup meant to come close to the café sits on the milky side, hojicha with a little milk on the other. Until the aim is settled, the ingredients cannot be.

Nor does brewing stronger necessarily bring the café closer. A bōcha shop's latte is more than twice the strength the same shop gives for everyday brewing, yet keeps the milk low. After strength comes the milk itself.

What may be called milk

Turn a carton from the fridge and small print runs down the side. If the type box says milk, it is raw milk only, with nothing adjusted.

According to a dairy industry body, only raw milk above set levels of milk fat and non-fat solids may be called milk. Low-fat milk, non-fat milk and milk drinks are each defined separately. For a day aimed at the café's impression, a carton whose type box says milk is the place to start.

One convenience store's hojicha latte is also described as finished with raw milk only. The sources go as far as the labelling rule and that description. After the carton comes heating in the pan.

Just before the skin forms

As milk warms in the small pan, a thin skin begins to form. Heating can be thought of with that skin as the boundary.

Like yuba slowly thickening, the skin on milk thickens with the time and heat. A dairy association explains that it starts to form above forty degrees, and that the first skin is mostly fat. A home recipe's whisking up to just short of boiling is also a way of keeping the heat from gathering in one place.

A fork will do for a whisk, a small saucepan for a milk pan. Dedicated equipment matters only for matching the café's speed. Warmed milk tasting sweeter is not a matter of more sugar. Sweetness other than fructose is said to be felt more strongly near body temperature.

The sources go as far as how temperature, skin and sweetness relate. Last comes plant-based milk.

Oat milk for baristas

At the end of the shelf stand cartons of oat milk. Those marked for baristas are described as having been tested for how readily they foam.

The criteria given are ease of foaming, fineness of foam and stability. The description says nothing about how it behaves with tea.

How it suits tea comes up again under blending. The figures and labelling rules are gathered just below. The milk in the small pan settles, leaving small bubbles at the rim.

Hold the tea, milk and sweetness

Set two cups of the same shape side by side and their white rims line up. What has to be held steady folds into three: the tea side, the milk side and the sweetness.

The tea side is weight of leaf, amount of water, water temperature and waiting time. The milk side is amount and temperature. On a day for comparing roast depth, everything except the kind of tea stays as it was.

Listing what to hold steady does not, by itself, make a comparison. Next comes the leaf being swapped.

More than depth changes

Open the two packets: thin stems in one, leaf in the other. Swapping the same amount does not compare roast depth alone.

In the shop's own product list, the original sits under roasted stem tea and the replacement under roasted leaf tea. Roast depth is not the only thing that changes. The part of the plant and the blending differ too; these are two different products.

The sources go as far as what the product list shows. After the products come the words on the packet.

Light and dark, a shop's words

Across the front of the packet, dark roast is printed large. Light roast and dark roast are each shop's own words.

Much as a river's name changes with the land it runs through. The labelling rules settle names such as hojicha, and roast-depth names are not among them. No reference temperature for roast depth appears in the documents consulted either.

Buy packets from different shops, and partway through it stops being clear what is being compared. After the words comes keeping a record.

Make it twice, write both down

A ruled table goes into the kitchen notebook. Make it twice under the same conditions, and record both.

One run cannot tell a difference between teas from the variation of the day. On a day when even one thing cannot be held steady, it is more useful later to give up the comparison and call it a cup for the record.

There are no records yet; what can be written here is the design of the comparison. Results will go first to the association's members as records, as they come in, and be added to this piece afterwards. The figures and product names are gathered just below. The table in the notebook is still blank.