Table Of Contents
Is Hourly Granularity Really Necessary? Isn’t Daily Granularity Good Enough? When Is Hourly Granularity Most Useful? Hourly Granularity: A Gold Standard In Cost Management

Cloud spenders of all sizes, but particularly those who have exceeded the $50,000 a month benchmark, know how challenging it can be to extract meaningful insights from their spend environments.

As a digital-native company grows, its spend environment naturally gets more complicated — and without a strong organizational framework, what could be a treasure trove of data turns into a rat’s nest of semi-organized, confounding figures.

Trying to extract business insights from a messy spend environment is a little bit like trying to find a drill bit in a disorganized garage. You’ll spend a lot of time sifting through uncorrelated, non-relevant items to find what you’re looking for.

Businesses looking to turn their cloud spend into a rich source of business insight must start with complete, real-time cost allocation. They need a way to allocate 100% of their cloud spend — including cloud, PaaS, and SaaS providers, as well as shared, multi-tenant, and untaggable costs — quickly, flexibly, and without relying on tags (which are impossible to perfect, and extremely difficult to maintain).

Moreover, businesses should expect to visualize their spend at a level of hourly granularity. “Granularity,” in this context, refers to the intervals of time at which you can parse your cloud spend. The vast majority of cloud cost management solutions max out at daily granularity; CloudZero is the only platform to offer hourly granularity across all spend sources, and in all platform views.

Most platforms don’t offer hourly granularity for one simple reason: It’s hard. The more cost data a company produces, the more difficult it gets to allocate it all — and often, organizational difficulty grows at an exponential, not linear, rate. Most platforms — especially incumbent solutions built early in the existence of the cloud — are not equipped to handle the difficulty associated with hourly granularity.

When we built CloudZero, we knew data granularity was of the essence for digital-native businesses. So, we handled the difficulty upfront, so that we could provide as much value to end-users throughout their engagement with CloudZero as possible.

In this blog, I dive deep into why hourly granularity matters, who needs it, and when it’s most useful.

Is Hourly Granularity Really Necessary? Isn’t Daily Granularity Good Enough?

Because daily granularity is as deep as most platforms go, cloud spend managers get in the habit of thinking of it as the industry standard — or, at least, good enough.

It’s certainly not the industry standard. And there are two main reasons why CloudZero believes it’s not good enough — and why ours was the first platform to present spend data at hourly granularity (even preempting AWS Cost Explorer):

1. The more data you have, the more insights you can find

The difference in data volume between daily granularity and hourly granularity is substantial — 24x, to be exact. If your cloud spend tends to spike at a particular time of day — from 3:00 PM to 5:00 PM, say — hourly granularity would reveal that and give your engineers clues about how to optimize. Daily granularity would omit it, showing a single cost representing the entire 24-hour period.

2. For engineers to own their cloud costs, they need relevant, granular data

CloudZero’s philosophy is that engineers belong at the center of cloud cost management. Engineers are the ones who build — and therefore, buy — cloud infrastructure, so the more cost-efficient their building decisions are, the more manageable the organization’s costs will be.

To be good cloud cost managers, engineers need near-real-time updates on the cost of their building decisions — and clear optimization paths. If an engineer sees a blanket daily cost that isn’t changing much, they won’t pay attention to it. But, if they see a recurring cost spike at a particular time of day, they’re going to investigate it, and, if possible, optimize it.

3. CloudZero users have reaped millions of dollars in savings from features using hourly granularity

The proof, as ever, is localized in the pudding —and hourly granularity regularly produces substantial savings opportunities for CloudZero customers. The CloudZero feature most directly tied to hourly granularity is Anomalies, which compares hourly data on your full spend environment from the past 72 hours to hourly data from the prior 90 days.

Through the first half of 2023, CloudZero has detected 5,558 anomalies, just four of which added up to nearly $1.2 million in annualized savings.


When Is Hourly Granularity Most Useful?

1. Breaking down Kubernetes costs

The chief benefit of Kubernetes is its ephemerality. When you need compute power, it spins resources up; when you don’t, it spins them down. This is great from a container orchestration standpoint, but can pose problems from a cost analysis standpoint.

Unless, that is, you have hourly granularity. If you have a few pods running for an hour when virtual machines (VMs) are most expensive, and fewer pods running later when VMs are less expensive, hourly granularity will show you exactly what your Kubernetes environment costs, and when. Daily granularity would smooth it out — and obscure potential cost optimization opportunities.

2. Understanding user activity

Users access your system at particular points in time. But if you spread that user data only to a single day, it’s impossible to understand how specific user activity correlates with your spend.

For example, it’s possible that a majority of your users log into your platform outside of business hours —likely the case for an entertainment platform like Netflix. Or, maybe you get a big surge of users at 2:00 AM — night owl musicians logging into Pro Tools to do some after-midnight audio mixing, say.

Daily granularity would spread this activity out across the whole day, leading you to provision cloud infrastructure in such a way that it ran all day at an average rate. But hourly granularity would allow you to gauge IT capacity to common user activity, and spend only what you need to spend to support your users.

3. Detecting and remediating spend anomalies

AI and machine learning (ML) are steadily edging their way into the cloud. On the cost management side, one of the most powerful early applications of AI and ML is in anomaly detection: Identifying unusual spend activity and notifying the relevant engineers.

CloudZero anomaly detection compares hourly spend data from the past 72 hours to hourly spend data from the previous 90 days. It is thus constantly training and retraining itself on what “normal” looks like, at an unparalleled level of precision, and getting ever more adept at singling out aberrant spend events.

Moreover, when it finds something unusual, it sends an automatic alert directly to the engineer responsible for the affected cloud infrastructure. This makes it easy for the engineer to learn about the problem, investigate it, and decide whether something should change.

Because daily granularity smooths out intra-day cloud spend spikes, it will miss at least some spend anomalies that an hourly system catches.

4. Engaging engineers in cloud cost management

In their yearly surveys, the FinOps Foundation regularly reports engineering engagement as the biggest challenge organizations have in instituting FinOps programs. In short, it’s hard to get engineers in the habit of managing their own costs.

But why is that? It’s not because they don’t want to — 60.6% of the FinOps Foundation’s members report to a CTO, meaning they’re technical people who want to do a better job managing their costs. When engineering engagement flags, it’s because they don’t have relevant, timely cost data that’s easy to explore and act on.

Hourly granularity means cost insights will be targeted, actionable, and rewarding for engineers to engage with. The more useful cost insights are to an engineer’s daily work —and the more positive reinforcement they get, via confirmed cost savings — the more likely they are to stay engaged over time.

Hourly Granularity: A Gold Standard In Cost Management

In summary, hourly granularity gives you:

  • Data volume: 24x the data (compared to daily granularity)
  • Insight volume: A higher number of more targeted insights and savings opportunities
  • Engineering engagement: A positive feedback loop that increases engineering engagement in cost management

But the story goes beyond hourly granularity. AWS Cost Explorer presents its data at a level of hourly granularity, but it still doesn’t provide category-leading power in cloud cost management.

For that, cloud users need a platform that offers complete, real-time cost allocation — and that doesn’t rely on tagging quality to do it. That means ingesting all of your cloud PaaS, and SaaS spend, allocating 100% of it —untaggable and shared costs included — in a framework that mirrors your business, and presenting it in a dynamically explorable view that makes sense to cloud engineers.

Engineers need relevant, timely data, displayed in a context that mirrors the structure of your business. They need the ability to drill into that data, seeing its trends at an hourly level, and understanding what it means for cloud infrastructure needs. With the right data at the right time, presented at a level of hourly granularity, they can take cost accountability into their own hands.

For all this and more, CloudZero is far and away the best option for digital-native businesses. CloudZero ingests and normalizes 100% of your cloud, SaaS, and PaaS spend in a single pane of glass, allocates it in real-time without any dependence on tagging quality, and makes it dynamically explorable in CloudZero Explorer.

This way, you know you’re getting insights from your entire cloud environment — and seeing it in a view that engineers love.

Our team would love nothing more than to show you our platform in action. .