Monday, 27 May 2013

Amazon IT capex - spending like it's gone out of style

Another interesting figure in the Amazon 2012 10K filing: capitalisation of IT expenditure:

"Capital expenditures included $381 million, $256 million, and $176 million for internal-use software and website development during 2012, 2011, and 2010."

There are a couple of points of interest in these numbers. Firstly, how small they are! There's an old Forrester benchmark for retail IT expedenditure which roughly suggests that, for a "traditional" retailer, 1% of revenue is spent running the show, and a further 1% enhancing it. And then they suggest that for a digital business, this latter number could be anything up to 8%.

By contrast Amazon's capital expenditure as a % of revenue looks like this:

Although it sounds odd to observe that $381M of IT capex is a small sum, when viewed against overall revenues it really is. You might actually argue that Amazon is underinvesting in IT... Given that it also makes a loss and has a growth strategy, this looks quite odd. What might they spend any extra on?

Well improving the online experience for fashion would be a good start. One of the "features" of the Amazon site is the extent to which it is one-size-fits-all. Essentially it still sells books, and if the product isn't a book, the user experience can be quite peculiar. (Take a look at the utterly bizarre online grocery beta in the UK or Germany for an example of really really peculiar).

Fashion is evidently something they want to push. Although my home page is doubtless heavily personalised, the fact that fashion is even worth personalised banner space is indicative:

And yet the site doesn't really tackle the whole issue of variants - sizes and styling - at all well. It's trying to bend a books-based system into clothing. Try this part of the UI/UX for a shoe product (I'm currently working for Dr Martens so shoes are front-of-mind right now!):

What the hell does that price mean? And what's with the size selector? A clear case of - to be honest - not spending enough on the IT to adapt it for footwear. Take a look at how Zalando handles this same challenge (much better, cutting a long story short).

Anyway, the other point of interest in the IT spend is that it is increasing, both as a % of revenue (from 0.5 to 0.6) and massively in absolute terms. And given that it doesn't seem to be on the "products" side of the business - or at least not on selling shoes - then it seems reasonable to guess that it might just be on services.

This would be reasonably consistent with another interesting note in the 10K - a comment on gross margins:

"Gross margins increased in 2012... primarily due to service sales increasing as a percentage of total sales." 

Gross margins increased from 22.4% to 24.8%, a hefty hike. Services sales increased to $9.4Bn from $6.1Bn in 2011 and $3.4Bn in 2010. Suddenly the IT spending numbers start to make a little more sense. If that increase in IT spending has been primarily to support increased services sales, then an extra $200M of IT spend to support an extra $6Bn of services sales - 3.3% - looks much more like the kind of benchmark numbers suggest by Forrester.

But then it makes the "other" IT spending look even more woefully inadequate...