Title should maybe be switched to "Google Cloud vs AWS Onboarding Comparison for YC companies".
Not to say that isn't a useful article on its own, but it's hard to draw too many meaningful conclusions for the rest of us when the first line of the AWS bullet points is "reach out to dedicated YC email".
I was able to get 130k free GCP credits (over 3 years) just by filling out the startup forms... as a complete nobody. You don't even need an affiliate to sponsor you.
I'm not saying anything for or against AWS or GCP. All I'm saying is most startups experience with AWS won't begin with "Send email to dedicated YC email address" so it's hard to draw conclusions from this article about what AWS support is like for "regular" start ups.
While I agree that this is a confound and as someone who also went through the same process, GCP also knows that you're applying as a YC company, and is a PITA regardless - in that way, there's a normalization factor of applying to both entities' startup credit programs with similar social factors. I've heard the same pains with GCP from non-YC founders about the free credits process and from non-YC technologists about bad customer service.
The grants start out small (I think the first one was $3k) -- but then once you spend 75% of them you can apply for the next round, which for me they gave me $17k out of a possible $30k based on previous usage. After I spent around $15k for that over the next year I applied again and got $100k.
Just one note though that I already had an MVP that I was running on GCP at the time I applied (but I was within the $300 free credits that I started with so I didn't pay out of pocket).
If you hit it big, you’ll spend on their platform for a long time. You don’t have to explicitly promise anything because you’ve built everything around their platform for months/years.
vendor lock-in. Good luck switching away from GCP once you're tied in to their cloud services. Might as well spend the VC money and not worry about moving.
Maybe things have changed since I applied - but I have not talked to a single sales person and only communicated via email with the google cloud for startups team. The process was so quick and painless that I at some point felt like I was being scammed and had to make sure I was not going to end up with a huge bill after they pulled the rug out from under me.
It's not too bad if you use the firebase store. But yeah using the firebase realtime db for anything that requires relationships/indexing can be kinda cumbersome.
I've built a prototype on top of it and it's fairly rudimentary but you can make it work. My biggest concern is that there are no case studies I can find of people using it past the prototype stage and how the pricing/scalability works out at that point.
We've spent a lot of time working around the limitations of FireStore, but it does work reliably. Pricing is VERY hard to extrapolate from early use; all it takes is one feature request and your assumptions are blown.
Love to hear from anyone who has gone beyond the prototype phase.
In 2015 I worked at a start-up and was in charge of applying for various cloud start-up benefits at AWS, GCP, etc. At the time getting GCP credits was the hardest and required meeting with a program coordinator after getting a referral from an industry recognized VC. It may have changed since then, bottom line YMMV.
Are there similar forms for AWS? I got 100k credits for my prior YC startup but I'm exploring new projects for a possible future startup and if I could even get 10k credits it would be a big help.
If you're planning on building your business on the cloud, the availability of free credits seems like the worst way to evaluate a cloud provider. Generous "free" credits might even be a warning that they have to pay people to use the platform.
If you truly need cloud hosting then that is $130k of free money no? That's more than many pre-seed rounds. If you use that for a bunch of Linux machines running some kinda containers and a non-proprietary database then I don't see any downsides.
Doesn't the same logic apply -- that new upstart is unproven and is essentially paying people to use their platform. The last thing your burgeoning new business needs is trying to debug issues on a brand new platform, you can't even hire industry experts to help because there are no industry experts.
So if you want to work on building your business and not debug cloud provider issues, avoid the new upstart.
This is also pretty specific about needing more free credits than GCP provides out of the box.
If you just want to get started with GCP, you just sign in with your Google Account for $300 in credit. When you run out you can just start paying with a credit card.
Exactly. Google: gives me $300 of credit no questions asked and tells me what things will cost upfront.
AWS: makes me apply to a program to get credits and if I want to price out anything it's almost a full-blown research effort where I have to dig through documents and cross reference tables between different services and I still probably miss something important. Free tier is opaque enough to frustrate even "use it at small scale and see," because you still have to do the research to see if they are pulling a "your first hit is free but try to scale and WHAM." Oh, but they're customer obsessed, so after much begging and pleading they'll refund half of WHAM, one time.
I don't really agree with that comment. I think all of the megaclouds have undesirably complex pricing structures, but they also all provide pricing calculators. AWS has one here: https://calculator.aws/
If I punch in the services I want to use, I can quickly see how much they will cost, and it's easy. I recommend clicking "Advanced" inside services on the calculator if you want to be sure you're not going to run into an edge case that costs lots of money unexpectedly.
If you use any of the megaclouds without first understanding the price structure, good luck.
Personally, I think a lot of companies would be better off using DigitalOcean or some other medium-size cloud. The pricing is much simpler, and pretty much all medium-size clouds charge $0.02/GB for egress bandwidth, either globally or in the US, depending on the provider.
In contrast, the megaclouds make shocking amounts of money off of their extremely pricey egress bandwidth, among other highly profitable aspects of their business.
Even still, AWS is a solid cloud platform, and they really do seem customer obsessed compared to what I've seen happen on GCP.
I don't really agree with this comment, because we've gone through year after year of new AWS cost analysis and pricing tools that always seem to miss the mark in a major way that's only obvious in hindsight and always seem to be consistently worse than GCP and Azure. They never seem to fix the flaws in old tools, they just tack on new tools with new blindspots. Hence my comment about cross-referencing. If you're willing to put in a lot of effort across tools, you can get a complete picture, but it really does take a lot of effort and foresight into the exact structure that your solution is going to take (which involves cross-referencing documentation). Amazon doesn't make any effort to quote you a price once they have enough information, they always make you work for it. If a price transparency tool is gated behind enough effort, is it really a price transparency tool?
That said, I'll grant you that AWS is not the only megacloud leveraging opaque cost structures. They just leverage them more.
My experience with AWS customer support was not so good. At the time I had very little knowledge about what I was doing so I can't recount all the details here, but the short version is I had a wildcard cert tied up with a resource that wasn't mine (something on Amazon's side was referencing it) so I could not delete the cert which prevented me from doing something with the associated domain name. AWS has no mechanism AFAIK to report a bug in their system, so I had to pay for developer support (which was expensive to me at the time). Even then, the issue was never actually resolved for me. I think I actually switched to a totally different domain name just to get around it. Here's a support thread where people are having the same issue:
> pricing tools that always seem to miss the mark in a major way that's only obvious in hindsight and seems to be consistently worse than other clouds.
Can you provide an example of a service that the AWS calculator doesn't compute the correct price for?
I honestly can't remember ever being surprised by what things on AWS cost, and I don't think I'm that shockingly good at detecting hidden costs.
Cloudwatch costs have been the ones my coworkers complain about. The one that got me was sagemaker -- the console had a bug where if you went to a region without endpoints, it would pop up a tutorial screen and the tutorial screen would "stick," hiding endpoints in other regions. Which led to hanging endpoints, which were covered by free tier for a few days before costs exploded. We had alerts active, but they alerted when we span up the resources, which was expected, and didn't make it obvious that paging through all the regions ensuring there were no running resources (itself painful) and watching daily costs for a few days was insufficient to ensure that we weren't going to have a $700 bill at the end of the month (or maybe $1400 -- I forget if $700 was the half refunded cost).
When I shared this anecdote at Re:Invent, the sentiment at the table was "lol that's cute, here's my story with an order of magnitude higher price tag." There were 5 or 6 of us, and my story was the smallest surprise cost except for one other person who was even greener than I was.
> Can you provide an example of a service that the AWS calculator doesn't compute the correct price for?
Can you tell me how I should have used AWS calculator to prevent my surprise charge? You can't, because AWS calculator assumes you know exactly what you're asking for, and the problem with opaque pricing structures is that you sometimes don't.
Other clouds tend to be much more upfront about "this will cost X," "this is costing X," "you're out of free tier," etc.
Yeah, that was weird to me also. It just sounds like YC companies collectively use enough AWS to get some premium tier support. I assume GCP has something similar, but YC companies don't spend enough there to get it.
AWS has, for a few years now, had a team dedicated to ensuring that AWS wins all of the successful startup business, which they know may turn into huge amounts of revenue for the few startups that succeed. The team was headed up originally by Paul Zimmerman (https://www.linkedin.com/in/paulsloanzimmerman/); he may be a good person to reach out to if you're hoping for some credits.
I definitely raised my eyebrows at "access to AWS YC slack"
It's really nice to hear from high-prestige companies about their experience because it gives outsiders a sense of what prestige might fix. If you're a small startup looking for help using AWS, finding someone to introduce you might really help! Not so much with GCP.
For what it's worth, as a non-YC company, I found the GCP onboarding process to be far superior in every way, and my experiences were essentially reversed. For context, we started with GCP for Startups 1.5 years ago when we were just getting beyond a demo, and with AWS Activate 1 year ago, when we had almost no revenue and just about 1,000 users. We were based in the Santa Fe Business Incubator (NM, USA).
All in all, it had taken our incubator at least 1.5 years to get enrolled in AWS Activate, and 4 days in GCP for Startups.
When we applied for GCP for Startups, we were approved in a week, and were contacted right away by an account manager and a support engineer who offered us help with anything we needed. When we applied for Activate, it took a month with lots of followups, and being shuffled between 4 support team members who were communicating with the Activate team (there was no way to contact them directly).
It definitely felt like our business was important for GCP, and for AWS, it wasn't. It was an easy choice.
Yeah, I think having a YC dedicated mail address at AWS helps. As I am still waiting for a response from AWS regarding joining one of their programs while GCP response was within the day. All access sorted the next day.
Based on other stories though, the customer experiance between the 2 companies appears to be simliar even if not a YC Company.
Google has never had a good reputation for Customer service, they believe they can solve all Customer Service with bots and Automation, this will ALWAYS lead to a lower level of customer service.
Amazon started in retail sales where customer service is king, so it naturally has batter customer service philosophy than Google.
Google has shown zero signs it even desires to have a customer service philosophy that is remotely similar to Amazon, or even Microsoft Software which is somewhere between Amazon and Google on the customer service scale
Of course Google could offer the same thing. My proposed title change would be even more recommended in that case.
My point is that obviously AWS thinks that on average being in YC is going to result in more revenue for them and therefore they prioritize support for YC companies. As a non-YC company I won't get the same treatment which makes any conclusions from this article less useful.
Not to say that isn't a useful article on its own, but it's hard to draw too many meaningful conclusions for the rest of us when the first line of the AWS bullet points is "reach out to dedicated YC email".