Skip to content
Imperial Star Gems

Buying guidesEngagement9 min read

Choosing an engagement ring stone

An engagement stone is the one diamond most people buy without ever having bought one before, and it is worn every day for decades. That combination — no experience, high stakes, hard wear — is what this guide is built around.

Start with the shape

Shape is the only decision here that is purely a matter of taste, and the only one that cannot be quietly improved later. Grades can be traded against each other; a pear is a pear. So settle it first, and settle it honestly — on what the wearer likes, not on what is currently good value.

Round brilliants are roughly two thirds of the market. They are also the only shape a laboratory issues an overall cut grade for, which makes them the easiest shape to buy well at a distance. Everything else is a fancy shape, and buying one well means reading the proportions yourself or looking at the stone.

Elongated shapes — oval, pear, marquise, emerald — face up larger than a round of the same weight, because more of their weight sits across the face. That is a real saving, and it comes with a real catch: elongated brilliant cuts are prone to a bow-tie, a dark band across the centre that no report mentions.

Round
The safest buy and the most expensive per carat. Cut-graded, so you can judge it from the paperwork alone.
Oval, pear, marquise
More size for the weight, and more taste required. Ask for a video and check the centre for a bow-tie before anything else.
Emerald, Asscher
Step cuts. Open and architectural, and unforgiving — buy them a grade or two higher in both clarity and colour.
Cushion, radiant, princess
Brilliant-faceted squares and rectangles. Forgiving on clarity, and the ratio changes their character considerably.

Where the budget actually shows

There is a spending order that holds for nearly every engagement stone, and it is not the order the four Cs are usually listed in. Cut first, because it decides whether the stone is bright or dull and nothing else can compensate for it. Then clarity — but only as far as eye-clean. Then colour, matched to the metal. Carat last, because it is the C that costs most and shows least per pound spent.

The single most common mistake is paying for grades that only exist under a loupe. Clarity is defined at 10× magnification, so a VVS1 and a well-chosen VS2 can be indistinguishable to anyone who is not holding a loupe — and the price difference between them will comfortably buy a better cut, or another quarter of a carat.

The second most common is buying exactly on a magic size. Demand clusters at 1.00, 1.50 and 2.00 ct, so the price per carat steps up as a stone crosses them. A 0.92 ct is about 2.5% narrower than a 1.00 ct — a difference of roughly a fifth of a millimetre — and meaningfully cheaper.

1. Cut
Excellent or Ideal on a round. On a fancy shape, sensible depth and table figures, tight symmetry, and eyes on the stone.
2. Clarity, to eye-clean
VS2 and SI1 are very often eye-clean. Stop at the first grade where the stone is clean to the eye, and spend the rest elsewhere.
3. Colour, to the metal
G–I in white metal, H–J in yellow or rose. Above that you are paying for a difference that only shows face down against a master stone.
4. Carat
Buy just under a magic size. Judge size in millimetres, not carats — the cut decides how much of the weight you can see.

Match the colour grade to the setting

Colour grading happens face down, in a neutral tray, under standardised light, against master stones. It is a laboratory comparison designed to be repeatable — it is not a description of how the diamond looks on a hand.

Once a stone is set, the metal around it drives the apparent colour more than a grade or two does. A warm setting lends its colour to the stone, so in yellow or rose gold an H, I or J reads as white and a D is money spent on something nobody will see. In platinum or white gold the contrast is less forgiving, and G to I is the sensible range.

Fluorescence is the related lever, and the one most buyers avoid for no reason. In an I or J, faint to medium blue fluorescence can make a stone look slightly whiter in daylight, and such stones usually sell a little cheaper. It is worth asking for specifically — just judge the stone in daylight rather than judging the word on the report.

The durability questions no grade answers

A grading report describes a loose stone. It says nothing about how that stone will survive twenty years of doors, handles, gym equipment and washing-up — and an engagement ring takes more knocks than any other piece of jewellery a person owns.

Three things are worth asking about before you commit, and none of them appear as a grade. First, girdle thickness: Extremely Thin is fragile at the setting bench and afterwards, and Very Thick parks weight in a rim you cannot see. Medium to Slightly Thick is the comfortable range.

Second, on any pointed shape — pear, marquise, heart — ask specifically about the girdle at the points, and have the setter protect them. Third, ask whether any feather reaches the surface, and where it sits. A large surface-reaching feather under a claw is a question about durability, not about clarity, and the grade will not flag it.

Girdle thickness
Medium to Slightly Thick. Avoid Extremely Thin, particularly on pointed shapes.
Surface-reaching feathers
Ask where they sit. Under a claw or at a point is a different conversation from tucked at the girdle.
Knots and cavities
Both break the surface. A knot can catch a fingernail; a cavity traps dirt and dulls the stone between cleanings.
Culet
Should read None on a modern stone. Medium or larger can show through the table as a grey dot.

Natural or lab-grown

They are the same material. A lab-grown diamond is diamond by every physical and optical test, it is graded on the same scales by the same laboratories, and it is inscribed as laboratory-grown on the girdle. What differs is origin, scarcity and price.

The practical consequence is straightforward: the same budget buys a substantially larger or higher-graded stone lab-grown. If size is what the wearer wants, that is the honest route to it. If the origin of the stone is part of what the piece means, that is a reason to buy natural, and it is a perfectly good one.

What we would not do is choose between them on resale. Neither is bought back at anything close to retail, and lab-grown prices in particular have fallen a long way as growing capacity has expanded. Buy the stone you want to look at.

A specification that works

A sensible engagement stone for daily wear, written as ranges rather than single grades. Adjust one line at a time, and spend what you save further up the list.

Shape
The wearer's, without compromise
The one decision that cannot be traded against another.
Cut
Excellent or Ideal (round)
On a fancy shape: depth and table in range, symmetry Excellent or Very Good, and eyes on the stone.
Clarity
VS2 – SI1, eye-clean
Step cuts one to two grades higher — an emerald hides nothing.
Colour
G – I in white metal, H – J in yellow
The setting metal does more for apparent colour than a grade does.
Carat
Just under a magic size
0.92 rather than 1.00, 1.42 rather than 1.50. Judge the size in millimetres.
Fluorescence
None to Medium
Faint to medium blue can help an I or J, and usually costs less. See the stone in daylight.
Girdle
Medium to Slightly Thick
A durability figure, not an optical one. Check the points on pointed shapes.
Report
GIA or IGI, inscription checked
Compare GIA with GIA and IGI with IGI — not across the two.

Terms used in this guide

The short version. Each one is defined in full, with what a report prints for it, in the glossary.

Cut grade
An overall verdict on how well a round brilliant has been cut — issued for rounds only.
Eye-clean
No inclusion visible to an unaided eye at normal viewing distance.
Fluorescence
A glow, almost always blue, that some diamonds give off under ultraviolet light.
Girdle thickness
How thick the band at the stone's widest point is, graded Extremely Thin to Extremely Thick.
Magic sizes
The round weights — 0.50, 0.75, 1.00, 1.50, 2.00 ct — where price steps up sharply.
Bow-tie
A dark band across the centre of an elongated stone, shaped like a bow-tie.
Step cut
A faceting style of long parallel facets running around the stone like steps.
Lab-grown
A real diamond grown in a reactor rather than in the earth — chemically and optically identical to a mined one.

Frequently Asked Questions

What buyers ask us most often about this particular purchase.

How much should an engagement ring cost?

There is no rule, and the ones you have heard — two months' salary and its variants — began as advertising. Set a figure you are comfortable with, then use the spending order in this guide to get the most stone for it. A well-cut 0.90 ct at G/VS2 will outshine a poorly cut 1.20 ct at D/VVS1 every day of the week.

Is a lab-grown diamond a real diamond?

Yes. It is the same crystal, the same hardness and the same optical behaviour, graded by the same laboratories on the same scales. The report states it is laboratory-grown and the girdle carries a matching inscription. The difference is how it was formed, what it costs, and how scarce it is.

What clarity grade is safe to buy without seeing the stone?

For a brilliant cut, VS1 and VS2 are almost always eye-clean, and SI1 usually is. For a step cut such as emerald or Asscher, move up: VS1 or better. But 'usually' is not a guarantee — clarity is graded at 10×, eye-clean is not graded at all, so ask us to confirm it on the specific stone.

Does the colour grade matter less in yellow gold?

Considerably less. A warm setting lends its colour to the stone, so H, I and J read as white in yellow or rose gold while a D would show no visible benefit. In platinum and white gold the contrast is sharper, and G to I is the range we would work in.

Should I buy the stone and the setting together?

Buy the stone loose. It is the only way to examine it from every angle, weigh it, and check the inscription against the report before a setter touches it. Choosing the setting afterwards also means the mount is built around the actual measurements rather than a nominal weight.