Saul,
Once again, thanks for a great write-up! It’s obvious that you don’t do these monthly reviews simply for the benefit of those of us who read your posts and try to emulate your style of investing. I’m confident that they assist you in sharpening your own thoughts about your investments. I’d be curious to know if you did something similar prior to starting this board. I don’t recall that you specifically recommended monthly reviews in the Knowledgebase (it’s been a while since I reread it), but if not, you might consider adding this practice. It’s something I have not really engaged in, but I think I will start, if only for my own benefit.
Separately, I’ll add just a few words about some the companies we hold in common where I think I can add some understanding about what they do and what gives them a moat:
Twilio. They provide communication services and they seem to have no viable competition in what they do besides do-it-yourself.
I’ll just shed a little more light on that subject. Much was said about Uber deciding to to roll their own. But not much was said about what might have motivated them. Communications and telephony reside at the very heart of the Uber business model. In fact, IMO ride hailing is actually an outgrowth of what they do, obviously, an important one and I’m hard pressed to come up with a better one, but it is the communications functions that enable the service. In that Twilio has a transaction based fee structure and Uber is a communications transactional enterprise it undoubtedly made economic sense for them to self provide in order to escape the fees.
But this is far from the norm for most companies. Most companies, are engaged in some other business and rely on Twilio enabled functionality primarily to provide customer support or sales functions. For those companies digging deep into the weeds of communications protocols (and these are very deep weeds) would require a hiring a handful of specialists who would inevitably become an IT bottleneck as their workload would be sporadic. The tendency would be to under-staff these high paid positions.
Enter Twilio with a library of APIs that most programmers can master after an initial training period. And there’s no annual subscription fee for the libraries, you only pay for what you use when you use it. What’s not to like? Virtually every company big enough to have an IT department will need some communications functionality at times. Twilio makes it easy and standard. No risk of two specialists coming up with different solutions for the same problem.
How about competition? I already discussed the special case of Uber. And yes, a competitor could develop similar functionality, it’s not that esoteric. But if you’re hiring IT staff why would you go with brand X? Twilio has the network effect of being first, best and getting better with an ever growing body of programmers already familiar with their APIs and how to use them. And their pricing model might be matched, but can’t really be beat, a new entrant would have to have a lot of funding to undercut their prices and survive for the long haul while playing catch up. If you’re a venture capitalist would you bet on brand X?
And just a brief word about Nutanix. Not all that many years ago many companies went through a major IT overhaul by replacing big blue mainframes with smaller primarily UNIX boxes. This IT transformation was called “distributed computing.” I was one of several managers on such a project where I worked. Not well recognized at the time was that one set of problems related to centralized mainframe computing was being exchanged for another set of problems related to multiple application and data servers. In a nutshell, as the number of applications and the volume of data escalated, the number of servers required multiplied as well. What used to be the I/O bottleneck (and vendor lockin) of mainframes was traded for a nightmare of networking, capacity allocation and load balancing among myriad boxes. Enter Nutanix. In a nutshell, they vastly simplify the management of all those boxes, interconnections and capacity requirements. The management job is not eliminated, but with Nutanix it is no longer necessary to pay attention to every server and storage unit as an individual device. All the devices and traffic between them is hidden from view and the management is raised up to a macro level. And it’s not just the management, but forecasting and capacity planning are made much simpler as well. These important functions become more accurate and gain expanded time frame windows.
New Relic - APM/NPM are not new. I managed the TPNS group where I worked in the 80s (TPNS stands for Teleprocessing Network Simulator, an IBM product for the IMS DB/TP environment). Oracle has offered database performance monitoring tools for a number of years. But, the problem should be fairly obvious from what I’ve just related. The monitoring tools that have been available are all domain specific. What sets New Relic apart is that it monitors at higher level of abstraction. And as Saul noted they are in the process of adding NPM tools. I don’t have much to add other than watch this space. I don’t know much about it other than what’s been discussed here, but to some extent Elastic plays in the same field. That may pose a serious competitive challenge - or it may not. I don’t know enough about it yet to make a determination. Maybe there’s some folks here who can shed more light on it than I.