Ask HN: How to sell a script that will save companies $100k a month?
30 comments
As someone in charge of cloud infrastructure I would look at your claims with a lot of skepticism. Very unlikely a single script will apply to the infinite variations in cloud infrastructure, and the business requirements and organizational compromises implemented and implied. The "$100k a month" claim right off communicates a bespoke solution with savings based on a single specific company setup.
AWS and GCP (I don't use Azure) show me possible over-provisioning and unused resources, and have tools to do more digging and identify savings. I have seen ads/articles about tools that claim to reduce cloud hosting costs. Where does your tool fit? Why would I trust it with access to my hosting account?
AWS and GCP (I don't use Azure) show me possible over-provisioning and unused resources, and have tools to do more digging and identify savings. I have seen ads/articles about tools that claim to reduce cloud hosting costs. Where does your tool fit? Why would I trust it with access to my hosting account?
+1 on turning this into a consulting offer.
Companies won't care that you just run some scripts, in fact they won't care much how you do it at all. They just want testimonials that you've done it for other companies and then price it to match the value they get.
E.g. if it's really $100k/mo and you were viewed as a reputable consultant in this area, so long as you can prove that savings, charging $100k to do your routine is possible.
Companies won't care that you just run some scripts, in fact they won't care much how you do it at all. They just want testimonials that you've done it for other companies and then price it to match the value they get.
E.g. if it's really $100k/mo and you were viewed as a reputable consultant in this area, so long as you can prove that savings, charging $100k to do your routine is possible.
He can’t tell them anything before he actually sits down and has a long conversation with the customer and determines their current state. “Consulting” is as much about understanding the customer as understanding technology.
There are tons of products that do this. Are you familiar with the market you are trying to enter?
Look at what they are doing.
Look at what they are doing.
This. I've also worked at several places that had their own in-house systems that acted like the equivalent of a smart thermostat, spinning up/scaling down resources across massive deployments targeting AWS, Azure, etc. to hit target monthly goals.
Try to do some preliminary research in the area to see if your script is sufficiently niche from the established players in the area of cloud optimization.
Try to do some preliminary research in the area to see if your script is sufficiently niche from the established players in the area of cloud optimization.
spinning up/scaling down resources across massive deployments targeting AWS, Azure, etc.
Isn’t this, like, a primary feature of AWS, etc? Why need an inhouse system for that? Or why need AWS then?
Isn’t this, like, a primary feature of AWS, etc? Why need an inhouse system for that? Or why need AWS then?
No. Cloud providers, like AWS, Google and Azure, make most of their money off of poorly configured enterprise infrastructure -- unused drives, VMs and K8S cluster. They provide controls that allow you to ramp down resources, but it is up to you to do the ramping.
This is not true. AWS has all sorts of tools and guidance to help you optimize spend.
I can say a lot of negative things about AWS and the culture from having been there 3.5 years. I worked in the consulting department (Professional Services).
While we wanted the customer to bring more workloads into AWS, we were all encouraged to help save the customer money where we could. No one got brownie points for coming up with a more costly solution than necessary.
I can say a lot of negative things about AWS and the culture from having been there 3.5 years. I worked in the consulting department (Professional Services).
While we wanted the customer to bring more workloads into AWS, we were all encouraged to help save the customer money where we could. No one got brownie points for coming up with a more costly solution than necessary.
I concur. I work in architecture for a large public cloud provider. I'd definitely double-check that something like this isn't already in the market (or open-source).
It would be interesting to see a novel script though.
It would be interesting to see a novel script though.
1. Fix it in your company.
2. Create a Youtube video of how you fixed it.
3. Spread that video far and wide.
4. If you get traction consider adding an offering
Be the go to person on Youtube for infrastructure cost savings.
The enterprise is a shitty place to find recognition. Seek it elsewhere.
2. Create a Youtube video of how you fixed it.
3. Spread that video far and wide.
4. If you get traction consider adding an offering
Be the go to person on Youtube for infrastructure cost savings.
The enterprise is a shitty place to find recognition. Seek it elsewhere.
My immediate thought is consulting, or you build some wrapper tool which monitors cloud configurations and looks for issues.
That said, if all it takes to fix this is a script, I’m guessing other developers/IT will or have done it on their own. It sounds a little like it’s your job to fix this and you’re simply declining to do so lol
That said, if all it takes to fix this is a script, I’m guessing other developers/IT will or have done it on their own. It sounds a little like it’s your job to fix this and you’re simply declining to do so lol
First thought is: how complex is the actual implementation?
Second thought is: how are you determining how applicable these "misconfigurations" are to other companies? How do you not know this $100K in savings is due to idiosyncrasies at your company that doesn't extrapolate to others?
Second thought is: how are you determining how applicable these "misconfigurations" are to other companies? How do you not know this $100K in savings is due to idiosyncrasies at your company that doesn't extrapolate to others?
My 2c: don’t do it as a ‘software company’, do it as a ‘service company’. You would be selling an outcome to your clients, your script would merely be an ‘internal only’ tool used by you to deliver an outcome to clients.
GTM would be a combination of cold outbound and content marketing. Start with your personal network, do warm outreach, and try to reach one/two degrees of separation to get potential starter clients. Give 2 to 5 starter clients a discount(or for free) in exchange for testimonials(social proof). Once you’ve completed delivery to your starter clients, leverage the testimonials for cold outbound and content marketing.
GTM would be a combination of cold outbound and content marketing. Start with your personal network, do warm outreach, and try to reach one/two degrees of separation to get potential starter clients. Give 2 to 5 starter clients a discount(or for free) in exchange for testimonials(social proof). Once you’ve completed delivery to your starter clients, leverage the testimonials for cold outbound and content marketing.
Look at existing companies in that space, eg https://antimetal.com/, for inspiration and to see if what you are doing already exists.
This sounds like a good idea a software engineer came up with, while his QA and DevOps colleagues are both amused in horrified.
Your script is probably going to work 85 to 90% of the time, optimistically, 5% of the time it will do nothing, and another 5% of the time it will cause mass outages, and a massive loss of profits.
It's better to over allocate resources, let's say you run an e-commerce platform that makes about 30 million a year. Right now you're spending $10,000 per servers per month, and in a typical month you really only need to spend $7,000. But a new product just went viral and it's totally selling out. Does it feel like a good idea to lose millions of dollars to try to save a few thousand bucks ?
Your script is probably going to work 85 to 90% of the time, optimistically, 5% of the time it will do nothing, and another 5% of the time it will cause mass outages, and a massive loss of profits.
It's better to over allocate resources, let's say you run an e-commerce platform that makes about 30 million a year. Right now you're spending $10,000 per servers per month, and in a typical month you really only need to spend $7,000. But a new product just went viral and it's totally selling out. Does it feel like a good idea to lose millions of dollars to try to save a few thousand bucks ?
Context: I work at a 3rd party cloud consulting company now as an “staff software architect” and I have worked at AWS in the Professional Services department in the past. I find it really dubious that you could save companies $100K a month with a magic script seeing that cost savings come in many varieties and often you need context to know why they seem to be spending more than is necessary.
It often takes refactoring an implementation after doing a lift and shift after determining whether it’s worth it.
And honestly, the way you presented your question, it doesn’t seem like you have any customer facing experience working with medium to large businesses to gauge whether your script is generally applicable to different use cases.
It often takes refactoring an implementation after doing a lift and shift after determining whether it’s worth it.
And honestly, the way you presented your question, it doesn’t seem like you have any customer facing experience working with medium to large businesses to gauge whether your script is generally applicable to different use cases.
You would need to make that script compliant with what the company wants to be compliant with plus their own devops and security practices. The real sell is figuring out how to implement the bespoke human project for each company to save the money.
Also if your script just say for example picks better VM sizes you better be sure there are no outliers where the existing legacy VM size is for a particular reason, be it a reserved instance, negotiated pricing or some technical constraint. Whatever your script does I doubt you can run it as is on any client as root aws and walk away and the customer be happy.
Good luck!
Also if your script just say for example picks better VM sizes you better be sure there are no outliers where the existing legacy VM size is for a particular reason, be it a reserved instance, negotiated pricing or some technical constraint. Whatever your script does I doubt you can run it as is on any client as root aws and walk away and the customer be happy.
Good luck!
Don’t try to monetize it at your existing company. It’s a conflict of interest.
Sleep on it. It’s easy to get excited. Your script could be amazing or maybe it’s not. Take your time evaluating the situation.
Like others have mentioned, this is a consultant or services company opportunity. However, getting your first customer is extremely difficult, no matter how much money you promise to save them.
Think about offering your solution as a SaaS offering. It will take some elbow grease to build out. This will determine how dedicated you are to the whole thing.
Sleep on it. It’s easy to get excited. Your script could be amazing or maybe it’s not. Take your time evaluating the situation.
Like others have mentioned, this is a consultant or services company opportunity. However, getting your first customer is extremely difficult, no matter how much money you promise to save them.
Think about offering your solution as a SaaS offering. It will take some elbow grease to build out. This will determine how dedicated you are to the whole thing.
You’re sure your company wants to save $100k/month?
It would take away from what some VP can say is the budget the manages so saving 100K would be actively negativ for said VP and thus not happen.
Well that’s not necessarily true, but that was the point of my question.
Make sure this person understands the situation they’re in and whether it’s actually a problem before thinking they have some surefire business.
Make sure this person understands the situation they’re in and whether it’s actually a problem before thinking they have some surefire business.
I'd be happy to talk to you more about this. I manage a cloud team, oversee a sizeable budget, and have a lot of experience optimizing cloud spend (along with startup/unicorn/mega corp experience).
The play is, like what others have said, it's your expertise and providing consulting services to others struggling with cloud spend.
hn handle @gmail.com
The play is, like what others have said, it's your expertise and providing consulting services to others struggling with cloud spend.
hn handle @gmail.com
Consulting, you have to show them they are spending the extra 100k first then have them pay you to improve it. Nobody needs to know upfront it’s a script or whatever other tools you use.
If you developed that script in your employer's time it already belongs to your employer, I'd strongly advise against 'clever' ideas like turning something built during work time to independent gig.
Well, if he worked at AWS as either an SA or in the consulting department (ProServe), it’s easy enough to legitimately do work on the company time and have the code accessible to him
after he leaves.
For those departments, once you know the process, you can submit your code - that doesn’t contain customer information - to be approved to be released as open source on AWS Samples with an MIT license.
https://github.com/aws-samples
The approval process doesn’t go through your manager. There is an entirely separate team and the turn around for approval is usually less than a week.
I was able to have access to all of the frameworks and utilities I wrote from scratch for customers after I left by going through the process. It was all above board.
For those departments, once you know the process, you can submit your code - that doesn’t contain customer information - to be approved to be released as open source on AWS Samples with an MIT license.
https://github.com/aws-samples
The approval process doesn’t go through your manager. There is an entirely separate team and the turn around for approval is usually less than a week.
I was able to have access to all of the frameworks and utilities I wrote from scratch for customers after I left by going through the process. It was all above board.
Put a very fancy GUI or web front end on it and provide “enterprise” support.
I helped my company save $25,000 a year, which is small compared to the potential savings your script could achieve. Despite that, all I received was a simple thank you—which, as you mentioned, is typical corporate thinking. Remember, where there’s a will, there’s a way. Don’t give up on your script! Wishing you the best of luck with your endeavor.
Now, I could do it in my company and maybe I will get a promotion for it and that's it. But to be fair, I am tired of this corpo BS I have to tolerate every week and would love to have some capital to quit it and focus on solving actually interesting problems.
I am certain that other companies also have this problem. But as it is solvable with a script, I feel like it would be really difficult to sell this solution to other companies...
So I come to you, HN, for advice. How would you make this into a business/saas/get out of corporate hell hole plan?
I was thinking maybe there is a way to make this open-source but with a cc-by-nc-sa-4.0 license to make big companies pay for this solution. But I feel like if there is no way to detect this being used, there is actually no way to enforce it...