Where the value sits
Green tea is unusual among teas in that the cheapest options are often genuinely good. A plain, machine-made, unnamed green from a decent province is a real tea with a real flavour, and it costs a fraction of a famous one.
The value sits with leaf picked slightly later, from a large producing area, made by machine. Loose pan-fired greens from Anhui and Sichuan, Japanese sencha sold in 100 g packs, and gunpowder-style rolled greens all sit in this band and all drink well.
Shape is where the price hides. Flat, needle or spiral shapes have to be made by hand or by a machine set up carefully, and both cost more than simply rolling the leaf. A broken or loosely rolled leaf brews the same cup but sells for much less.
The spread is wide enough to be worth knowing before you shop. A plain pan-fired green sells for 20 to 40 yuan per 100 g at the producing end, while a famous-name first flush from a core zone runs to several hundred for the same weight.
Region without the name is the other cheap seam. Leaf grown next to a famous area, processed the same way, sells without the premium. You give up the protected designation and most of the resale story, and you keep most of what is in the cup.
What is not worth the premium
The clearest premium to refuse is the pre-Qingming label on a cheap tea. Genuine early pick costs more at the farm gate, and a bag sold at everyday prices carrying that label is either late leaf or a smaller share of early leaf than the name implies.
Gift packaging is the second. A tin, a box and a ribbon add cost without touching the leaf, and the tea inside a gift set is frequently chosen for how it looks rather than how it drinks. Buy the leaf loose and put it in a tin you already own.
So-called wild or ancient-tree green tea is the third, and it is the one that costs people the most. The wild and old-bush green category is small, most of what is sold under that description is not it, and the premium is large enough to be worth doubting on principle.
Finally, be careful with famous names attached to low prices. A Longjing from outside the 3 recognised zones is a perfectly decent flat green and is sold as one by honest sellers. It is not a bargain West Lake Longjing, and no price makes it one.
Spend on freshness
If you have a fixed amount to spend, put it towards freshness before grade. A fresh second-grade green from this spring will beat a top-grade lot that has sat since last year, and the gap is not close.
That means buying smaller amounts more often. Two 100 g packs bought 6 months apart cost roughly the same as one large pack and give you fresh leaf twice, which is the difference between a tea you finish and a tea you abandon.
It also means checking how the seller stores it. Leaf held at 25 °C fades noticeably faster than leaf held at 4 °C, and refrigeration costs the shop money every week. A seller who pays that bill is spending on the thing that actually matters.
Spend the rest of the budget on one thing you cannot get cheaply: a tea whose origin you trust. One good Longjing from a seller who can name the area is worth more to your understanding of green tea than a cupboard of anonymous leaf.
A simple budget rule
A rule that survives contact with real prices: split the money three ways. Most of it goes to an everyday tea you brew without thinking. A smaller share goes to one seasonal purchase, ideally the first flush of something you already like.
Keep the smallest share for experiments. Buy 50 g of something unfamiliar, brew it properly, and write down whether you liked it. Failed experiments are cheap at that size, and they are the only way your taste actually develops.
Buy twice a year rather than once. Green tea is at its best within about 6 months of processing, and the main harvest lands in early spring, so a spring purchase and an autumn top-up keep you inside that window without cold storage.
And do not chase a grade you cannot taste. Most people drinking at home cannot reliably separate grade one from grade two in a blind cup, while anyone can tell fresh from stale. The budget should follow the difference you can actually detect.