It’s Valentines Day today, and it made me think back to the pricing model we built here at Anderson Analytics two years ago today. To get a background on the model and why I decided to build it you can check out Anderson Analytics’ 02/09/06 press release “A Diamond is Forever – But at What Price.”
It looks like we have had 3,649 unique visitors to the diamond pricing calculator in just the past year alone. We put up a public tracker on the site sometime in 2005, so you can check out the statistics yourself if you like. I always enjoy looking at the GEO tracking for some reason, you can see that since we started tracking about 73% of our visitors are from the US, interestingly Singapore is #2 at 6% (though this may be due to someone’s post about it on a more popular blog in Singapore, not the popularity of diamonds in Singapore. Still we know Diamond culture is growing fast in East and Southeast Asia. I’m sure there will be several last minute visitors today due to Valentines Day.
When building a model such as this you need to learn a bit more about the market and the product, it isn’t sufficient to look solely at the data, which in this case were all key variables and priced on well over 40,000 diamonds screen-scraped from the internet plus a selection of non internet diamonds. To get this additional knowledge in order to build the model I visited with experts in the New York Diamond District. I also had discussions on websites such as pricescope.com where diamond merchants discuss pricing.
It was apparent that these diamond experts wanted no part of an accurate pricing tool, at least not one that is available to the public. There is no doubt in my mind that DeBeers marketing campaign over the past several years has been the most successful campaign in history of Marketing. They have successfully driven something which should rightfully be a commodity into something extremely emotional. “Can you put a price on love”, “Each Diamond is as unique as a snowflake” — BS!
Diamonds are in fact easier to price than cars, and for cars we have Edmunds.com and KellyBlueBook.com. Before building the Anderson Analytics Diamond pricing model I took a careful look at the most widely used pricing tool in the market now, the Rapaport Wholesale Diamond Report. Apparently Martin Rapaport who invented this model also got a lot of heat for it before everyone started using it, you can read more about him on Wikipedia. But it was apparent to me when looking at the Rapaport that we could build something much more powerful using advanced data mining and statistical software.
The first version of the model we put out was actually much more accurate than the model we have up now. The first version predicted one single fair market value price for a diamond given all the variables from the 4 C’s to certification, depth, table, symmetry etc. This was actually the fair market price of online retail diamonds of course since this is where most of our data came from. So these prices were lower than what you would find in your local brick and mortar jewelry store.
After receiving a lot of hate email from jewelers we decided to change the model slightly to make it less accurate. Now instead of giving you an exact price it lets you compare two diamonds and tells you which is a better overall value.
Regardless of what jewelers have said about it (and not all response has been bad, some have asked to receive the code and/or put it on their sites), I’m very confident about the accuracy of the model. We have also had it reviewed externally by a group of MIT graduate students who have expertise in modeling, and they were impressed with the logic and accuracy.
There’s nothing wring with emotive marketing. But come on! Don’t give us that crap about each diamond being a unique snowflake. Whatever happened to selling quality product at a fair price?
Happy Valentines Day!
0 responses so far ↓
There are no comments yet...Kick things off by filling out the form below.
Leave a Comment