How Unit Economics Can Improve Software Company Margins

Here’s an engineering perspective and outline of how this entertainment services and software company implements FinOps unit economics.

In an attempt to improve profit margins by reducing cloud costs as a % of revenue, FinOps practitioners (at the request of senior leadership) began measuring the following unit costs:

1. Cost per compute core
2. Cost per GB of storage
3. Cost per 1000 streaming * hours.

Background

Using the initial unit costs as a baseline, the FinOps team embarked on a program to improve cloud resource rates (unit prices) and cloud resource consumption/usage.

Data Sources

AWS & GCP billing data were used as the top-line (numerator) for the unit costs. The following metrics were used as the denominator for the unit cost metrics:

1. # of compute cores
2. Volume of data stored
3. Volume of streaming-hours.

Maturity Levels

Prior to this initiative, our company’s Unit Economics maturity level was Pre-crawl and now we consider ourselves to be running with unit economics due to the self-serve ability and awareness our engineers have on unit cost metrics as well as the automation in place to calculate the metrics.

Executive Sponsorship

Executive sponsorship was critical to the success of this initiative. The CFO, CTO, Director of Cloud CoE and engineering leaders were all involved in sponsoring the initiative to use unit economics to improve profit margins

Business Impact

This initiative took place over a 12-month period. It resulted in a 28% reduction in cloud costs per 1000 streaming-hours.

In addition to the reduction in cost per 1000 streaming hours, our engineers are now much more aware of cloud costs, and have self-service access to granular cost data. Engineers are able to self-serve the unit cost metrics in Slack.