Is your Google Cloud bill spiraling out of control? You’re not alone. Businesses waste up to 30% of their cloud spend due to inefficiencies, costing the industry billions annually. From cloud architects and DevOps teams to IT leaders and executives, reducing cloud costs without sacrificing performance has become a top priority. 

Whether you’re a startup scaling rapidly or a seasoned cloud engineer looking for efficiency gains, there are actionable strategies that can help you achieve significant savings. In this article, we’ll walk through the top 5 Google Cloud cost-saving strategies for 2024—giving you practical insights to maximize efficiency, trim expenses, and enhance your bottom line.

1. Right-Sizing Resources

Overspending in the cloud often results from over-provisioning resources, which happens when organizations allocate more capacity than they actually need. Right-sizing involves continuously adjusting your resources—like virtual machines (VMs), databases, and storage—so you only use and pay for what’s necessary. This is especially critical for organizations transitioning from on-premises infrastructure, where over-provisioning was often required to handle peak loads. In cloud environments, resources can scale dynamically based on demand, so failing to adjust to this flexibility can lead to significant overspending.

The goal of right-sizing is to align resource allocation with actual demand. Google Cloud Platform (GCP) makes this straightforward with built-in tools that allow you to scale resources efficiently. For example, if a VM is underutilized in terms of CPU or memory, it can be resized to a smaller instance type. Similarly, if a database is processing minimal amounts of data, you can switch to a smaller, less expensive instance without sacrificing performance.

By right-sizing your resources, you avoid paying for idle capacity, which is one of the largest contributors to wasted cloud spend. According to a report from Flexera,  27% of cloud spending is wasted due to underutilized resources and inefficient provisioning​.

Key Tools for Right-Sizing

Google Cloud offers tools like Google Cloud Recommender, which provides tailored recommendations based on your resource usage patterns. The Recommender can suggest resizing VMs, moving data to more cost-efficient storage classes, or even suggesting instances that can be shut down based on idle time. These recommendations are generated using machine learning models that analyze your actual usage and predict future needs. This means that, with minimal effort, you can continually optimize your environment to ensure it’s both cost-efficient and high-performing. You can learn more and access the tool here: Google Cloud Recommender.

Another useful tool is Cloud Cloud Operations Suite (Formerly Stackdriver), which helps track CPU, memory, and storage usage across your environment. With this data, you can quickly identify underutilized resources and take action to scale them down. For example, if a VM is consistently running at less than 20% CPU utilization, it may be a candidate for a smaller, cheaper instance type.

Finally, you should check out the Compute Engine Rightsizing Recommendations: Google Cloud’s Compute Engine offers built-in rightsizing features, which automatically suggest optimal machine types for your VMs to avoid over-provisioning. This feature can significantly reduce unnecessary spending on idle resources. You can explore more about rightsizing in Compute Engine here: Compute Engine Rightsizing.

Best Practices for Right-Sizing

1. Regular Audits: Conduct frequent reviews of your resource usage, especially after deploying new applications or services. Tools like Google Cloud’s Recommender will help you identify inefficient resource allocations, but manual audits will ensure you don’t overlook less obvious areas of waste.
2. Leverage Custom Machine Types: GCP allows you to create custom machine types where you can allocate just the right amount of CPU and memory for your workloads. This level of granularity helps avoid the over-commitment of resources common with predefined instance sizes.
3. Dynamic Scaling for Databases: Use Cloud SQL Autoscaling to automatically adjust the resources for your database instances based on current workloads. This ensures you’re not paying for more resources than necessary during low-demand periods.
4. Schedule Non-Critical Resources: If certain VMs or services are not needed during off-hours, you can schedule them to shut down and restart at specified times, avoiding unnecessary costs for idle resources. Google Cloud’s Compute Engine offers native scheduling tools for this purpose.

By right-sizing your resources and making use of GCP’s optimization tools, you can potentially save up to 40% of your cloud spend without sacrificing performance .

2. Autoscaling to Manage Traffic Fluctuations

In today’s digital landscape, traffic and demand for applications are often unpredictable, fluctuating based on factors such as time of day, user behavior, and specific events like marketing campaigns or holiday promotions. This is particularly true for eCommerce businesses, media platforms, and any customer-facing service that may experience sudden traffic spikes. To effectively manage these variations, cloud autoscaling is essential. It enables you to dynamically adjust your computing resources in real-time, ensuring that you have just the right amount of capacity to handle the workload—no more, no less.

Autoscaling in Google Cloud Platform (GCP) helps businesses avoid the high costs associated with over-provisioning during low-traffic periods while preventing performance degradation during peak times. This capability allows resources such as virtual machines (VMs), databases, and containerized services to scale up when demand increases and scale down when it drops, saving money and optimizing resource usage.

How Autoscaling Works
Google Cloud’s Compute Engine provides horizontal autoscaling for virtual machines, which automatically adds or removes VMs from an instance group based on metrics like CPU usage, memory usage, or custom metrics defined by the user.  Similarly, Google Kubernetes Engine (GKE) offers cluster autoscaling, adjusting the number of nodes in a Kubernetes cluster based on the resource requests of pods. 

For instance, if an eCommerce business launches a major promotional event, traffic to the website could surge by several hundred percent in a matter of minutes. Without autoscaling, they would need to provision servers to handle peak traffic at all times, resulting in wasted resources and excessive costs during normal operations. With autoscaling, however, they can start with a minimal number of resources and allow GCP to scale up automatically when traffic spikes, then scale back down as traffic returns to normal levels. This not only ensures performance reliability but also minimizes cloud costs. 

Google Cloud allows you to set specific autoscaling policies, including:

• CPU-based scaling: Automatically adds or removes instances based on CPU utilization.
• Request-based scaling: Adjusts the number of instances based on the volume of incoming requests.
• Custom metrics: Enables scaling based on any custom metrics relevant to your business.

This flexibility helps organizations tailor autoscaling to their unique needs, whether it’s handling web traffic, API requests, or database queries.

Cost-Saving Potential
By autoscaling, companies can achieve significant cost savings. As we said in the previous section, a Flexera study found that cloud users overspend by 27% due to idle or underutilized resources . Autoscaling addresses this issue by dynamically adjusting the number of instances and reducing the need for pre-allocated, often unused, capacity. Businesses no longer need to pay for over-provisioned infrastructure that’s sitting idle during low-traffic periods, as autoscaling ensures they only pay for the resources they actively use.

Real-World Example: Autoscaling for eCommerce
We have a real life example of an eCommerce business preparing for Black Friday. A typical Black Friday sale may cause a massive surge in traffic, with users flocking to the site in a short time window. Without autoscaling, the business would need to provision enough infrastructure to handle the maximum expected load, resulting in high costs during off-peak times when the infrastructure is not fully utilized.

With autoscaling, the infrastructure automatically adjusts. When traffic spikes, additional compute resources are provisioned to handle the load. When the sale ends and traffic decreases, resources are scaled down. This ensures that the company is not paying for unused resources when they’re not needed,  In this case study, we achieved an 85% reduction in server costs during off-peak times when compared to what was required during peak times. Moreover, autoscaling guarantees a consistent user experience, ensuring that the website remains responsive and performant during the peak sale period. We have a case study about this here.

Key Tools and Features

• Google Kubernetes Engine (GKE) Autoscaler: Automatically scales the number of nodes in a GKE cluster based on demand, ensuring efficient resource use.
• Compute Engine Autoscaler: Adds or removes VMs from a managed instance group based on CPU utilization, HTTP request load, or other custom metrics.
• Cloud Functions and Cloud Run: These serverless services automatically scale based on the number of incoming requests, so you only pay for the time your functions or containers are executing.

Best Practices for Implementing Autoscaling

1. Set Baseline Metrics: Define clear metrics for when to scale up or down. CPU usage and request rates are common triggers, but for more granular control, custom metrics can be used.
2. Use Cooldown Periods: Set cooldown periods to avoid frequent scaling actions that could result in instability. This helps ensure that scaling actions are based on sustained demand rather than temporary spikes.
3. Optimize Scaling Limits: Define minimum and maximum scaling thresholds to avoid over-scaling or under-scaling. These thresholds help control costs and ensure you have enough resources during peak traffic.

By leveraging autoscaling, businesses can significantly reduce costs, improve performance, and ensure seamless customer experiences during high-traffic events like product launches or holiday sales.

3. Committed Use Discounts

One of the most effective ways to reduce your Google Cloud costs is by taking advantage of Committed Use Discounts (CUDs). Google Cloud offers substantial savings for customers who commit to using specific services, such as virtual machines (VMs), databases, or GPUs, over a period of one or three years. In exchange for this commitment, businesses can reduce their cloud spend by as much as 57% compared to on-demand pricing.

Unlike traditional pay-as-you-go models, CUDs provide businesses with predictable pricing, making it easier to budget and plan cloud expenditures over time. These discounts are particularly beneficial for companies that have stable, long-term workloads, where the resource demands are consistent and can be forecasted.

How Committed Use Discounts Work
Google Cloud’s CUDs allow you to purchase a specific amount of resources (measured in vCPUs and memory for VMs, or in other relevant units for other services) at a discounted rate, committing to a set usage level over one or three years. In return, Google offers steep discounts over the equivalent on-demand pricing.

For example, a business running a 24/7 application on Compute Engine could commit to a specific number of vCPUs and memory for a three-year period. In return, Google Cloud reduces the cost of those resources, potentially saving the business tens of thousands of dollars over the course of the commitment.

According to a Flexera 2022 State of the Cloud Report, one of the top concerns for organizations using cloud services is cost control, with 69% of businesses identifying cloud cost management as a top priority . CUDs directly address this challenge by locking in predictable pricing and allowing businesses to avoid fluctuations in their cloud bills.

Key Benefits of Committed Use Discounts

1. Significant Cost Savings: The primary benefit of CUDs is the potential for substantial cost savings. By locking in rates for long-term usage, you can reduce your cloud spend by up to 57%. This makes CUDs one of the most effective strategies for controlling cloud costs, particularly for businesses with predictable workloads.
2. Predictable Budgeting: CUDs offer price stability, allowing businesses to forecast cloud expenses accurately. This predictability is crucial for long-term financial planning and helps avoid the variability often associated with on-demand pricing.
3. Flexibility in Resource Commitment: While committing to a fixed level of resource usage might sound rigid, Google Cloud offers some flexibility. For example, if you commit to using a certain number of vCPUs and memory, you can apply those resources across different machine types, regions, and operating systems within Compute Engine. This flexibility ensures that your CUDs can adapt to changing infrastructure needs without penalizing you for shifts in workload distribution.

How to Maximize Your Savings with CUDs
The key to maximizing savings through Committed Use Discounts lies in accurately predicting your future resource needs. Here’s how to get started:

1. Analyze Your Long-Term Resource Requirements: Start by reviewing historical usage data to identify which workloads are stable and predictable. These workloads are ideal candidates for CUDs. Services that run consistently, such as databases, batch processing jobs, or backend services for web applications, are perfect for commitment-based pricing.
   Example: If you know that your database instances will require a constant level of compute resources, you can commit to a specific amount of vCPUs and memory for three years and lock in the savings.
2. Use Google Cloud’s Pricing Calculator: Google Cloud provides a pricing calculator that helps estimate the savings you can achieve with CUDs. This tool allows you to input your expected resource needs and see how much you can save by committing to specific services. It’s essential to experiment with different scenarios, such as one-year versus three-year commitments, to find the most cost-effective option for your workloads.
   Pro Tip: Use historical billing data available in Google Cloud Console to forecast your long-term usage accurately. This data will give you insights into your average consumption over time, helping you determine the ideal level of commitment without over-provisioning.
3. Select the Right Commitment Period: While three-year commitments offer the highest discount, a one-year commitment might be more appropriate for workloads with shorter life cycles or where future needs are harder to predict. Assess the trade-off between the higher discount of a longer commitment and the flexibility of a shorter-term commitment based on your workload stability.
4. Monitor and Adjust: Even after committing to a specific level of usage, it’s essential to monitor your actual resource consumption. If you notice changes in your workloads, consider modifying your future commitments to match new requirements. Google Cloud provides detailed billing and usage reports to help track your CUDs and adjust them if needed for future commitments.

Ideal Use Cases for Committed Use Discounts
CUDs work best for predictable, long-running workloads. Common use cases include:

• Production Environments: Applications that run 24/7, such as databases, backend services, or APIs, are prime candidates for CUDs. Their usage patterns are stable, making it easier to forecast resource requirements.
• Batch Processing: If you run large batch jobs regularly and can predict the compute resources needed, CUDs can offer significant savings.
• Big Data Workloads: For businesses using services like BigQuery or Dataflow, committing to a predictable level of resource consumption over time can lead to considerable cost reductions.

By accurately predicting your resource needs and taking advantage of Committed Use Discounts, you can significantly reduce your Google Cloud spend, making your cloud operations more cost-efficient. This strategy is ideal for organizations with stable, non-variable workloads that are looking to optimize their cloud costs while maintaining performance.

4. Leverage Cloud-Native Solutions

Adopting cloud-native technologies is a powerful strategy for reducing Google Cloud costs while optimizing resource efficiency. Cloud-native solutions like Google Kubernetes Engine (GKE) and Cloud Functions allow you to focus on building and running applications without worrying about managing the underlying infrastructure. These services are designed to scale dynamically and automatically, driving significant cost savings by eliminating the need to maintain idle or over-provisioned resources.

What Are Cloud-Native Solutions?

Cloud-native solutions are applications designed specifically to run in cloud environments, leveraging the inherent flexibility and scalability of cloud platforms. Unlike traditional applications that often require fixed infrastructure, cloud-native technologies like Kubernetes and serverless functions automatically adapt to the demand of the application, ensuring that you only use the resources you need when you need them.

Google Cloud offers several cloud-native services, including:

• Google Kubernetes Engine (GKE): A managed Kubernetes service that enables you to deploy, manage, and scale containerized applications seamlessly.
• Cloud Functions: A serverless computing service that runs code in response to events and charges based on execution time.
• Cloud Run: Another serverless platform that runs stateless containers and automatically scales based on the number of requests.

By leveraging these services, businesses can dramatically improve efficiency and reduce costs.

How Kubernetes (GKE) Drives Cost Efficiency

Kubernetes, an open-source container orchestration platform, has become the go-to solution for deploying and managing applications in cloud environments. Google Kubernetes Engine (GKE), Google’s managed Kubernetes service, allows you to run containerized applications with dynamic scaling and self-healing capabilities.

One of the most significant cost-saving benefits of using Kubernetes on GKE is autoscaling. GKE supports cluster autoscaling, which automatically adjusts the number of nodes in your Kubernetes cluster based on the resource requirements of your applications. This ensures that you only run the infrastructure needed to meet current demand, preventing unnecessary over-provisioning.

Example: If your application typically requires 5 nodes but spikes to 20 nodes during peak times, GKE will automatically scale up to handle the increased demand and scale down once the traffic decreases. This prevents you from paying for 20 nodes during off-peak times, translating to substantial savings.

Additionally, Kubernetes allows you to take advantage of bin packing—the practice of running multiple workloads on a single node to maximize utilization. This feature ensures that nodes are used to their full capacity, reducing the need for additional nodes and lowering costs.

Cost Savings with Serverless Functions (Cloud Functions)

Serverless functions represent another cloud-native approach that offers significant cost benefits. Google Cloud Functions is a fully managed, event-driven compute service that allows developers to run code without provisioning or managing servers. The key advantage of serverless computing is that you only pay for the actual time your code is running, which makes it ideal for workloads with unpredictable or sporadic usage patterns.

For instance, instead of maintaining a dedicated virtual machine to run occasional batch jobs or API requests, you can deploy your code to Cloud Functions and pay only for the compute time your function actually uses. Cloud Functions can handle requests at scale, automatically scaling from zero to thousands of instances depending on traffic, and then scaling back down when demand decreases. This makes serverless a highly cost-effective solution for handling unpredictable or bursty workloads.

Example: An application that processes images or videos only when users upload files can use Cloud Functions to handle the processing. Instead of running a VM 24/7 to wait for uploads, Cloud Functions will automatically spin up and execute the processing code when an upload is detected, and you’ll only be billed for the time the function is executing.

Why Move to Cloud-Native Services?

Moving workloads to cloud-native services offers several advantages, including:

1. Automatic Scaling: Both Kubernetes and serverless platforms like Cloud Functions scale resources dynamically based on demand. Whether you’re handling fluctuating traffic or intermittent workloads, you only pay for the resources you actively use, minimizing waste.
2. Reduced Idle Capacity: Traditional infrastructure often involves maintaining idle or underutilized resources to accommodate peak loads. With cloud-native technologies, resources are provisioned and de-provisioned as needed, ensuring you never pay for idle capacity.
3. Improved Resource Utilization: Kubernetes uses advanced resource management techniques like bin packing to optimize how workloads are distributed across nodes, ensuring you get the most out of your infrastructure.
4. Lower Operational Overhead: By using managed services like GKE or Cloud Functions, businesses can focus on building and running applications rather than managing the underlying infrastructure. This reduces the need for specialized operations teams and lowers associated costs.
5. Efficient Handling of Microservices: Cloud-native platforms are ideal for deploying microservices, where different components of an application are deployed independently. Kubernetes and serverless platforms are designed to handle microservices efficiently, scaling each service independently based on its resource requirements.

Real-World Impact: Moving to Cloud-Native Solutions

Many organizations have seen significant cost savings after migrating to cloud-native technologies. A case study by Google highlights how Citrix saved 45% on infrastructure costs by migrating its workload to GKE and utilizing Kubernetes autoscaling features . By moving away from traditional VM-based infrastructure and embracing a cloud-native approach, Citrix was able to reduce waste and optimize usage based on demand.

Similarly, businesses leveraging serverless architectures for event-driven applications have seen reduced infrastructure costs, particularly for sporadic workloads. According to a study by Deloitte, companies that adopted serverless computing reduced their cloud infrastructure costs by up to 70% for specific workloads by avoiding idle resource costs .

Best Practices for Implementing Cloud-Native Solutions

1. Use Horizontal Pod Autoscaling (HPA): If you’re using Kubernetes, set up Horizontal Pod Autoscaling to dynamically adjust the number of pods in your deployment based on resource utilization metrics like CPU or memory usage. This ensures you scale up to meet demand but avoid paying for idle pods when demand is low.
2. Consider Function-as-a-Service (FaaS): For workloads that run intermittently or in response to events, consider moving to a serverless model with Cloud Functions or Cloud Run. Pay only for the time your code is running, which can be a significant cost saver.
3. Monitor Resource Utilization: Continuously monitor resource usage across your Kubernetes clusters or serverless functions. Use Google Cloud’s Cloud Operations Suite to ensure that your cloud-native services are optimized and to identify any underutilized resources that can be scaled down.

5. Optimize Storage for Infrequently Accessed Data

Data storage is a critical component of any cloud infrastructure, but it can also be one of the most significant contributors to your overall cloud costs if not managed properly. Many businesses inadvertently overspend on storage by keeping all of their data in expensive, high-access storage classes—even when much of that data is rarely accessed. To address this, Google Cloud offers different storage classes optimized for varying data access patterns, enabling you to lower your costs without sacrificing data availability when needed.

Google Cloud’s Nearline, Coldline, and Archive storage classes provide cost-effective solutions for storing infrequently accessed data. By migrating less-frequently used data to these lower-cost storage tiers, organizations can achieve significant cost savings while maintaining access to data when required. This strategy is particularly useful for businesses that handle large volumes of backup data, archival content, or logs that need to be stored long-term but are rarely accessed.

Understanding Google Cloud Storage Classes

Google Cloud offers several storage tiers, each designed for different access patterns and use cases:

1. Standard Storage: Ideal for data that is accessed frequently (hot data), such as content for websites or active databases. This storage class is optimized for low-latency, high-access scenarios and carries the highest cost.
2. Nearline Storage: Designed for data that is accessed less than once a month. Nearline is a good choice for backups or datasets that are rarely queried but still need to be readily accessible. Nearline storage costs significantly less than Standard Storage.
3. Coldline Storage: Best suited for data that is accessed less than once a year. Coldline storage is even cheaper than Nearline and is ideal for long-term backups, disaster recovery, and archival storage. Although the cost to store data is very low, retrieval costs are higher than in Nearline.
4. Archive Storage: The most cost-effective option for data that is rarely, if ever, accessed. Archive storage is perfect for data that must be retained for compliance or historical reasons but is not needed for day-to-day operations. While retrieval from Archive storage takes longer and is more expensive, the storage cost itself is minimal—up to 80% cheaper than Standard Storage.

How to Optimize Storage Costs

Optimizing storage costs requires understanding your data’s access patterns and choosing the most appropriate storage tier for each dataset. Here’s how you can maximize savings by leveraging Google Cloud’s storage classes:

1. Identify Infrequently Accessed Data: Begin by auditing your data to determine how frequently different datasets are accessed. This is where Google Cloud Storage Insights can help. Storage Insights provides detailed metrics on data access patterns and allows you to identify datasets that are rarely or never accessed.
2. Move Cold Data to Lower-Cost Tiers: Once you’ve identified infrequently accessed data, migrate it to more cost-effective storage classes like Nearline, Coldline, or Archive. For example, if you have large volumes of backup data or logs that are only accessed during audits or disaster recovery, Coldline or Archive storage will offer significant cost savings. Archive storage is up to 80% cheaper than Standard Storage, making it an excellent choice for long-term retention of infrequently used data.
   Example: A company that stores terabytes of monthly backup data on Standard Storage may find that these backups are rarely accessed. By transitioning this data to Coldline or Archive storage, they could reduce storage costs by 50-80%, while still retaining the ability to restore backups when needed.
3. Use Lifecycle Management Policies: Google Cloud allows you to set lifecycle management policies to automate the process of moving data between storage classes. For example, you can create a policy that automatically moves data from Standard to Nearline storage after 30 days of inactivity and then to Coldline or Archive storage after a year. This helps ensure that you aren’t overpaying for storage, even as data access patterns change over time.
   Pro Tip: Set up automated transitions to ensure that hot data is moved to cooler storage classes as it becomes less frequently accessed. This reduces the need for manual intervention and guarantees cost savings as your data ages.

Real-World Example: Optimizing Backup and Archive Storage

Imagine a media company that generates massive amounts of video content. While new videos are actively used for editing and distribution (requiring Standard or Nearline storage), older video files may only need to be stored for compliance purposes or long-term archival. By moving these older, infrequently accessed files to Coldline or Archive storage, the company can cut its storage costs by more than 50% without losing access to the files when needed.

Similarly, businesses that maintain daily or weekly backups can move older backup files to Coldline or Archive storage, ensuring they’re paying significantly less for data that is accessed rarely, if ever.  You can see a case study about this topic here.

Cost Comparison Between Storage Classes

Here’s a high-level comparison of Google Cloud’s storage classes and potential savings:

• Standard Storage: Best for high-access data, but the most expensive option.
• Nearline Storage: Costs around 50% less than Standard Storage, ideal for data accessed once a month or less.
• Coldline Storage: Costs up to 80% less than Standard Storage, suitable for data accessed once a year or less.
• Archive Storage: The least expensive option, offering savings of up to 80% compared to Standard Storage, but with higher retrieval costs and longer retrieval times.

Pricing for each of these storage options can be found here.

Best Practices for Optimizing Storage

1. Evaluate Data Access Patterns Regularly: Continuously assess how often your data is being accessed and adjust storage tiers accordingly. Data that was frequently accessed in the past might now be rarely used, making it a candidate for migration to Coldline or Archive storage.
2. Use Versioning and Retention Policies: If you store multiple versions of the same data (such as backups or logs), use versioning and retention policies to limit how long older versions are stored. This reduces the total amount of storage needed.
3. Monitor Retrieval Costs: While Nearline, Coldline, and Archive storage offer lower storage costs, they come with higher retrieval fees. Make sure to balance your retrieval needs with the cost savings to avoid unexpected charges when accessing archived data.

Optimizing your storage strategy by moving infrequently accessed data to Google Cloud’s lower-cost storage tiers can lead to substantial cost savings—up to 80% in some cases. Whether you’re managing backups, logs, or archival content, transitioning to Nearline, Coldline, or Archive storage ensures you’re not overpaying for data you rarely access while still maintaining the ability to retrieve it when necessary. With smart data management, automated lifecycle policies, and ongoing monitoring, businesses can significantly reduce their storage costs without sacrificing accessibility.  Check out Google Cloud’s best practices for storage for more guidance on how to implement efficient data storage strategies.

Conclusion

Managing cloud costs doesn’t have to be overwhelming. By implementing these five proven strategies, you can take control of your Google Cloud expenses and reinvest those savings into what matters most for your business.

But why navigate this complex landscape alone? Our team of Google Cloud experts is here to guide you every step of the way. If you are interested in getting help, you can contact us here.

Don’t miss out on this opportunity to optimize your cloud spend and boost your bottom line. Use the scheduling widget below and schedule our free consultation.

Take the first step toward significant savings—your future self will thank you.