Ebay at the crossroads (part 2)
In many ways, what ebay has achieved in the business model that it has pioneered is astounding. It has built an iconic brand that has changed our purchasing behaviour, and it has done this on a foundation of many people that it has never met. In other words, it has built trust on the back of millions of people who have bought and sold through it – a minority of whom are not trustworthy. It has done all of this as an intermediary and has relied on general honesty, to which it added a feedback system. This is an awesome achievement.
Ebay is also in a very rare position in the business world – it continues to dominate the market that it largely created, and has no significant competitors. The result of this is that it is very passive in the way that it drives sales. It relies on people coming to the site to search for items to buy. This is backed up to an extent by marketing – it used to spend a lot more on paid search and its television adverts have been visible but have not been successful (according to Marketing Week). In this way, it is missing the greatest potential of having an online trading platform – especially two things: recommendation algorithms and viral exposure.
This post will look at recommendation algorithms, the next will look at viral exposure. These algorithms drive a greater volume of sales per customer. In The Long Tail, Chris Anderson argues that there is a long tail of demand in many things. In other words, if you chart the sales volume of available variants of a product in descending order of popularity, there will always be a few products at the head that account for most of the sales, and less demand for the more niche products further down the tail. The aggregate demand for all of these niche products is significant (and frequently with less competition and better margins), but a traditional retail model cannot meet this demand easily due to the cost of shelf-space and inventory, so choice becomes limited. For online retail, the dynamic is very different - fewer physical constraints mean that it is easier to meet very niche demand.
Everyone has seen Amazon’s solution to this problem – the recommendation algorithms. These appear in a number of different circumstances and use the combination of excellent product data, what you are looking at (and have looked at / bought), what products other people who viewed that product also viewed / bought, and amateur reviews and recommendations of those products. This creates the ability to treat people as niches of one and give them highly relevant products (people like you have also bought…).
Ebay could do something very similar. Sellers already categorize their items in order to get a much better sale (something, by the way, that they could do a lot better if they had the free-text and suggestions technology that Facebook and others have), so it wouldn’t take a lot to categorize items far more effectively. Imagine that – people would be prompted to buy far more things if the recommendations worked much, much better. Now I know that people might not be very happy with Ebay suggesting rival auctions at the bottom of the items that they are trying to sell, but it could go at the end of the bidding screen (assuming that the auction is not about to end in the next 30 minutes), or even better, after bidding has closed it could be displayed to the losing bidders. The point is that this would cost very little for them to develop, and even less to run.
Imagine the power of the Long Tail of Stuff (as Chris Anderson called Ebay) proactively recommending relevant stuff to you – it is enough to make you wonder why they have not already done this.