Diamond dealers vs retailers: bottom line up front

Short answer: for a serious natural diamond, go to a manufacturer, a cutting house or a dealer.

  • Cheaper. A retailer buys its stones at exactly that level and then applies keystone, roughly doubling the price, with a branded showroom at about three times. Between one trade level and the next the steps are thin, commonly 3 to 5 percent, so the whole mark-up sits in one hop and you can buy before it.
  • Better. The specification a careful buyer describes when asked is about five times more common one step up the chain: 52 percent of trade-access stock against 9 percent of retail-held stock.
  • Better treated, which is what this page measures. Across 216 classifiable South African businesses, the trade group carries a mean Google rating of 4.77 against 4.53 for retail, and 2 percent of trade businesses sit below 4.0 against 13 percent of retail.
  • They do the rest too. Setting, bespoke commissions, resizing, re-tipping, insurance valuations, repairs and buy-back, usually on their own bench. A counter without a workshop sends that work out to a bench like theirs.

Retail earns the sale on three things: the piece exists today, you can slide a wall of finished rings onto an actual hand, and some people want the brand. A 4.53 mean is a good score and nothing below argues otherwise.

Now the finding in full. South African businesses that Google labels as trade rate better than the ones it labels retail, and the gap is not small. Across those 216 classifiable businesses, sitting at 4.5 or above: 84 percent of trade, 69 percent of retail. Sitting below 4.0: 2 percent of trade, 13 percent of retail. In the written reviews, one and two star reviews make up 2.5 percent of the trade pile and 8.2 percent of the retail pile. What the study measures is satisfaction and the language buyers use, and the limits of that are set out in full further down.

What I measured, and the one method point that matters

I pulled businesses from Google Maps across Johannesburg, Sandton, Pretoria, Cape Town and Durban on 3 August 2026. 400 businesses were captured, 312 were in the jewellery or diamond trade, and 216 had at least five reviews and were classifiable. From 179 of those businesses I collected 6,482 reviews, of which 4,993 contained written text. Reviews were sorted newest first rather than by rating, so what follows is representative recent experience rather than a complaint-first sample.

Here is the part that decides whether you should believe any of it. The trade versus retail split uses Google’s own category labels, not mine. I did not sit and decide who counts as a proper dealer, because that judgement is exactly where a study like this normally goes wrong. The groups are:

GroupGoogle’s own category labelsBusinesses
TradeDiamond dealer, Jewellery manufacturer, Wholesale jeweller44
RetailJewellery store, Jeweller, Jewellery designer172

That is the whole classifier. No opinion in it.

The results

MeasureTradeRetail
Mean Google rating4.774.53
Rated 4.5 or above84%69%
Rated below 4.02%13%
1 to 2 star share of written reviews2.5%8.2%
Price mentioned positively95.6% (n=68)80.5% (n=215)
Trust or honesty mentioned positively92.9% (n=42)82.4% (n=102)
Stone quality discussed17.3%12.6%
Delay or turnaround complaints0.8%1.5%

Sample: 216 classifiable South African jewellery and diamond businesses, 6,482 reviews from 179 of them, 4,993 with written text, collected 3 August 2026. Price and trust percentages are the share of reviews that raised that subject at all, with n shown.

Price and trust are where the two groups actually separate

The star ratings tell you there is a gap. The written reviews tell you what the gap is made of.

When price came up at all in a trade review, it was raised positively 95.6 percent of the time (n=68). In retail reviews it was raised positively 80.5 percent of the time (n=215). Notice the sample sizes as well as the percentages: price is a far more common talking point in retail reviews, and a far less settled one. Trust and honesty run the same way, 92.9 percent positive for trade (n=42) against 82.4 percent for retail (n=102).

Stone quality gets discussed in 17.3 percent of trade reviews against 12.6 percent of retail ones, which fits a straightforward reading: dealer customers arrive already thinking in colour, clarity and cut, so that is what they write about afterwards. Delay and turnaround complaints are low everywhere, 0.8 percent for trade and 1.5 percent for retail, so the idea that going direct means waiting longer does not show up here at all.

By exact Google category

Splitting the same businesses by their precise label complicates the picture, and the complication is worth more than the headline. Every category with at least three businesses is below, including the ones that cut against the argument this page has been making.

Google categoryBusinessesMean rating
Jewellery buyer124.89
Jewellery designer (retail side)224.87
Jewellery manufacturer (trade side)164.79
Diamond dealer (trade side)254.76
Wholesale jeweller (trade side)34.73
Gold dealer144.69
Jeweller (retail side)274.56
Jewellery store (retail side)1234.47

The trade labels do not come top. Jewellery designers, who are firmly on the retail side, rate 4.87, above all three trade categories, and jewellery buyers rate higher still. If this page showed only the five labels that suit its argument, the trade would appear to sit cleanly above retail with no overlap. That is not what the data says, and presenting it that way would be dishonest.

What the data does say is narrower and, I think, more useful. The largest retail category by a distance, jewellery store at 123 businesses, is also the lowest rated of the lot at 4.47. That is the category most buyers actually walk into. So the finding is not that the trade beats everybody. It is that the high-volume retail experience is the weak part of this market, while small specialist operations do well on both sides of the trade counter.

Two warnings attached. Wholesale jeweller contains only 3 businesses, far too few to rely on alone. And jewellery designer, at 22 businesses, is a small self-selected group of people doing bespoke work, which is a different job from running a shop.

What this study does not prove

These limits are not fine print. They are the reason the rest of it is worth reading.

  1. It measures satisfaction and sentiment, not rand prices. This study is not where the price argument comes from. That rests on the trade’s own keystone convention, which is published and needs no dataset. What the reviews show is that buyers who dealt at trade level were markedly happier about what they paid.
  2. There is a selection effect, and it probably favours dealers. Dealers serve fewer, better-informed, self-selected buyers who often arrive with a spec and a budget already set. Retail absorbs walk-in traffic and gift buying, which is a harder audience to satisfy. Some of this gap is who walks through the door, not how well they were served.
  3. Wholesale jeweller had only 3 businesses. Too few to lean on alone.
  4. Google categories are often self-assigned by the business. A label is a claim. Treat it as the start of your checking, not the end.
  5. No business is named and no review is reproduced verbatim. This is aggregate data only, published as aggregates deliberately.

What I take from this

If you are buying a serious natural diamond, the evidence points toward buying at dealer or manufacturer level. Every measure in this study runs the same way, the star gap is backed by the language gap underneath it, and the strongest single finding is that price complaints all but vanish in the trade group.

Two other pieces of research on this site arrive at the same door without measuring satisfaction at all. On stock: of 230 GIA-certified natural diamonds listed by six South African sellers in June 2026, 52 percent of the 93 reachable through trade access were D to G in colour and VS2 or better in clarity, against 9 percent of the 137 held in South African retail. On price: the July 2026 listing work, 73 rows across 18 sellers of which 17 are South African, put a finished one carat ring in a premium showroom at R105,000 to R428,000, a band that includes the setting and rests on three listings of three different specifications. Set that beside the published list rather than beside another observed band. Rapaport’s round list of 20 March 2026 prices G/VS1 at 1.00 to 1.49 carats at $5,400 per carat, which at a dealer discount off list, R16.50 to the dollar and 15 percent VAT puts a loose one carat at roughly R72,000 to R92,000. One of those is a finished ring and the other is a bare stone, so it is not a like-for-like subtraction. The distance between them still needs no villain to explain it. The trade’s own convention is keystone, roughly double the trade cost at a retail counter and roughly three times at a branded showroom, against margins of commonly 3 to 5 percent between one trade level and the next. The doubling happens when a stone leaves the trade for a counter, and it pays for rent, stock, staff, insurance and a brand.

Three angles agreeing is worth more than any one of them alone, and it is worth saying plainly that not one of the three is a price comparison between named sellers. The price research sets out how those figures were collected.

Retail is still a legitimate choice, and I am not going to pretend otherwise. A 4.53 mean is a good score, and 69 percent of those businesses sit at 4.5 or above. A showroom you can walk into on a Saturday, with staff who talk you through a finished piece and handle sizing and after-care, is a real service many buyers genuinely want. Buy it with your eyes open. Just do not assume the shop floor is also where the best stone at the best number is, because nothing in this data supports that.

The label on the listing is a genuine signal, it is free to read before you drive anywhere, and it is one of the few signals in this market that is not written by a marketing department.

Checks you run yourself

None of these need a shop name, or me.

One thing first, because it is where most people lose an afternoon. The category label is a good signal to read, but “diamond dealer” is a poor thing to search. That phrase and its cousins draw close to no search volume in South Africa, perhaps thirty a month between them, and the handful of results they surface are mostly not the businesses this study is about. Search the price question instead: 1 carat diamond price, diamond price per carat South Africa, diamond price South Africa. That is where the sellers who publish real per-carat numbers turn up, and it is where the demand actually sits, roughly 28,410 searches a month against 30. Then, on whichever listing you land, run the checks below.

  1. Read the category label on the listing itself. It sits under the business name. Diamond dealer, Jewellery manufacturer and Wholesale jeweller are different animals from Jewellery store.
  2. Sort the reviews newest first, not by rating. The default view is not the honest view. Recent experience is what you are buying into.
  3. Ignore any listing with fewer than five reviews. That was my threshold for including a business at all, and it is a sensible one for a buyer.
  4. Count what the reviews are actually about. Reviews praising price, honesty and the stone itself are worth more to you than reviews praising the decor. Do it for two or three sellers and the difference shows quickly.
  5. Ask whether the seller holds the stone right now, or sources it in after you pay. Ask it in exactly those words, then use the diamond buying checklist for the rest of the appointment.
  6. Get the GIA report number before payment and verify it yourself at gia.edu/report-check, then match the girdle inscription to the report.
  7. Ask for SADPMR registration and any trade body membership, then check it rather than accepting the claim.
  8. Get the loose stone price and the setting price as separate line items, VAT included, in writing, from at least two sellers.

Method, and how to challenge it

Data collected 3 August 2026 from public Google Maps listings and public Google reviews, grouped by Google’s own category labels. Sentiment counts are shares of the reviews that raised each subject, with n shown against every one. No business is named, no review is reproduced, and nothing here is a recommendation of any particular seller.

If you think a figure is wrong, say so. Corrections go to [email protected] and the process is set out in the editorial policy.

See also