So one of the things that's interesting here is the Amazon component. And the anonymity of Amazon vs a smaller provider. Let's assume for a minute the founders actually wanted to keep it alive (I'm guessing they didn't but that's a whole other story) You're generating revenue of say $40,000 and your Amazon costs are $30K. If you went to a small or mid tier sized provider and said "Good news/bad news. We're generating $40,000 a month and growing (oddly revenue isn't which is again another story) You're costing us $30,000 a month. If we can cap you at $20,000 a month for 12 months that frees us up $20,000 and we can keep 3 employees on in the short term and still grow the business. We can get through the rough patch and continue to grow and in a year from now we can adjust this upward and be worth more. If you insist on $30K, which is your right, we shut the whole thing down because we can't keep the lights on.
Now the provider is either looking at generating $240,000 for the next 12 months or losing everything from your account. He/she is probably doing the deal with you because much of their costs are fixed. But I don't know if/how you have that conversation with Amazon.
> But I don't know if/how you have that conversation with Amazon.
I do. I've had a conversation like that with Amazon. What I found is the Amazon hosting team is excellent: they moved mountains for my account to get us to success.
The Amazon account exec arranged immediate discounts and bridge services, even working through the holidays, then brought in multiple engineers for phone calls with me to teach me how to optimize our project to cost less to run. Every one of the Amazon people who helped me was 100% on board with finding ways to make the project successful.
That conversation is always worth having with vendors in my experience. They may be able to come up with ideas, or offer rebates, or promotions, or even convertible note financing toward the next fundraising round.
Can vouch for this as well. So far Amazon customer service has been great for us as well. We just got introduced to our account manager a couple of months ago but she has been very helpful in getting us connected with the right people and helping us save money on our monthly usage.
I can vouch for this with Rackspace as well. Our account is nowhere near $30k per month, but we do pay the $100/month managed service fee. I've had multiple conversations in regards to new services and infrastructure changes that might save us money.
It's always good to approach any of your business-critical vendors and service providers to discuss pricing. You might find that they are more than willing to negotiate.
I loved Everpix and wanted you guys to succeed. But when I finally heard the bad news, I thought Amazon should buy Everpix and make their cloud drive super awesome with it (Right now cloud drive is nearly terrible).
Everpix was a great product and I still can't believe there is no one who sees value in it.
[Everpix founder here] Interesting idea and no we didn't try negotiating with Amazon. Not sure how much room there is here?
However it's important to keep in mind in a situation like this, and at this scale of burn rate, you're not looking for a solution to shave off a few thousands $ left and right, which is kicking the can down the road (if even), but for a long term solution.
The key problem here is that I don't think you can realistically switch a business built on the VC funding track to the bootstrap track without the collateral damage of killing your company's essence i.e. the product.
For instance: when suggesting to have fewer employees, that is implying you have room for that or that your employees are interchangeable, which is far from the truth. It's not like we had 5 tech support people, and we could have downscaled to 3. We had hired quite slowly and were already as lean as it gets i.e. 1 person per project: iOS, Web, Infrastructure, Science, Design, Imaging / QA. All these folks were working at full capacity managing a multi-platform service managing hundreds of millions of full-res photos with ~100 servers, displaying photo collections of 10,000+ photos in very responsive apps, and running cutting edge computer vision science at a scale no other consumer photo startup was doing (the closest was the semantic search in Google+ Photos). You can certainly cut salaries (which we did and only works up to a certain point in the very competitive SV hiring market), but you can't remove positions which would be akin to extending a bridge to nowhere.
Bootstrapping it with the existing team would cost about 1.5 million a year by my extremely rough estimate. So you'd have to make $30 per user per year. Facebook's ARPU is about $6 per year (much higher in the US though).
If you make as much in ad revenue as Facebook, each of your 7000 paying customers would have to foot a yearly bill of $171 to keep the service going (but that number would go down as you grow).
So I think you're right. Bootstrapping it at this point is probably unrealistic. You may need like 10 times the number of users for it to be a viable business.
Thanks for sharing. This is a truly inspirational insight. As a novice startup wannabe, I have this idea that whatever imperfection I start off with, I can just tidy up when and if I get huge. But from a look at Everpix, I can see this is probably a fatal approach. You've got to be scalable from the start. Do you agree?
Yep totally understand. I meant this more as a general exercise perhaps more applicable to bootstrappers or even seed funded folks who had a single large cost that wasn't currently scaling.
When I first read their story, I wondered exactly what you wrote: how much was their amazon hosting bill and could they have at least kept the lights on if they went with a smaller provider?
I know when I was doing my music start up it would cost is many many times more to be on amazon than the two high power dedicated servers.
I don't know the story in their case but I find it shocking how many high bandwidth start ups have never even explored the option of not using AWS. I mostly blame this problem to the easy availability of funding which encourages you to throw money at problems. The issue though is that the more you build out your infrastructure around AWS exclusively, the harder it will be to explore other options later.
I think some people only see one side of the trade-offs here.
A counter example would be a start-up I worked with that started out on a smaller provider. When we launched a beta of the application, we unexpectedly got invitations for national TV coverage with less than 24 hours notice. We asked the provider for everything they could give us, but they only had a limited amount of unused capacity themselves so there was little they could do. Of course we couldn't come near to soaking up the traffic spike, and lost a huge potential number of positive user experiences.
That sort of thing doesn't happen to you on Amazon or the other largest on demand providers.
That's why you start with Amazon. And then once you reach scale you slowly start building your own private cloud. Its much more efficient and costs a lot less. We have been doing the same in our startup.
Something else I think people fail to consider is what I would call a hybrid approach.
For example, even if you mostly operate on your own infrastructure, keep a replica of your data tier going on EC2, along with freshly configured images for the rest of your moving pieces. This way you have an instant pop up B site for disaster scenarios.
There are a lot of other possible concepts. Keeping a dataset at amazon to use for map reduce jobs is convenient. Or even if you want to keep your infrastructure virtual, being able to price arbitrage across a couple providers could be a powerful cost reduction. The price of entry for all this is devops automation, but I think people are now mostly convinced that's worth it.
I think it's the best way to go, but also for a resilience approach, I want to be able to ditch a datacenter in 15min. If you can ditch the datacenter, why not ditch the whole provider?
So we tried to keep the deployment scripts ready for a competitor at all times, having backups we could restore there etc. And if you grow big enough, this becomes a bargaining chip with your provider too: I can ditch you in 15 min.
Or simply take those $30k and get some colocation and hardware - get the monthly operational costs down to sub-$10k easily.
It strikes me every time how much money people sink down on Amazon(AWS) when they can do it themselves for a lot less. The usual counter-argument is "but we didn't have money for a sysop", well yeah - your devs/devops needs to manage the AWS system as well, it's not much harder to manage real hardware / servers and network.
Another post that striked me was $300k in consultants, thats more than 3 years of full-time employees - what did they do? Was it worth it?
People keep saying this, and I can see how it's easily true if we're talking about EC2 or one of Amazon's database services. But what's the hardware equivalent of S3 (the majority of the costs in this case) that's cheaper and "not much harder to manage" than S3? I feel like once you want to get past a couple TB of (redundant!) storage in colo/dedicated you're back to the land of sysops, no? I mean, I am decently comfortable in my ability to keep a server running, but I've never maintained a SAN (is that what I would need to replace S3?) and wouldn't trust myself to provide the level of resilience that Amazon is able to provide.
GlusterFS on Supermicro Storage servers and loads of raid-grade harddrives, install once - scale forever. And it's not very hard to learn one additional software, and everyone knows that you need to setup a simple backup solution or system.
What I'm saying is that if you pay a cold $30k to AWS it is probably time to look at other solutions - or your COGS will keep growing out of hand.
I understand the greatness of AWS and Cloud systems when starting up, especially to keep CapEx down, but only to a certain point, you still have to manage systems (servers), security and storage/mount-points etc, so why not take the time to learn how to do it yourself and save a lot of money? Why not get an experienced DevOps instead of pure Developer?
And this company was already paying $300k to consultants, why not use a consultant or outside company to setup/manage their own hardware and service contracts as well? It would still be less than $30k and would still scale a lot better.
Have you tried running the equivalent of S3 using GlusterFS? Because I've seen mixed results with Gluster in practice and am not sure I'd feel comfortable going that route. Also, which one of their staff would have had to learn how to scale that? How many mistakes would they have made along the way and how much would that have cost them?
I think you're not understanding that you DONT have to eventually manage all of the systems on your own.
S3 requires effectively 0 management. Dynamo requires effectively 0 management. The list goes on, and grows every quarter.
Sure, if you're just running EC2 nodes, there's some sysop requirement there. But if you run those on ElasticBeanstalk instead, you don't even have to manage that!
It's very hard for server and hardware-focused folks to 'get' the dramatic difference of running with AWS or GCE or other large cloud ecosystems. It's trivial for a, say, Java developer, to have a large, scalable system and have 0 access to each try to manage the actual machines and software that everything is running on.
You CANNOT replicate that in a private environment today, for anything less than a large infrastructure management cost.
What a hugely useful thing for you guys to do. Thank you, truly.
The lesson I see from this: AWS hurts bad if you are using it heavily, but salaries... salaries kill. Automate, outsource and self-build as much and as long as you can possibly stand it, and then push on a lot longer than that.
>The lesson I see from this: AWS hurts bad if you are using it heavily, but salaries... salaries kill. Automate, outsource and self-build as much and as long as you can possibly stand it, and then push on a lot longer than that.
Going off on a tangent here, but if (a) "software is eating everything" and (b) "salaries kill", wouldn't that mean a society of mass (30%+) unemployment in the future?
I wonder who would have the money to pay for stuff like Everpix or any other offering at that point...
My theory is that when everything is automated, so will food production also be, assuming there is enough energy in the world for that. Then everyone will need very little money and Everpix can charge much lesser amount of money...thus still keeping number of users the same.
Ideally, but who says things will be automated in that order?
The crazy thing is that when people are suffering, there is work to be done. If the 30%+ unemployed in that scenario could reorganize themselves into servicing each others needs, they would surely want to do that.
I also worry about centralization (of both government and corporate power) which get at the primary political problem to be dealt with in my opinion: how to mitigate the sociopathic element.
My best answer, at this point, is to break large states up into lots of small states, spreading the sociopathic ruling classes broadly and making it difficult for corporate interests to completely dominate a given political unit.
Within a given smaller and sovereign political unit, I hope that decisions will be made that encourage full, useful employment (since there is less room for both social services shenanigans as well as corporate indifference in a compressed society.)
I'm currently someplace where we are treating AWS as a VM provider and it's freaking nuts how much it costs. (We might be looking at the problem wrong or failing to appreciate the fact that we can always scale with enough money.)
If you have a fixed workload (i.e. X servers always online) then take a look at reserved instances. The break even point is about 6 months for the light instances and a bit more for the medium ones. I don't recommend the heavy ones as the cost structure is different (you pay regardless of whether tt's on so ... if you later change your instance types you still have to pay hourly for the full term of the reserve).
Even better is if you can plan your infrastructure on AWS around using spot instances. They can be really cheap (we're talking a 5x times cheaper then on-demand and 3-4x than reserved). If your instances are used in a stateless fashion with all persistent state saved externally (DB, S3, etc) then you can do some pretty cheap scaling with spot instances.
One setup I've played with is a core set of non-spot instances phalanxed by a number of spot instances (at a couple different price points) for stateless web traffic. As long as the spot price stays below your bid you have significantly more instances available (which should give your users better response times). When the spot price rises your spot instances die and things slow down, but your app would still be alive thanks to the non spot instances. Again it takes quite a bit more engineering to get a setup like this but this is the kind of thing you need to do to take advantage of elastic computing.
At Everpix, we were already extensively using spot instances for the vast majority of our EC2 needs (which were pretty intense considering the type of computation and image science we were doing). That did save quite a bit of money when we did the switch and it was easier to do than we expected, so strongly recommended!
Did you guys look at bare metal at all? (either dedicated or colocated)? Spot gets you in the same ballpark of price on a "unit", which can be mentally deceiving since an AWS unit is pretty weak.
In our experience with CPU or IO intensive workloads, Spots are still 2-5x slower than what you can get at decent dedicated provider..
No, there was no time nor did we have the in-team experience. EC2 is one thing, which you can get down to pretty cheap if leveraging spot instances, but most of our infrastructure costs was on storage.
We had already very efficient image optimization for full-resolution photos using 4X to 5X less space than competitors. Without that our AWS bill would have been insane. So as far as it goes for pure storage on S3, we were as low as possible.
However, on the database side of things, there was a lot of room for optimization (and we only got to some of it).
You could move some of the pics to your own server, leaving AWS to handle growth in storage only. The raw costs could be very significant, though most likely you would need another person to create that system make it fast, which in SF probably means losing all the gains at that scale.
There are _a lot_ of dedicated options. In terms of price, Softlayer's at the high end. If Softlayer gives you a price/performace improvement over AWS of 2.5x, you can get another 2-3x that by picking a cheap (but not necessarily bad!) provider.
On the cheap end, you're looking at OVH and Hetzner. They've both been reliable for me, but I still wouldn't run anything mission critical on them (from anecdotes). Hetzner is Europe only, while OVH does have a DC in Quebec. If you're already using spot instances, which can vanish pretty quickly, you might have the code in place to handle these potentially flaky providers.
Over the years, I've had good experience with WebNX which has good connectivity to Asia. They've been supposed to open a NY location for a while now...Speaking of NY, there's reliablesite.net which is ok. You have Hivelocity in Florida (which I didn't use specifically for South America, but I imagine they'd be good for that).
There's a handful of companies which I've never used but that I've been hearing good things about for years. NetDepot, SingleHop and LimeLight Networks for example.
I have used 100TB in both Singapore and Washington. They are a Softlayer reseller. The extra level of indirection sucks when something goes bad, but they are much cheaper (what you can get from 100TB for $400 might cost $1200 at Softlayer). A few of the locations they actually have staff at, so I imagine that's much better.
I recommend you spend some time on webhostingtalk.com. It's great. There's a forum specifically for deals which can be useful to pick something up.
For anyone who's only familiar with AWS/clouds: most of these take 1-2 days to set up machines, their control panels are horrible, they don't have an API and you might not even get a dedicated KVM switch.
Wow, I feel like a huge advertising billboard this morning!
Hi.
Do you know why the above "cheap" services do not user something like OpenStack to bundle their services?
This would supposedly provide them at least with some dashboards and API's.
thanks.
Many of these are older than OpenStack by a wide margin. OVH in particular is huge, they own many data centers, thousands of employees...you just don't make that sort of switch overnight. The cost and risk might not be justifiable.
I did some spreadsheet-ing to look at the "profitability" of their userbase. If the business were sustainable, then Monthly Recurring Revenue [1] would be higher than the cost to service the customers [2]. If not, then growing the userbase just increases total monthly losses (assuming a steady freemium conversion rate).
In Everpix's case, it appears they were growing unsustainably.
Their data shows that all the way up until February 2013, they were spending $3 on AWS for every $1 of MRR. That's scary, because it means that a new customer bringing $1 MRR was actually increasing their monthly losses due to usage costs!
Things got better though. In March and April 2013, a surge in premium users seems to have shifted the balance in the right direction. But still, for May 2013 until the shutdown they were spending $1.20-1.40 on AWS to service every $1 MRR.
So growth in their userbase was actually increasing their total burn rate, which makes growth unsustainable.
Your analysis is certainly logical, but with all due respect relies on the faulty assumption that Everpix's infrastructure, responsible for almost all of variable costs, was somehow "frozen".
The reason we were getting closer and closer to being positive on variable costs (looking at revenues on a sales recognition basis, since from the sales volume perspective, things were already positive in the last few months) is, yes, improved monetization, but more importantly AWS optimizations. We had squeezed a lot out of S3, then EC2 and our last step, the one we were working on before shutting down, was RDS where there was a ton of room. Looking at our trajectory, I'm pretty confident we would have been positive even on AWS (but unlikely much). You can even compound that with AWS discounts which apparently even startups can get looking at other comments in this thread.
To re-iterate, you would not build a freemium business like Everpix, with intense computing and storage requirements, at scale on AWS. Large photo platforms have their own storage and servers. Everpix was never intended to grow out of AWS either and you know you can cut your infrastructure costs to at least half. That was the plan post Series A. Most CTO / CEOs of large established photo companies I talked to said they were doing much better than this 1/2X, so that's conservative (I've been told ranges of 3-8X in savings vs S3 for instance).
Anyway, Everpix was bringing about $6 revenue / user / year IIRC, which is the part that really matters considering the denominator, infrastructure costs, is something you know you can really bring down. One relevant reference point is 500px for instance which is doing about $1 revenue / user / year [1]. I would be surprised if they were growing "sustainably" too ;)
Oh, yes you're absolutely right. Growth didn't have to remain unsustainable... reducing costs on AWS seems like the right idea (and kudos for huge efforts there with spot instances, etc!)
My attempted point was just that during the lifetime of the company, growing the userbase was increasing burn rather than decreasing it. That's a normal VC-funded approach to things, and occasionally works dramatic wonders like Facebook. But in the case of Everpix, the negative operating margin helps explain why an exponentially growing, paying(!) userbase can still fall apart without additional funding.
Btw, I really appreciate you guys being willing to share all these numbers. Very brave, and extremely fascinating. Best of luck in your next endeavors!
Why would it be unusual for the majority of costs to be salaries for a company that sells services and develops products? This has nothing to do with the "talent war." That's just the way things work. What else would they be spending money on?
> AWS - $394,588.35 - or 14% of total costs
How much higher would that number have been had Everpix built out their own infrastructure?
My points weren't criticisms, they were merely highlighting the sentiment that comes from outside "the bubble" that is SV. It is unusual to people outside SV* to see such a high number (the % was just an additional data point that signals to non-tech people how much salaries constitutes tech startups). This has everything to do with the talent war - the market dictates that an engineer's salary must be $foo because Facebook/Google/etc are willing to compensate that much. People are generally curious why that "that's just the way things work" who aren't familiar with it.
> company that sells services and develops products
Here is my criticism now - not all products and services are equal, yet the costs to develop them (in SV) are. FB/Google/etc.'s high revenues dictate that their products and services constitute a developer that costs $foo - they set the market rates. Talented developers are fungible, but their salaries are not - thus the disconnect.
> How much higher would that number have been had Everpix built out their own infrastructure?
I don't have the answer to that question, and probably very few people do. There are two points to be addressed there - (1) could they save money by building their own? mature companies deal with that question all of the time (2) this is a real life indication of how much a software services money can make on one client by "selling pick axes in a gold rush" - thereby singalling a reinforcement to that sentiment.
I am intimately familiar with high-level metrics at software businesses in many locales other than Silicon Valley, and direct cost of employees is the largest expense at all of them. (Not true generally, since marketing budgets get arbitrarily high at some firms, but true of all dozen or so that I have direct experience with.) It can easily range to 80%+.
Also, the imputed salaries being paid here are very much not Google/Facebook numbers. The payroll for a team of five Goolgers is plus or minus $100k a month, almost double what it was for Everpix. Everybody was likely taking far below-market salaries.
I did a bit of Excel modeling of this, based on Everpix's hiring timeline and published numbers. My count is that they had 157 employee-months prior to shutdown at a fully-loaded cost of approximately $9,200 per employee-month. An employee-month is one month worked by a founder or employee, but not a consultant. Fully-loaded costs includes recruiting, healthcare, salary, and taxes, but (for the purpose of this calculation) does not include pro-rated costs for rent, hardware, software licenses, or what have you. Those are all numbers I'd be careful to bake into my planning as a founder but which I don't need to reverse the math here to salary packages.
I would ballpark the average package at Everpix as what Americans would understand as "$90k plus healthcare plus equity." This is not straight-line comparable to "$90k plus healthcare" at other US-based software firms, even those with baseline perks packages for working professionals rather than those competing in the Valley perks arms-race with Google and Facebook. For example, if one were to take "$90k plus healthcare" at an insurance company doing Java EE XL mappings every day, it is virtually beyond question that there'd be 401Ks with employer matching and whatnot in the picture as well.
If you were to do a straight-line comparison with e.g. Google or Facebook, this is substantially below their pure-cash offers for fresh graduates with no professional experience, and both would add both substantial cash bonuses and incentives plus equity grants plus perks with non-trivial cash value. The compensation packages employees capable of solo-shipping e.g. iOS applications or non-trivial image recognition software would be offered are substantially north of that. $140k cash-on-the-barrel plus industry standard perks isn't even an eyeraising number for intermediate engineers with no skill more specialized than shipping working software.
I'm mostly doing this math to say that "Lavish expenditures on employees were not a contributing factor here. They paid substantially below-market wages for employees doing similar jobs locally." I know that this might cause people who are not located in the Valley to feel a bit of heartburn given the absolute numbers, but that's another discussion entirely. (It should be noted that I live in an area where $90k cash packages would be unusual for engineers with 20+ years of professional experience.)
This breakdown is characteristic of software startups, not SV in particular.
I was involved in the Web startup community in St. Louis, Missouri for years before moving to SF. It was the same for those companies: salary costs dwarfed all other costs.
Unless I'm mistaken, workforce is the largest single expense of most businesses, not software startups in particular. If anything, as a percentage of operating expenses, 52% sounds kind of low.
A million dollars is a lot, but I have no reason to think it's crazy. Without very important context, hearing "a million dollars on salary" isn't useful.
Aren't these numbers also so high because they took VC funding? The VC wants to throw the company against the wall as fast and hard as possible to see if it sticks. That means no time to roll out custom infrastructure, and hiring rockstar consultants just to build the product as fast as possible.
The idea is not terrible, it's definitely better than doing everything the right way (tm) for 5 years, just to find out there is no way to fit the market.
>That means no time to roll out custom infrastructure
It is sad that the industry is so anti-intellectual that we're actually afraid of computers now. Giving dell a CC# and giving them your hosting providers shipping address is not a complex or difficult task. There is no need for "custom infrastructure" to run a trivial web app. You buy servers, you plug them in.
Depending on your service provider (hosting, msp, mttos, whatever), the up-front msa can be truly painful if you want "real" service. I.e., redundant power and multi-provider network connectivity with multiple connectivity points that have a massive capital cost but will be as reliable as your commercial or home utilities.
Couple the cost of establishing the relationship (especially with startup credit) with the NRC and MRC of hosting services, smart hands rates, reserved cabinet or cage space for those surprise boxes on the dock, and incremental pricing models, and it's very difficult to build out and manage poorly predicted growth in a financially constrained environment.
Saving money on this means you're allocating additional engineer hours to dealing with this yourself.
The 'middle road' is probably an operator in a decent NAP or superNAP that manages the big lease and then carves out a margin by offering smaller space, possibly even colocating your hardware in the same rack as someone else's (and maybe not in the same rack with your other gear), and charging more than a macminicolo but offering significantly improved and variable services.
There are a number of walls you can hit that will present themselves as an almost insurmountable obstacle when you own your own gear. There are mandatory professional services when setting up some gear (SAN vendors have been notorious for this) and it's just a bunch of attention you have to devote to something for which there is generally a good enough solution for a company that may not exist long enough to depreciate the servers that ate a huge amount of their startup capital.
It's not that we're 'afraid' of computers. It's like saying we were afraid of electricity or telephone lines 40 years ago. We can manage it, but it's almost irresponsible and risky to do so.
You are inventing a problem that doesn't exist. I can literally have a cabinet in a first class data center in 2 days and it takes 15 minutes of my time. You save assloads of money this way vs AWS which is incredibly expensive. You can buy an equally performing server for the cost of a month of renting the VM from amazon (for the big ones obviously, you can't buy servers as slow as a ec2 large instance.)
Alright, who in your company knows what hardware to buy at which datacenter for how much? What software is going to run on it? who picks that software? How do you get data onto it? How do you get data off of it? What happens when it goes down? How do you monitor it? How do you scale it to two servers? Or more? Who is going to do that? Who is going to know all that? Who else is going to know that so your company doesn't fail when the first person isn't around?
>Alright, who in your company knows what hardware to buy at which datacenter for how much?
Everyone? I don't understand how you can imagine a problem like this. "Who in your company knows how to tie their shoes?"
>What software is going to run on it? who picks that software? How do you get data onto it? How do you get data off of it? What happens when it goes down? How do you monitor it? How do you scale it to two servers? Or more? Who is going to do that? Who is going to know all that? Who else is going to know that so your company doesn't fail when the first person isn't around?
AWS solves precisely zero of those problems. It isn't magic pixie dust, it is renting Xen VMs.
In addition to being fast to deploy, cloud infrastructure lets you easily scale up and down to try things out. This is useful if you're still finding traction or distribution channels that work best.
>> Giving dell a CC# and giving them your hosting providers shipping address is not a complex or difficult task
Financially, cloud servers have properties that might make them attractive relative to this. Also -- buying and managing colo is a fairly specialized discipline. There are a ton of business reasons to not take on a risk in this part of your business when you can outsource it at nominal cost. A COGS of 14% is quite nominal.
>In addition to being fast to deploy, cloud infrastructure lets you easily scale up and down to try things out
And you can buy a single server for $5k that will outperform $5k/month of EC2 VMs.
>Also -- buying and managing colo is a fairly specialized discipline
What? That is like saying "buying and managing phones is a fairly specialized discipline". No, it isn't. You give money, you get service. Not complicated. People have been able to handle this just fine for longer than you've been alive.
I wonder how much of the resentment to owning hardware comes from the "fail fast" approach - if you're trying to vet a business model quickly or move onto the next thing, having actual physical assets to take care of if the company ends up failing could be an off-putting extra complexity for some people vs just clicking "delete instance" with some virtualized provider.
I think it has more to do with initial fixed cost.
Personally I used my cofounder's existing array, but that's not an option for most companies, particularly starting out, and cutting out server purchasing costs pays in the short term. However, if you have the initial funding, as in this instance, it pays to buy the servers you need.
Plus, it's an asset that, should your company fail in 6 months, can be sold to at least salvage some of the initial investment.
[edit]In the last boom, though, people bought millions of dollars worth of servers and equipment that are now less than worthless; in many cases you have to pay to get rid of them. So, I guess there can be a significant downside.
> It is unusual to people outside SV* to see such a high number (the % was just an additional data point that signals to non-tech people how much salaries constitutes tech startups)
It's not unusual at all for payroll to be the vast majority of a company's expenses. Normal profitable business usually expect payroll to be 30~40% of gross revenue (and that's for companies actually earning money, not during bootstrapping), and 50% exists although it's usually considered dangerous for profitability.
And of course that depends on the domain, a hairdresser and a refinery will have very different payroll rates and structures. Pretty much all of software and saas company activity and value comes from the people working there, in fact I'd expect payroll %-age to be higher (not lower) outside of SV: lower rent and likely lower utilities.
Economically increased supply will decrease the price. Supply & demand. Another way to increase supply and thus save money is to hire remotely, and not restrict yourself to a small geographic area. If you can hire a guy in upstate new york where a house costs $40k all in, that $1000-$2000/month saved in after-tax rent costs (~$17-34k/yr pre tax) alone will significantly decrease their required wage.
If the supply of labor is reduced, the price of labor will rise until demand falls to match supply. That means that some positions will go unfilled; positions whose returns don't exceed the cost of the labor. But we don't know the total value of that labor. The community could be worse off altogether with those positions unfilled, because price doesn't capture all the information - we don't use money for every social exchange, and nor can we reliably estimate the present value of future contributions.
For example, many US companies are created by immigrants (18% of Fortune 500, or 40% if you include children of immigrants). Steve Jobs was the child of an immigrant. Hiring a guy who is so enthusiastic about getting a job that he wants to live in a rural area and live on a pittance, vs hiring someone who's ambitious and hungry - who's going to create the most value?
If you're paying H1s less than locals you're violating immigration law. The idea behind the H1 program is to help you when you cannot find local expertise in your specialty and need to recruit globally. It's a very sensitive issue in the USA if you are bringing in less expensive foreigners to "steal" jobs from locals.
That said, it's prone to abuse and I have seen some H1s with really poor compensation due to a weak negotiating position (the difficulty of switching jobs while in-country, sponsored adjustment of status, etc.).
This is pretty dense stuff. We're working on putting together a more extensive and user-friendly site to make sense of this heap of data. Stay tuned. :-)
6800 subscribers. 2,665,192.34 in expenses.
50000 users.
$395 per subscriber per year. $35/mo.
Convert 10% more of users.... 5000 and it's still $20/mo.
Was this really $5/mo or $40/yr?
There had to be a story about HR plateauing and subscribers increasing an order of magnitude in 12 mo. - and it's still under water for another 2 years.
Who saw the writing on the wall and when did they start raising the warning flags? What happened from then on? What warning signs (talent leaving, investors glaring) showed up? When? How did they take form?
I went through an exercise to recast the numbers as a bootstrapped company here in New Zealand. we don't, for example, have any healthcare costs, and items like legal fees, rent and salaries (including taxes) are substantially lower.
I have 3 Questions for the founders, if possible:
How much did you play with pricing to try to drive revenue per customer up? e.g. Did you consider/test no free customers, higher pricing and gold tiers?
The consultants cost of $272k seems very high - what sort of consultants was this spent on?
If you did it again as a bootstrap (i.e. founders are in control), what are the to 3 things that you would you do differently that would impact on costs and revenue?
The trend towards startup transparency is pretty damn awesome as a startup founder myself. Even with startups that have nothing to do with photo sharing, being able to see the guts of a (failed) startup is extremely insightful.
Thank you for this. Reading the VC letters was just so amazingly refreshing I can't begin to tell you. I love how they said "no" in almost all the same ways.
I'd really like to know the two that were so standup, that they wouldn't fund a competitor or someone who could, eventually, be a competitor. In all honesty, those are the types of funds you want backing you. They've bought your vision and they're not going to do anything to undercut you.
Are you kidding? The "we can't fund a competitor" VCs are just saying "no" like everyone else. There's nothing particularly noble (or ignoble) about it.
In the particular case of Everpix, I'm relatively certain I can guess who these VC funds are and who that competitor is. Given this information, it's not just some empty excuse for a VC to bow out.
It's not uncommon for VCs to have policies of not funding competitive startups, it's in their own interests as well as the startups.
If a VC has funded two competitors it means the startups become reluctant to share information with the VC, the founders of the first startup become unhappy with the VC and follow-on rounds become very messy and cause bad politics.
Certainly it happens (in some cases due to pivots, etc.) but I imagine it's something that most funds would prefer to avoid.
Could Everpix have been run by a very small team at the end, rather than shut down? It looks like it was near adding a net 1,000 subscribers per month.
Also, wouldn't it have been possible - with the understood pain points - to substantially reduce costs by moving off of AWS? Trading the ease of AWS for the critical cash for operations.
Yes they probably could have. I'm purely guessing here but I suspect the founders had a decision just like the VC's. Does anyone want to keep this alive and try and salvage it? It would be doable but a massive grind for 18 months and we're probably never going to build much more than a $5 to $10 million per year business even if we really do make it. Or do we just want to chalk this one up and walk away and use the experience to try and hit a home run next time?
Buried beneath a stack of preferences it makes little sense for founders to make any sacrifices to save the company (i.e. not take a salary for a while) since all money will go to investors first.
If you read through the VC correspondence the hang up was that the space simply isn't big enough (>$100M). According to the VCs the hockey stick would have hit a ceiling too low for them to invest.
Series A investors make very few investments that are time consuming and capital consuming. Often, a specific partner will make just 1 investment per year, and stick with that company for 7+ years. They'll also allocate ~$15m in capital to it (the $5m for the Everpix raise, plus another $10m for follow-on investments in future rounds). So they're not willing to take risks like this, it doesn't make sense with the model.
You know at some point we are going to have to sit back and ask "Where the F&*^K" does this $100M or $1B business number comes from.
The minute I get the opportunity to put my money my mouth is, I will invest solely in lifestyle businesses: one to two guys/gals wanting to build a profitable business with a yearly dividend, one where if the team makes $500,000/year we are all extremely happy.
You can already do that; invest in existing companies that pay a dividend. They're established and they understand that their investors value their continued profits more than risky behavior.
> "Where the F&^K" does this $100M or $1B business number comes from*
Funding rounds that want to raise millions in VC are for companies that aim to eventually be a $100M business.
A smaller business wouldn’t involve venture capital. And the bank is still a viable option if capital is needed to scale, as long as there is a realistic repayment plan.
This is what separates you from the real movers and shakers: they just want to get in, get theirs, and get out, and they believe they're sufficiently smarter than the rubes to do it.
$30K/month and ~100 servers seems like a huge waste. I'm in the business of managing similar services. I handle, not photos, but video streams. My work is for one of the largest companies in the world (Fortune-5) and I deal with things like superbowl traffic spikes, etc... Everything is handled in house but for spiky loads, we look to things like Akami.
I realize managing at Fortune-5 is different than 100% oursourcing but the cost constraints, I have to imagine, are very similar. For example, I have to imagine Amazon gets similar price breaks from Cisco as my company - I know Amazon has custom built servers and all so there may be some advantages but at $30K/month, I think any hosting company would happily take 1/2 or 1/4 of that.
I'd be curious to know what kind of bandwidth loads you were pushing through AWS and how much of the AWS cloud can be made redundant per client.
Yes, I took that into account in mulling things over. We do very similar stuff with video streams (post processing) before pushing to Akami. We also use SAN backed storage and Cisco CSM load balancers. All of our equipment is in the US and we use VMWare for everything. I know little about AWS and their virtualization technology but where I work, it would appear we pay a little bit more of a premium for VPS nodes if I assume Amazon rolls their own - the $30K just seems outrageous in this regard. And to top things off, we don't host just one web site; we have a myraid of applications spanning our data centers to deal with all sorts of stuff (many product lines each with their own web site with streaming, flash, user accounts, etc...)
Could anyone please clarify for me following things? I have just started to learn about startup capital.
1. Since company is now closed down what happens to Investor's money ? Do they just loose all or do owners have to return it ? ( sorry if this seems pretty noob but I would like to know it )
2. What does 1 year maturity mean in convertible notes ?
3. Shouldn't the Net Operating Income be negative ? Since they had income of $280696 and Expense were $2,665,192.34 ?
Did they count investor's money as income ? or how did they arrive at $2,384,224.67 ?
Thank you for the reply.
The elephant in the room that nobody is talking about here is
they gave away too much. They could have adjusted their freemium model. Why not contact people using a lot of resources ? Let them know the best way to keep the services up at the current clip is for them to start paying.
That's my feeling t0o. I had been using it for a few months before it shut down and was amazed at the amount I could host for free. I recommended the service to a number of people and would have been willing to pay for it if I needed to.
Also I don't get why it disappeared from one day to the next... Couldn't they massively restrict free usage? Or have some form of donation/fund raising à la Reddit Gold? people were attached to the product and it was disappointing to see it just go away /rant
These are very interesting numbers. Obviously the financials are not well suited to the VC-backed startup model, but I wonder if Everpix could have been successful as a small business run by a small handful of people.
In the "About Everpix" part, the text make it seem like the app was doing ok and the reason they had to shut it down was because investors didn't want to invest in it anymore for whatever reason.
Well, the lack of investors was surely a big symptom of their failure but the cause it was not.
When you burn > $2 million in funding ($200k alone in PR and "Promotional Expenses" as their spreadsheet shows) and still can't get through the timid mark of 50 thousand users, then maybe there is something very wrong with your business model that you should think about.
I'm confident that's something that crossed all the investors minds during the series A conversations. They didn't mention that fact directly because it could sound too harsh or rude and they wanted to keep the relationship in good spirits, obviously.
The startup game is tuff, no doubt about it. I hope the Everpix guys learn their lesson and to better next time, they look like cool and competent people.
The PR expenses (109,552.34) seem pretty high and I wonder what that line item entails. Did you guys use a PR firm or run any expensive marketing campaigns?
No marketing campaigns (we spent maybe ~10K on online ads mostly for testing). We did use a national PR firm though for several months and PR contractor for a couple months as well before that.
My message to all startups out there: despite Valley folklore, it is ok to have a marketing budget and spend money to advertise to users - especially initially. Not every successful company or service grows "organically".
First - this is a brilliant gift to the startup community - thank you. I am interested to know more about the marketing strategy of the company. It looks like you had some very successful campaigns with CyberMonday and Daring Fireball but spent very little? http://pollenizer.com/everpix-have-open-sourced-their-startu...
Why is the main focus on Aws costs when in comparison the HR costs are multiple of the Aws costs? Is it because were too technically minded here? Is it cause the charts in the article only showed the Aws costs whereas including the HR costs would have changed that perception. I don't know but i just keep seeing people focus on the Aws cost which was not the real problem in comparison.
Note my other comment. It hints at this. You make the point clearer.
The technical costs are not damning. Where is the concern about the other costs, which aren't being picked apart? If AWS was 0, it's still unsustainable growth until someone (swisspol?) dispels our concern there.
Another big void is who was pulling the plug and how/what are the founders and direct employees feeling? I'd like to know if there were ship-abandoners, etc.
What is the turning point and what were they decision makers watching that caused the shift?
This doesn't look like a technical learning here. This looks like a business learning. So, while they're being transparent and it's still fresh :) share.
We can A/B test buttons. How do you A/B test business decisions?
It looks like a bunch of people can armchair this thing, technically, to the point where it makes me wonder if someone could NOT take on nearly the funding and make the $100M company on this idea with just some bootstrapping.
Also... how was there not a pricing model that scaled with use?
What was the roughly $300,000 in consulting fees for? How do the typically help startups? And did you find them yourselves or were they recommended by the VCs? Is it business/process consulting? Do you feel not having a non-programming founder was beneficial or a hindrance? This is awesome that you guys did this btw. Really eye opening. Thanks.
Everpix team, thanks for being generous. These metrics and insights are a gift to the startup community. Thanks also for choosing SendGrid to deliver your transactional and marketing email. Judging by the different categories of email you sent, you put a lot of thought into your email program. Question: How essential was reliable email delivery to your business expansion? What lessons did you learn about using email effectively?
(BTW, would you consider publishing more of your email metrics on GitHub? For example, SendGrid reports the type of email clients that read the mail you sent (e.g. desktop, mobile, webmail), and the mailbox providers that you sent the most mail to (e.g. Gmail, Yahoo, Hotmail, etc.) Here's where to find and download those reports: http://www.screencast.com/t/5mW9bm4Ueq)
with all the lamenting about how hard running a consumer photo service is, I wonder how imgur does it. their traffic must be insane, and they have no subscribers, so how come they can do it and these relatively upmarket services can't?
Imgur is financed by advertising, and they offer pro accounts, admittedly subscribed by a sliver of users only. Revenue per user are orders of magnitude lower than for a for-pay service, but they have far more users. Their servers (and costs) are optimised to serve a limited amount of photo fast, host a little more and serve them not so fast; they most likely have far less photos, at far lower resolutions, than Everpix who hosts entire shooting sets for pro-am photographers.
Everpix's product was loved by their users. They absolutely built "something people want". Unfortunately the promise they made their users was very expensive to deliver.
I just read all their pitch decks and they never mentioned exactly how Everpix was - or could be - better than Dropbox, which IMO would be the first obvious competitor.
The overall problem they were attacking is real, so maybe if they were more focused and had distilled better their solution, with a smaller and lest costy team, they could have made it.
Just my opinion from a totally outsider perspective. Take it for what it's worth. I want to congratulate the team for trying and wish better luck next time.
Do startups at a similar stage usually measure things as diligently as Everpix seem to have done?
If they kept an eye on of all this on a monthly or weekly basis while they were still running would they have been doing too much or too little?
I've worked with some organisations that are reluctant to invest in this level of analysis. Some of them might even now say "It didn't help Everpix to spend time on that stuff."
Do startups at a similar stage usually measure things as diligently as Everpix seem to have done?
Frequently, no. There exist companies with revenue in the $X0 million range who have less numbers floating around the company. You'd be surprised.
If they kept an eye on of all this on a monthly or weekly basis while they were still running would they have been doing too much or too little?
I'm going to respectfully decline to answer counterfactuals about someone else's business. In general, software businesses have two tasks: Make software. Sell software. Analytics is a very useful thing to have if it lets you accomplish one of those two tasks. There are many software businesses at which it does. There are some at which it does not.
So is there another service that does what everpix did? I had it in the back of my mind to jump in because it's a service I've really wanted, but didn't get to it before the great shutdown. I've looked at Loom and I use Trovebox, but neither seem to support videos very well, Loom is iOS only, and Trovebox is moving to a different business model.
After Everpix shutdown I looked around and found both Loom and Trovebox but I also found PictureLife. I tried both Loom and Picture life for a little over a month and settled on PictureLife. Check it out, I am very happy with it.
This is very interesting, thanks for taking the time to compile it all.
What I'm wondering now is whether you plan on open-sourcing the technology built so people can potentially plug it into their own storage providers and take up the cost of storage, or adapt the photo analysis tech. You've built a wonderful base here, will it be put to some use?
To me it looked like Everpix expected the kind of growth usually seen with free services from a paid-subscription model.
On the one hand, paid subscriptions "make more sense economically", on the other hand, as long as we are talking about high risk capital anyway, a free service model might have made more sense "financially".
This is a huge help the community, thank you! The completeness of this data is a testament to the team's diligence and in-depth understanding of startup metrics. Regardless of the outcome of Everpix, this analysis shows that the team has developed some good processes and best practices for their future ventures.
Interesting to see Bertrand Serlet (EDIT: shit, I confused my ex-Apple Frenchmen. Jean-Louis Gassé was BeOS), the guy behind BeOS, among the early investors. He seems to have an affinity for good products that just don't quite make it.
This is a REALLY useful treasure trove, thanks for posting
Looks like the largest expenses were payroll and hosting. I wonder if tbis is the reason they shut down even as the number of subscriptions was rising.
From [1], the vast majority of the AWS cost was Storage ($9k in Oct 13), a large ECC instance from Feb-March 13 ($7k per month), and backups ($5k in Sept 13)
AWS costs seem reasonable to me and it is definitely the smart way to build a service like this. Saves a ton on time early and services like this can optimize later when they are much larger.
I wonder why they didn't simply downsize on the hosting. I mean if that's a huge part of the operating cost, so much you need to shut down the entire operation, doesn't it make sense to go with a different host? even resorting to a dedicated box or choosing a different cloud provider?
It's really hard to believe because I bet the end user didn't really give a damn about whether it was hosted on amazon or elsewhere. It's also hard to believe why not just outsource the generic stuff. I really don't see anything unique about everpix that a North American can only do and not some guy across the ocean. Was it absolutely necessary to have in-house talent for everything when it was a threat to the continued operation of the company?
I wonder if they were simply bootstrapped and grew as a small business focusing on net profit, would they have made a better decision.
I keep reading comments from this angle, so I'll throw in my 2 cents as an Everpix founder :)
I can assure you that without the facilities of AWS, which saved us a ton of time and development overhead early on, there would have been no Everpix. Period. Actually, the very first Everpix was on Google App Engine, but that's a different story.
As an early stage startup, whose core business is not building infrastructure, do you really want to be in the business of managing for instance your own S3 type storage with the same performance and reliability at 200+ TB scale? We never lost a single user photo or account. Never lost a database server either. When your users trust you with their life photo collection, that does matter.
Then of course, if your company starts gaining traction, your bills increase so things change. As a matter of fact, post Series-A, our plan was to switch out of AWS, concurrently to our infrastructure's redesign to go from managing 100s of million of photos to billions.
Doing the switch earlier would have been premature IMO: there was no one the team with such experience, even less bandwidth. The opportunity cost would have been significant on all the other aspect of the company. Let's not forget the required upfront capital which takes some time to recoup so savings would not kick in from day 1. Finally there's also the fact our infrastructure was still rapidly evolving (type of EC2 machines used, database architecture, etc...), so even buying reserved instances was very difficult as there was a good chance of buying machines we would not need anymore 6 months down the road ultimately wasting money (it did happen).
So what we did instead what focusing on driving continuously driving down AWS costs through various optimizations - if you were to look at our dashboard graphs, you would notice AWS costs were growing slower than users / photos, which reflected that work.
Long story short, the infrastructure was paying for itself through subscription revenues. The real "killer" was payroll. I use quotes because in this type of business, it's expected to have payroll your largest expense - there was no surprise. BTW note that the team was already 1 person = 1 entire product component like iOS or infrastructure so quite lean and highly productive. For a VC backed consumer business, you pretty much need the VC money to cover these fixed costs until you reach profitability.
> it's expected to have payroll your largest expense - there was no surprise
Only when you are small haven't marketed the service and your subscriber numbers are still small. You could probably support 10x the subscribers with only 2x (or less) of an increase in staff.
Thanks for clearing that up. I'm quite surprised to see how large the salary cost is. That and the 400k for consultants ( was this for computer vision experts?)
I'm kind of impressed looking at the landing video, it's a neat idea, it's sophisticated using CV to organize photos. However, the market just didn't respond enough to cover the payroll costs.
Did having a VC funding in anyway impact your decision making to grow faster than you should?
I guess it starts with an idea, does it fit? if you have VC money you can see the answer faster but I fear that it puts an expiry date on your product if the answer is still no by the time you can't find another investor. Shame because investors are not always rational. Maybe the next investor could've seen the vision behind your product and pushed. Sad to see a seemingly great product disappear from the market and reducing consumer choice because they had to 'go big or go home'.
I also wonder what it would've been like if you started in a market where tech salary was on the low side, such as Vancouver, BC, where there are laws set in place so that you can work your developers and designers to death out of fear of losing their job in the scarce market and not having to pay any bonuses or overtime.
I wonder what the outcome would've been if you weren't funded, and grew customer by customer. Doing this would've fine tuned the margin between cost and profit and forcing you to grow at a rate that is right for your size at the right price at the right time (where supply and demand curves intersect) just like mom and pop's business.
Thanks for sharing this boatload of information on how a company works for those that never ventured out to startupdom. It really is in the spirit of hackernews. Brutal honesty.
Now the provider is either looking at generating $240,000 for the next 12 months or losing everything from your account. He/she is probably doing the deal with you because much of their costs are fixed. But I don't know if/how you have that conversation with Amazon.
EDIT: Oh and fantastic that they shared.