Related reading: Ask HN: How did you significantly reduce your AWS cost? (2017) [0]

The top comment is great. Two easy wins:

* Putting an S3 endpoint in your VPC gives any traffic to S3 its own internal route, so it's not billed like public traffic. (It's curious that this isn't the default.)

* Auto-shutdown your test servers overnight and on the weekends

[0] https://news.ycombinator.com/item?id=15587627

For an (offline) alternative, check out → https://github.com/similarweb/finala (written in Go)

I started using Finala a couple of months ago out of experimentation, it works well, but you need to tweak the configuration a lot to help the tool understand your infrastructure, spending habits and normal usage. My colleagues are very happy with the results, we started saving ±1.8k per month, which is pocket change for our scale but is still money that we can spend on other things.

That looks really neat. I guess it doesn't even have to be offline if you deploy it somewhere. I doubt we're currently at the scale to utilise this to its full potential but I've saved it for later.

Thank you for posting an offline alternative!

I'd say the biggest difference is that we handle everything (so that engineers can focus on actual product dev) and also offer white-glove service to all of our customers. For midsized and up, we offer 24/7/365 access to an engineer for any questions or problems with the Usage dashboard.

pretty cool! did you guys (or anyone) posted anything somewhere so we can check out about the usability/deploy/etc...?

Maybe I'm missing the point, but why wouldn't a company hire a contractor to do the same thing you do for a one-off fee? I just don't see why this is a subscription model.

Of course, a company's usage will change over time, but I don't see how that's worth the monthly fees advertised.

> optimization process should be done automatically through a series of sophisticated algorithms

> ...then it's passed to a human that manually verifies each recommendation for correctness

So how much of the heavy lifting is actually done by the algorithms?

I guess many companies just have so much room to save on their AWS bills. So the service quickly pays for itself. It's fine to keep the subscription at least for a while and maybe later switch to periodically model if the only feature it offers is saving.

But usually it's also about visibility. Services like CloudHealth give you much better billing overview than what AWS gives you. Without that data you may have a lot of blind spots and budget management is much harder. I think the long term value is billing visibility even for companies with good AWS experts.

I think the math behind saving algorithms isn't too complicated. It's just hard to get it around your head specially when you have too many resources. So it's not practical to do it without these tools.

You will be surprised how much inefficiency developers can produce. Especially with infrastructure as code it's just too easy to create oversized resources and forget about them.

Many companies also have a lot of inefficiency with their architecture. So it makes sense to outsource savings on EC2 instances to a service like this and let your internal talents to focus on architecture optimizations where you can probably even save more money.

You hit it right on the head, stunt!

Hi petargyurov, great question!

Every recommendation is first produced by a model. Then for a recommendation, we add a human to the cycle that understands the client's infrastructure and adjusts the confidence score accordingly. The savings are then shown on a user's dashboard - alongside an analysis report.

When we onboard a new customer, we spend time understanding their infrastructure plans. Then we keep that information for our human-in-the-loop.

You missed the first question:

> why wouldn't a company hire a contractor to do the same thing you do for a one-off fee? I just don't see why this is a subscription model

This is something I am curious about as well.

Our recommendations break into two categories:

1. Annual recommendations (Reserved Instances, Savings Plans) Sure, if a company only wants these recommendations - they can bring on a contractor once per year to do these changes. This is the one-off fee type contractor the poster is talking about

2. Continuous recommendations (Rightsizing, Spot Management) These recommendations are typically not something you could do once and be done. Rightsizing (which includes zombie instance detection) can happen through time. Spot Fleet management is also a continuous play - spot prices fluctuate and Usage's strategy works hard to maximize savings while minimizing downtime.

You might ask, how does this compare to hiring a contractor full time?

If you're a midsized company, Usage charges $5k/month (or $60k/year). A fulltime contractor in SV would likely be at least $100k/year and that doesn't factor in the number of billable hours (through interviewing) it would take to find someone who knows 50%+ of all of AWS' features.

I don't agree with that cost comparison. I don't think you need a full time contractor for this, maybe just someone you pay to check things out for an hour or two per month. This is a task any reasonable skilled sysad or devops lead should be on top of as part of their daily work. I'd say managing our large(high 5 figures monthly) account takes less then two hours per month to audit.

It really depends on the size and configuration of the organization. It might be manageable at a few dozen people, but it becomes much more difficult to scale when you have 1,000+ engineers all working on different accounts with different architectures and requirements.

Additionally, our growth and enterprise plans all have guaranteed savings. 3x and 5x, respectively.

For instance, if you chose the growth plan ($5k/month) and saved less than $180k after 12 months, we'd give you a 100% refund.

Similarly, if you chose the enterprise plan ($10k/month) and saved less than $600k after 12 months, we'd give you a 100% refund.

imagine companies where all new products are being deployed on AWS, and for every new project, you submit the suggested AWS infrastructure to them for an evaluation.

That's the vision :)

Doesn’t AWS compute optimizer does the same? What’s the difference?

Hi emersonrsantos,

Good question!

With Usage, every enterprise customer gets their own fine-tuned model. AWS compute optimizer throws the same standardized model to all of their users. I can't compare our models to the ones AWS compute optimizer uses - but I can tell you that we've never had a complaint about a customer taking a recommendation and regretting it later on.

Additionally, we have humans look over every recommendation to ensure quality. Typically we have an understanding of a customer's cloud infrastructure before they sign up for Usage - and the human in the loop uses that information to grade recommendations produced by Usage's models.

[DISCLAIMER] software engineer @ AWS, not working on billing.

Interesting project! This product can unlock a lot of customer savings, and wishing you all the best.

I have a couple of questions:

- Are you focusing on EC2 compute costs or are you focusing on other AWS offerings as well? Curious to see how the savings compare to EC2 compute optimizer [1].

- How is your churn rate? What is the value of your product once a customer optimized their AWS bill?

- Do you have a screencast of your product?

[1] https://aws.amazon.com/compute-optimizer/

1. We're currently focused on EC2s (it's been the largest line item for our first few customers).

2. Rightsizing and spot fleet management keep our users coming back - but our customers get the most value the first month. Almost all of our customers have renewed their subscriptions after the first month so far.

3. I don't know how well the computer optimizer model performs, but we haven't had a customer take a recommendation and regret it later on. For our enterprise plan, we fine-tune our models to a company's historical usage (rather than throw a standardized model at all of our users).

Happy to share a screencast! Feel free to email me if you'd like.

How do you compare to CloudHealth?

I do cost optimisations for living and though there are a lot of tools, most of them are niche and doesn't scale well. For generic purpose I tend to use CloudHealth, but still we write a lot of scripts to do some custom managment.

AFAIK, CloudHealth is mostly about cost management and analytics. Usage takes it a step further and also shows recommendations.

I would be interested in hearing about your experience with cost optimization. My email is kaveh@usage.ai if you'd like to reach out.

This is very much needed. I started to create something similar a few years back in case it's helpful. https://github.com/crizCraig/benjamin/blob/master/benjamin.p... Mostly it dealt with recommending reservations.

I like this. What a great idea for a consultancy. There must be so many organisations that are paying way more than they need by not tweaking each aws service they are using. Anything from using the wrong S3 storage type to using ECS when they could switch to Fargate.

You're exactly right. We've even seen very small pre-seed companies (with just a handful of EC2s) throwing away a few thousand dollars annually in extra charges to AWS. The problem becomes even bigger at large companies.

If you could somehow automate this and provide a tool that did analysis of a companies aws stack usage over a month and suggest configuration changes that would be a big thing.

That's what Usage does - we aren't a consultancy to be clear. We automate all of our recommendations (and occasionally add a human layer above it)

That reminds me early stage startup’s betting big on Kubernetes

as far as i can tell from the website this is mostly about converting workflows to RIs. what about right sizing or underutilized instances? do you need read permissions from cloudwatch to look at CPU/memory utilization or just the cost and usage report?

I work at a company that started from cost optimization but now offers a platform to manage the customers cloud operations and i only have 1 suggestion - be very upfront about the permissions/data needed for Usage to do it's job. you don't want to start a demo with a big/enterprise client only to find out your permission needs are way beyond what is allowed in that company for outside services

Hi Thristle,

We're more than just RIs. We also offer Savings Plans, Rightsizing, and Spot Fleet management.

Our permissions include CPU, network, and memory logs. We can scale permissions back on a customer-by-customer basis, though (as we have in the past).

Think this is a growing problem.

Related:

https://segment.com/blog/the-10m-engineering-problem/

Looks like the calculator needs a little tweaking!

https://i.imgur.com/Lp8B62n.png

Oops, thanks!

Check out cloudchecker (spelled it wrong purposely, the 2nd e should be removed) it may give you insight for pricing and features.

For enterprises (Fortune 500), Tangoe does this and integrates with your accounts payable. Even does the payment processing.

That’s not the experience I have with AWS. They are really trying to help you get the most out of their platform - including costs.

I’m not saying their pricing is not confusing, but the account managers and sales engineers I’ve worked with always did a good job navigating us through it - and being honest when something can be achieved a lot cheaper.

Better have a customer optimize their usage, pay less, be happy, and stick with AWS than switching to a different platform.

There are so many cost optimization (recommendation or actual) companies and as far as i know none were bought by aws. Azure bought 1 company and used that to allow their cost management to connect and show AWS cost information. other then that, nothing.

Kinda like how Amazon acquired CloudEndure, improved the service, and lowered cost by 80%? They really broke that. /s

Full disclosure: I work at AWS. Not on CloudEndure or anything relating to pricing. But come on.

AWS != Oracle