How do I compare diamond quality from different online jewelers?

Last month a client named Daniel forwarded me six listings for near-identical 1.5‑carat round diamonds. Every stone was graded F/VS2, every one claimed “excellent cut.” Prices ran from $8,200 to $14,500. He wanted to know which one was the right one. The honest answer: only two of them were worth buying. The rest had something wrong that the video didn't show. Comparing diamond quality across online sellers isn't about the sparkle reel. It’s about reading the lab report like a jeweler - and knowing exactly which numbers matter and which pictures are lying.

Start with the lab report, not the photo

Pull the GIA or IGI report number - not the PDF, the actual number - and look it up on the lab’s website. If a seller won’t give you the report number before you swipe a credit card, close the tab. I don’t care how good the video looks. No report number means no verifiable third-party grading, and that’s a dealbreaker. For natural diamonds I want a GIA report. IGI is the standard for lab-grown and that’s fine, but I read IGI grades a half-step softer on color and clarity, and I price them accordingly. If you’re comparing one stone with a GIA report against another with an EGL USA report, you’re comparing two different grading scales. Don’t do it. Stick to one lab for the comparison.

Cut grade is the easy part - the proportions are the hard part

A GIA “Triple Excellent” or an IGI “Ideal” is a starting point, not a guarantee. The cut grade is a range, and that range contains diamonds with very different light return. For a round brilliant, pull the crown angle, pavilion angle, and table percentage. What I look for: crown angle between 34.0° and 35.0°, pavilion angle between 40.6° and 41.0°, and a table between 54% and 57%. A 34° crown with a 40.8° pavilion? That stone will return light like a hard drive. A 35.5° crown with a 40.2° pavilion? You’ll get dark patches under the table, even if the report says “Excellent.” Most online tools only show you the cut grade. A handful of specialty sites show you the full proportion set, and those are the ones I send clients to first.

Depth percentage matters too, but it’s a proxy. I like a total depth under 62.5% in a round. Any deeper and the stone starts to face up small for its carat weight. You’ll see a 1.00‑carat that looks like a 0.90 because it carries all its weight in the pavilion. That’s a bad deal and you can spot it in three seconds with the depth number.

For fancy shapes, the report won’t save you

Ovals, cushions, pears, marquises - GIA doesn’t assign a cut grade for those, and the numbers alone can lie. A 1.20‑carat oval with a 62% depth and a 58% table can still be a dead stone if it has a black-hole bowtie that spans the whole face. I once had a client, Nicole, order an oval from a big e‑commerce player. The 360‑degree video was shot with a single bright point source that washed the bowtie out completely. When she opened the box, the center of the stone was dark grey in daylight. We returned it inside 24 hours. The replacement had a 1.28 length‑to‑width ratio and the seller sent an ASET image - that made all the difference. With fancy shapes, you need an ASET or Ideal‑Scope image, or at minimum a video shot in diffuse, shaded daylight with the diamond on a hand. If all you have is that spinning jewelry-box video under a dozen LED spots, assume the worst and ask for more.

What the video hides

Online jewelers light their stones to sell, not to grade. A diamond can look top‑color under a video ring light and then show a faint brown or grey tint in mixed office light. This is especially true in the near‑colorless range - an H or I from a slightly loose lab can face up a full grade lower in person. Fluorescence is the other invisible variable. Medium blue fluorescence can make an I‑color diamond look like an F in sunlight, which some clients love. Strong fluorescence, though, can make a stone look milky or hazy. Reports mention fluorescence but not the transparency penalty, and no video I’ve ever seen online captures haze well. If a diamond has strong blue fluorescence, I want to see it with my own eyes before I set it. If you’re buying sight unseen, I’d cap it at medium.

Check the clarity plot on the report against the video. A VS2 with a single inclusion under the table in a critical spot can scatter light at certain angles; you won’t see it in a straight‑on video. A VS2 with a feather that breaks the pavilion surface is a durability risk, especially in an ultrasonic cleaner. The report will map the inclusion, but no listing will point out that a surface‑breaking feather near the girdle can become a crack. That’s the kind of thing a jeweler spots in six seconds under a loupe.

The return policy is your real inspection window

Every diamond I’ve ever bought online for a client came with at least a 10‑day return window, and I use every one of those days. A 30‑day policy is better. I tell clients: order the two strongest candidates, ship them to me, and I’ll put them under the scope, weigh them, measure them, and compare them side‑by‑side on a hand in daylight. More than half the time, we send both back and wait for something stronger. That’s normal. The sellers expect it. If a jeweler charges a restocking fee on a diamond, I’d look elsewhere - plenty of reputable vendors don’t.

If you’re comparing diamonds from three different sites and you’re stuck, email me the report numbers and the prices. I’ll tell you what I’d set and what I’d skip. That’s not a sales pitch. It’s exactly what I do for clients sitting across the bench - and sometimes the right answer is “none of the above, let’s keep hunting.”

The stone you end up with should make you stop worrying about the numbers. That’s when you know the search is over.