The other night, I took a close look at King Digital’s F-1 IPO filings. It made for some interesting reading.

But as I finished the post, I noticed something interesting. King seemed to have a much lower cost structure than Zynga. King is now much larger than Zynga in terms of users and revenues, but on a per user basis the two generated the same amount of revenue, roughly $1.50 per user per month. However, in terms of profitability, King makes a lot more, about $0.66 per user per month to Zynga’s $0.02. (I think I missed a zero in the previous post.) And in terms of cash, King generates $1.67 per user per month, to Zynga’s $0.26. 

It strikes me as very odd that the two companies employ a very similar business model, as demonstrated by the revenue per user figure, have such widely divergent profitability levels. I speculated that there were a few possible explanations. It could be some weird accounting difference, but the cashflow figures make that, at best, a partial explanation. It  also seems unlikely that the ‘cost of goods’ is different for the two companies, they are both selling virtual goods. I looked through the financial statements of both companies, and saw some interesting differences. Both companies have similar cost of revenue, about 30% for each. However, in 2013 Zynga spent 72% of revenue on R&D while King only spent 6%. So even though King had more than double the revenue of Zynga in 2013, they spent only $110 million R&D, while Zynga spent over $640 million. Now I suspect that there are some accounting differences here (King amortizes some of its development costs, Zynga does not appear to), but this is still a wildly different figure. 

Then it occurred to me that maybe there is something structural that divides the two companies. Zynga’s hit products were built on technology that is now ten years old.  They have gone through a lot of disruption and restructuring since then, so it is unlikely that they have made any major changes to their software stack. By contrast, even though King has been around longer, its hit games were all built very recently and probably use the latest technology. 

 So I dug a little deeper, and found that the two companies built their infrastructure in very different ways. Zynga was one of the first companies to get really big built on Amazon’s AWS cloud computing platform For several years, they were a poster child of AWS. In recent years, they have begun building their own data centers. I believe they now use AWS to start a new game, and then once they determine the actual load of users transition to their own servers, occasionally using AWS again for unexpected spikes in demand (aka Cloud Bursting). By contrast, KING claims to use its own servers entirely. The numbers bear this out too, in 2013 Zynga only spent about $8 million on capex, while King spent about $16 million.

I think this points to something important – the cost of computing has fallen so dramatically that companies using the latest software stacks may have a real cost advantage. This resonated with me because of a conversation I had a year ago with the CEO of another social gaming start-up. When I first met him five years ago, he was almost an evangelist for AWS, and talked a lot about how important AWS was to his business. But by last year, he had changed his tune considerably. At that time, he told me that if he were to start again today he would avoid AWS entirely. He had recently switched to using his own servers, and this had involved hiring no new employees. According to him, the tools have gotten so much better that it has gotten far easier to stand up your own servers. I mentioned this annecdote a few months back when I looked at just how profitable AWS probably is. AWS may be so profitable that it is pricing itself out of the market. [Disclosure: I own 50 shares of AMZN stock.]

I have long held that the advances in computing and software in recent years have the potential to shift the competitive landscape in many industries. Companies that build cheaper IT systems will gain a real world advantage over their peers. When analysts look at companies, we tend to pick our favorites. “This is a great industry, but Company A is just better run than its competitors.” It is often hard to pin down why one company can just seem so much better managed than others in the space. There is a tendency to just assume the C-Level executives somehow run a better ship. I think that advances in technology actually play a major role in this. The retailer that has the best software for getting customers in the door, or the casino with the best algorithms for picking bad gamblers will be able to steadily inch ahead of their competitors. I know for a fact that the CFOs of some major, recent Internet IPOs are watching this trend closely and trying to understand how some of their peers can get away with spending so much less in capex. It should come as no surprise that Internet companies are the first to experience this, but I think this trend will spread. This is how technology becomes a competitive advantage.

So I am not certain that this is the reason why King appears so much more successful than Zynga. There are probably a number of factors involved. But I also believe that having the best software stack is going to be an important differentiator for companies.