Trade development tempo for hard costs
The previous post in the series suggested considering new high-value objectives for challenging conditions. This post suggests other methods that will help you can increase the value of a BAD (Big Data/AI/Data Science) project.
Teams can spend the additional effort to schedule and organize project work to be more efficient. They can operate to minimize cloud-based development costs for tests of data pipeline functions, data model migrations, configuration freezes/tests, and storage consumption.
Another way that project teams can economize is to reconsider full training passes in favor of spot training/testing to lower overall costs (even if just weekly instead of multiple cycles per week).
Incremental deployment can be slowed for better testing, reduced training overhead, and to focus on current high-value functions or user stories. This raises the demands for active project management and lowers “build/integrate/cycle” instances to lower monthly cash expenses.
The focus on testing and training also pushes work to improve quality rather than number of build cycles (results versus activities). Teams developing on the cloud must manage their costs to conserve budget for the project.
Tomorrow, we'll consider how to lower technical project overheads and increase function delivery rates.
You'll find that post at www.ekalore.com/bad-project-blog
Comments