Selling Smart – the Art of Price Variations

In traditional brick and mortar shops, variations in price are a long established practice to identify the so called ‘sweet spot’ for the best price at which a product can be sold to a consumer. Supermarkets are known to be experts at this practice, where everything from the store layout to where a product is placed on a shelf is thought out, tested and evaluated to ensure the highest profits. They can do this because of the high volume of consumers using their stores, and because of the high footfall can get a good, statistically significant sample on a daily basis to test out their hypotheses.

Moving Price Variations to the Online World

Moving this analogy to the online world, it will not be surprising to learn that this has been used to much effect in the world of ecommerce marketplaces, where once again the high virtual footfall of consumers is providing massive amounts of data to merchants and marketplaces on how to sell, where to sell, and most importantly at what price to sell at. The marketplace model has also dramatically reduced the cost of this experimentation, so now it’s easier than ever to vary pricing across the catalogue, with very little expenditure. Obviously some marketplaces such as Amazon (where if a listing exists you will attach your own product) or FlipKart are not appropriate for this sort of pricing activity, however eBay is perfect, and in fact, at a point in 2009 for when data was obtained, it was found that there were over 240,000 pricing experiments from a total of 7.5 million listings.

What does a variational listing look like?

One example of a variational price listing is from a single seller, ‘budgetgolfer’, who listed the same item 31 times under different prices and sales methods; 11 of the listings were for a fixed price of $124.99, while the others were auctions scheduled to end within a week. The seller had also varied shipping fees between $7.99 and $9.99 for each of the listings. As eBay’s search algorithm will typically spread the seller’s listings over multiple pages, then it is unlikely the listings will compete with each other, and so this can be a good strategy to increase visitor numbers.

Are auctions any good?

In the same research, it was also found that typical auction final bid prices gave around a 13% discount to a customer as opposed to a ‘buy it now’ button, and only in around 5% of cases did the value of the bid go above the actual price of a product (that was listed on eBay elsewhere with a ‘Buy it Now’ button). eBay is now moving toward a more ‘buy it now’ centric model, and currently around 60% of listings on eBay are buy it now (although I still receive emails to see if I need to sell any gadgets after clearing my loft out).

How does shipping affect the overall price?

Interestingly, it was also observed that the increase in shipping fee by $1 was compensated by a decrease in auction price of $0.82, so it’s also suggested that the shipping fee is not entirely ‘internalized’ by the customer (e.g. the consumer doesn’t note it as a $1 price increase on the total cost), although this was seen to be different across different categories of product.

What can I do now?

As for action points from this, set out your own hypothesis that you want to test. For example; “A decrease in shipping fee will increase my overall sales”. Try varying your product and shipping prices on multiple listings that you make at the same point in time. Note that there can be differences between product categories (e.g. DVDs or jewellery). Make a record of these experiments, and analyse your sales for the time period they are running (note the number of sales that occur for each listing). Run them for a number of weeks (as long as the sale price doesn’t cause you to lose money), and see where they turn up in search results. If you can, run these across multiple marketplaces, and you might be surprised by the differences; you’ll get a very different customer on eBay to one on ShopClues or Amazon for example. At the end of the time period, see whether your hypothesis has been proven or disproven, and act upon it.

For further reading, delve into http://www.stanford.edu/~leinav/Seller_Experiments.pdf, which was used as a reference for this article.