Ask HN: Critique my startup: ZeroTier One
https://www.zerotier.com/
-- Status:
Tech: Emulates an Ethernet switch over a robust P2P protocol.
Product: VPN-like, make virtual networks everywhere.
Revenue: freemium
Growth: ~1% organic (no promotion), ~50% user retention after 1-2 months (measured).
Feedback: positive, >90% "very disappointed" for canonical "if this product disappeared?" question.
-- Issues:
* 1% organic growth tells me: I haven't quite achieved product/market fit, the "beta" status is putting people off, or product isn't quite "there."
* Freemium conversion rate is also low, tells me current value proposition is broken.
* No mobile version. Already know how to build it, but time consuming...
-- Hypothesis:
Virtual networking ("VPN" has become confusing due to HideMyAss, etc.) is a "sleeper" market not unlike file sync. "There's lots of things that can do this." "But do you like any of them?"
I've studied Dropbox closely. I'm trying to do that in virtual networking-- solve an old problem in a compelling new way.
Some validation e.g.: #1 source of organic hits is variations of "alternative to Hamachi" directly or via alternativeto.net.
Two parter then... (1) is this hypothesis correct?, and (2) what would push this from "good" to "category redefining?"
I have some bigger "pivot" ideas but I wanted to get feedback on what's there first.
All thoughts are welcome. Thanks!
12 comments
[ 2.8 ms ] story [ 37.1 ms ] thread500-600 users online at any given time, about 1000 total global users.
User base is highly global, with lots in China, etc. There's a lot of people using this to collaborate globally.
Paying customers are all consulting outfits with distributed teams. I can see this being compelling there. But there's only a couple of them.
Current development iteration is focusing on better OS integration, a bit of other polish elements, and fixing some very minor networking issues reported by a few users. On Windows networks now show up (in the dev branch) as "real" networks, and it shows the popup allowing you to select their role and firewall designation as home, work, or public. On Mac I'm working on making networks show up in System Preferences as real networks there as well.
Nothing comes immediately to my mind for why I'd want that same capability in 2014. Which is not to say there isn't an important use case. Rather: even people who have done stuff like this before probably need to be told how this service makes their life better.
Here's what some current users are doing:
* "Office LAN with no office" for file, printer, and other devices sharing and collaborative development. One comment was "cool, now we can use git as an actual peer to peer version control system."
* Allowing a distributed team to access a license server that must be online to use some commercial app.
* With bridging to extend an office into VPN-space (regular VPN replacement).
* A few users have used it to create a private backplane network between servers in the cloud. I wrote up a blog post on this:
https://www.zerotier.com/blog/?p=176
* One user is researching it as a way to link together small embedded devices with cellular connections in the field. I've been working with them on this. It builds and runs fine on Raspberry Pi and works over cellular Internet.
* It's got a bit of a following as a way to tunnel out from behind the Chinese Great Firewall, though I'm not sure if that's a sustainable or monetizable area. But I do have one paying customer from this market. It's a commercial user with an office inside China and an office outside that finds it very convenient to have a single LAN spanning their China and Oregon offices.
* The public network feature has a bit of a following in the decentralized networking community. This is not likely to be a direct revenue source but has resulted in a lot of high quality technical feedback.
https://www.zerotier.com/
Even my work internet connection is behind a carrier-grade-NAT box so I can't even port forward. I use it to access our Jenkins server which is physically located inside the building.
It's particularly useful for 2 reasons:
- each machine gets a very stable IP (unlike DHCP-assigned addresses) which simplifies hosts files / dnsmasq etc - it's better than port forwarding as every device can work on port 80
So my use case would be something like "simple LAN setup with no more port forwarding"
Of the converted users so far, the vast majority of them are distributed teams. It's a peer to peer virtual network, so it allows them to have an "office LAN without an office." There's a wide variety of reasons people might want that.
Nevertheless, I am getting the impression that that's a niche market, so I'm thinking about other directions I might go with this. The underlying engine is very powerful and versatile, so there are other markets I could potentially apply it to: SDN, p2p SDKs for software developers, etc. I'm starting a market research deep dive on this kind of thing.
I can't imagine that there are no larger niches for a networking engine that enables friction-free direct connectivity between any two devices, but I'm not completely sure what these niches are. I've got some hypotheses but I want to let the market research lead.
Thanks for your comments!
There are a lot of other areas you can apply this technology, but there are already tons of players in those spaces.
One user has a CAD app that must talk to its license server to work. With this, they can have a license server and their distributed team can use it. (Consulting company.)
There's a few users who just like the convenience of being able to do things like share screens, drop files to each others' machines, etc. as if they're all in the same room. There are other ways to do it, but they find this easy and straightforward and the services are already built into the OS.
A few users have security concerns that make them more comfortable with direct file transfer over their own encrypted LAN than with Dropbox, et al. These are medical and financial sector users.
But ultimately I think you are at least partly correct. Most things have moved into "the cloud," so a significant proportion of users have no pressing need for direct connectivity. This makes it likely to be a niche market.
Of course you can go further and ask "why has everything moved into the cloud?" There are multiple reasons, but one reason is certainly that direct connectivity is hard. But now you have a chicken or egg problem: nobody builds apps that require direct connectivity because it's hard, and there's little market for making direct connectivity easier because nobody builds apps for it. Those kinds of catch-22 scenarios are murderously hard to escape, either technically or marketing-wise.
One application that doesn't fit the mold though is private network backplanes for cloud servers that span data centers and cloud providers. I've got a couple users who have servers on multiple providers on multiple continents (Amazon, Digital Ocean, etc.) and this lets them have a single shared encrypted private backplane with a few minutes' setup time and no manual configuration.
I've considered following that line of thinking and building some kind of "meta-cloud" solution, also possibly integrating with Docker, OpenStack, etc. The idea would be "SDN over WAN and between providers." I've gotten a few positive comments on that concept, but I'm at the early stages of exploring it.