Product Unit Cost
Overview
Usage-based business model is entirely built around unit economics. As a critical part of unit economics, unit cost greatly impacts the sustainability of the business model, such as deriving pricing strategy or sales motions. For some segments of SaaS industries, the benchmark for product pricing and sales contracts involve careful unit cost estimations. Examples of such segments are data & analytics, machine learning products, developer tools or software infrastructure.
Paigo provides real-time, granular and automatic unit cost measurement for SaaS business. On Paigo platform, SaaS business can understand the unit cost of serving customers and leverage aggregated usage-based cost analytics to drive business decisions.
Terminology
Cost / Unit Cost refers to the cost of SaaS business.
Pricing / Usage-based Pricing refers to the price SaaS business charges SaaS customers.
Cloud Pricing refers to the price cloud platform (such as AWS, Azure and GCP) curages SaaS businesses for using their infrastructure.
Fixed Cost refers to the cost that typically doesn't change as every new SaaS customer is onboarded to the SaaS product. For example, if a monitoring component is running 24x7 to monitor the status of SaaS backend, onboarding another new customer to the platform typically does not change the cost of running monitoring component. Fixed cost is sometimes called shared overhead.
Variable Cost refers to the cost that changes (usually increases) as every new SaaS customer is onboarded to the SaaS product. For example, if a new virtual machine is allocated to each new SaaS customer, the cost of running virtual machine is considered a kind of variable cost.
Unit cost is a kind of variable cost, because as a new customers onboarded, their incurred usage will increase the cost of SaaS business.
Defining Unit Economics
To understand the cost structure, it is useful to define an atomic usage unit for a particular dimension that SaaS business wants to charge customers for, then find the unit cost structure for the usage unit. For instance:
For customers' data storage, the usage unit can be GB-Hour, meaning storing 1GB of data volume for a SaaS customer for 1 hour.
For customers' compute time, the usage unit can be Minute, meaning running 1 Minute of CPU and compute machine for a SaaS customer.
There might be more than one way of defining unit usage for a particular Dimension or SaaS business, such as GB-Hour vs GB-Month for data storage. If possible, it is preferable to define more granular units because they can be aggregated to coarse units. And as pricing evolves, different units might be used to improve on pricing.
Note that unit cost includes the software cost incurred to run SaaS application, it doesn't include other operations cost such as engineer salary, office rental, develop tools etc.
Paigo offers a courtesy consulting service to define the unit economics of SaaS product and unit cost. Contact team@paigo.tech to find out more.
Methodology for Cost Calculation
There are several critical challenges to derive unit for SaaS.
Multi-tenant environments: It typically requires careful attribution of cost to customers in order to derive a meaningful unit cost. For example, at the infrastructure level, a Kubernetes cluster is serving multiple customers at the same time. The cost attribution to each customer, or even each tenant, has to be calculated correctly.
Differentiating variable cost vs fixed cost: The complexity of the underlying infrastructure of SaaS applications has made the boundary of variable cost and fixed cost hard to define. For example, a load balancer serving all customers might be considered variable costs or fixed costs depends on the nature of business, the pricing model and the infrastructure topology.
Complex cloud pricing and cost structure: the cloud pricing could involve various cost all combined and not in a linear way. For example, when serving SaaS customers with data snapshots on AWS, the cost depends on various input such as the data block that has changed from last the snapshot, the amount of time to keep the snapshot (such as 0.8 months), etc. All of these parameters could affect the unit cost.
In order to calculate unit costs that takes into account all of aforementioned challenges, Paigo samples unit cost at different times and apply aggregation algorithms (such as AVERAGE) to derive the unit cost in the context of a particular business model.
As an example, a business uses AWS EC2 instance to run their applications for SaaS customers. They use fleets of on-demand instances, reserved instances and spot instances as the working node for their backend. The provisioning and orchestration of different instance types is done by a third-party software, and is subject to the availability of instant families they desired and going market rate of spot instances. In order to derive the unit cost for serving EC2 instances to customers, Paigo would measure unit cost continuously at a high frequency, and calculated average unit cost of serving one instance for one hour to customers based on their infrastructure topology and customer usage pattern.
See the use cases of usage-based costs in the following chapters for more details.
Difference with Cloud Cost Optimization
There are many cloud cost optimization software in the market. They would scan and monitor the cloud cost together to recommend or even execute cost saving tactics. Examples would be releasing idle resources, purging inactive data or arbitraging with cheapter resources. Paigo's usage-based cost analytics differs from them in several ways:
Paigo takes into account the usage-based business model and SaaS customers' usage topology. Therefore, Paigo would identify variable cost / fixed cost, multi-tenant / single-tenant in the context of the SaaS business to calculate cost.
Paigo focuses on increase revenue by calculating usage-based cost. Cost optimization software focuses on reduce the cost.
Paigo takes into account cost optimization done by other software when calculating usage-based cost. For example, a cost optimization software may actively manage the EC2 fleets by buying/selling reserved instances in the secondary markets or spinning up spot instances to save SaaS business on the instance cost. When Paigo calculates usage-based cost, it takes into account all the activities performed by the cost optimization software for the granular unit cost calculation.
Last updated