While the explanation of how regression analysis works takes a few pages, the reasons for using a regression analysis take just two sentences:
- Learn which factors are most influential on diamonds’ prices.
- Using a sample of diamond inventory from several vendors, develop pricing formulas, which predict the price of any particular diamond sold by such vendors.
- Using such formulas compare predicted prices of similar diamonds across various retailers. Such formulas also help identify which retailer has consistently lower prices compared to others.
So a regression analysis can tell you if the price of what you are buying is inline with competitors’ prices and which factors influence the price more than the others. Some retailers set their prices based on specific, quantifiable diamond attributes, such as the 4Cs (cut, color, clarity, and carat weight). Others pick prices based on what they think people will pay. So, not all pricing is logical, and a regression analysis can help steer you away from a jeweler who picks prices on a whim, as well as a retailer who follows a logical pricing pattern but charges 20-50% more than an equally reputable store.
If you are comfortable collecting data and running a regression analysis, that’s great, but not everyone is.
As an example of what regression analysis can reveal, we analyzed 6,000 diamonds from one reputable retailer (name withheld as we are just trying to prove the worth of regression analysis—not advertise for or against this particular company). The resulting formula looks like this:
Diamond Price = Intercept + (Carat x $4,020) – (Clarity x $122) – (Color x $90) – (Cut x $25)
To use this formula to predict this particular retailer’s diamond prices, just plug in parameters of a particular diamond. For example, price a diamond that is:
- Shape: Round (brilliant)
- Carat weight: 0.60 ct.
- Clarity: VS2
- Color: H
- Cut: Ideal
So, you would plug in $230 for Intercept (based on the actual regression analysis for this particular retailer), 0.6 for carat, 5 for clarity (see Regression Tables below, where 5 is equivalent to VS2 clarity grade), also 5 for color (color grade H is also equivalent to 5 in the Regression Tables), and finally, 1 for Ideal cut. Our formula for this retailer will then look as follows:
Diamond Price = $230 + (0.6 x $4,020) – (5 x $122) – (5 x $90) – (1 x $25) = $1,557
| Color | Grade |
|---|---|
| D | 1 |
| E | 2 |
| F | 3 |
| G | 4 |
| H | 5 |
| I | 6 |
| J | 7 |
| K | 8 |
| L | 9 |
| M | 10 |
| N | 11 |
| O | 12 |
| P | 13 |
| Q | 14 |
| R | 15 |
| Clarity | Grade |
|---|---|
| FL | n/a |
| IF | 1 |
| VVS1 | 2 |
| VVS2 | 3 |
| VS1 | 4 |
| VS2 | 5 |
| SI1 | 6 |
| SI2 | 7 |
| I1 | 8 |
| I2 | 9 |
| I3 | 10 |
| Cut | Grade |
|---|---|
| Ideal | 1 |
| Very Good | 2 |
| Good | 3 |
| Fair | 4 |
| Poor | 5 |
So, using the formula, the predicted total price is $1,557. The actual retailer’s price for this diamond is $1,516. Thus, in this case, the diamond is $41 (3%) less than the formula’s predicted price. With the assistance of regression analysis, you’ve learned this particular diamond’s price is inline with this retailer’s pricing plan.
By doing the math, you have also learned the key factors that weigh on the retailer’s pricing decision. Carat is by far this retailer’s largest concern. Carat accounts for more than $4,000 per carat or $2,400 for the 0.6-carat diamond used in our example.
The results of the regression analysis show the remaining Cs—clarity, color, and cut—are significantly less important to this retailer’s pricing. Each change in cut varies the price by $25; each color grade accounts for $90; and each clarity grade changes the price by $122. So, if you were debating downgrading the color or clarity by one grade, but were not certain which to drop, you’d see that a step down in clarity saves you $32 more than a step down in color (saving of $122 versus $90). Of course, price should not be your only consideration!
With the knowledge gained from this one regression analysis, and after you did you own homework you can walk into that retailer’s store and compare the “predicted” price for the vendor you have analyzed (assuming this retailer has one of the lowest prices in the market) against the price tags of the brick-and-mortar retailer you are considering. Comparing the actual prices with “predicted” will help you see if any diamonds are “good deals” compared to the low-price retailer.
You can also construct a similar regression analysis formula for other retailers, based on their inventory, to compare predicted diamond prices from store to store. All you need is a calculator, more information on how to compute a regression analysis, and time.