Amazon Supply has been making some waves in the laboratory products market since they got into what they refer to as “Lab & Scientific Products”. A lot of manufacturers have asked us what we think about their entry into the market and we have generally responded anecdotally that it’s most likely an extension of their current business model: Sell a huge variety of products, inexpensively, with easy ordering and fast shipping. The assumption is that Amazon, with all its efficiencies, would be able to offer lower prices than could its competitors. We said that enough that we started to wonder if it’s actually true.
To settle this once and for all, we did a little mini-study. We compared the stated online cost of 10 products sold by Amazon Supply vs. 5 of the major US distributors: Fisher, VWR, Daigger, Cole-Parmer and Thomas Scientific. We only considered products where the exact same product from the same brand was offered by Amazon Supply and at least four of the other companies. Included was plasticware (3 products), glassware (1 product) and equipment (6 products). Reagents were not included because Amazon Supply is weak in that area and carries mostly commoditized chemicals and buffers which are difficult to brand match across 6 companies. We admit, there is no good way of selecting products in a manner that is both random and practical, so we simply searched for popular items from common brands that we believed most large, general-purpose life science distributors would carry. It actually worked quite well.
A few notes before we get to the findings… The costs analyzed are the US costs. Prices in other countries will vary, and of course every country will have its own unique distributors. If you’re outside North America, you may very well only know 2 of the distributors we used as a comparison. We also tried to remove any influence caused by differences in pricing given to each distributor by specific manufacturers by having as little overlap in manufacturers as possible. In fact, the only manufacturer of more than one product used in our mini-study was Corning, who manufactures two of the products sampled.
We took all the prices for all 10 products, normalized the prices for each product, then took the average of the normalized prices for each distributor. This gave us one number – if our study is accurate (which it very well may not be since the sample size is quite small) this number will represent how much more or less expensive any given distributor is. A value of 1.050 would indicate the distributor is 5% more expensive than the average of these 6 distributors. Likewise, a value of 0.900 would indicate a distributor is 10% cheaper.
So, these are the averages of the normalized prices for our basket of 10 products:
- Amazon Supply: 0.896
- Fisher Scientific: 1.052
- VWR: 1.035
- Daigger: 1.077
- Cole-Parmer: 1.003
- Thomas Scientific: 0.950
Turns out that we very well may be correct – Amazon does seem to be competing on price. Their prices for these 10 products were, on average, over 10% lower than the average competitor. (For all you statistics nerds, the 2-tailed, 2-sample unequal variance t-test score on the difference in Amazon Supply’s prices was 0.033.) What was at least equally as interesting to us is that for every product – 10 out of 10 – Amazon Supply’s prices were lower than the average. In our sample population, the closest they got was a normalized price of 0.984 on an IKA orbital shaker. They also advertise free 2-day shipping on orders of $50 or more, which is just about everything, so taking that into consideration Amazon would be even more price competitive.
Something else that I found noteworthy was that there were only two companies that carried all 10 products (aside from Amazon Supply, which did by definition due to our study design) – VWR and Thomas Scientific. Fisher and Daigger each carried 9 of 10, Cole-Parmer carried 7 of 10. Again, this could very easily be an anomaly due to the limited sample size, and we didn’t bother to do any statistics, but I thought it was interesting nonetheless. If two makes a coincidence and three is a pattern, ten might even be called data, albeit not a whole lot of it.
We figured some people might be interested in the data, so we put it online here. It’s in excel format so you can play with it if you’d like. If you get motivated and add to it or do additional analysis, let us know! E-mail me at carlton.hoyt@[you know the rest].