Automation: When To Move Forward (Part Two)
Today’s technology affords enterprises the possibility of automating just about anything—and this may be just as much a curse as a blessing. Many companies, in their desire to compete, desperately want to view automation as that cure-all silver bullet that instantly levels the playing field. However, hastily adopting the “let’s automate everything in sight as fast as possible” approach can set companies back—negatively impacting cost and time to market.
Part One of this series examined when it makes sense for your enterprise to consider automation—but just because you can automate, should you? Part Two discusses when your business should pull the trigger on automation and tips to deliver the best results.
WHEN TO SAY WHEN
Mark Evans, in charge of Quality Assurance and Automation divisions at Ascendum Technology Services, emphasizes the importance of the time, quality and money triangle in the decision to automate. The balancing act of managing cost, maintaining quality, and maximizing speed weigh heavily in determining when to move forward with automation.
Mark identifies a key tipping point up front. “When business costs in the testing space become almost equal to development spend, it’s appropriate to prepare for a change because we are spending too much money and time testing something that should already work.”
Also, Subject Matter Experts (SME) in testing play a crucial role in identifying when to proceed with automation. Mark continues to explain the importance of this role. “An organization’s SMEs must determine when to execute proper automation scripts to deliver that perfect balance of time, quality, and money. Automation applied in conjunction with the correct expertise delivers the magic triumvirate—assured quality, faster or equal time to production, and a tremendous reduction in cost.”
TIPS FOR AUTOMATION SUCCESS
It’s not enough to know when—prosperous enterprises must understand what to automate and the best practices for success. The following suggestions increase the odds your automation program will deliver the desired results.
Automate Smoke Tests- Blow some smoke across all your systems by running simple automated scripts to determine their connectivity and physical function.
Automate Redundant Tasks- Run expanded smoke tests across all reoccurring actions for function by releasing code to confirm connectivity and that the entire system is navigable.
Conserve Intellectual Capital- Seek to automate items that require zero intellectual capital.
Aim for 80% Regression- This is the standardized target to deliver automation that can be executed automatically every hour, day and week. By hitting this 80% target, enterprises are ensuring the function of all the tasks necessary to stay in business all day every day.
Target Maintenance- Maintenance should be targeted to current scripting rather than everything every day. Best practices tie these scripts into what is already built and running—not creating for something likely to run only once such as a maintenance request.
Specialize SMEs- Remove your SMEs from toiling in the mundane and allow them to focus their attention exclusively on critical path items to maximize their value.
Target the Repeatable- Seek to automate all repeatable tasks with multiple scenarios.
WHEN AUTOMATION SUCCEEDED
Case Study One: Calculated Automation with Large Global Bank
- Automated 80% regression of bankcard applications
- Eliminated 60% of execution costs in the first year
- Allowed the presentation of greater scope for each release
- Automation removed the bottleneck of backed up projects
Case Study Two: Smoke Test Automation with Large Mortgage Lender
- Subject generated suite of tests ensuring all systems functional and transactional items in working order
- Tests run every four hours to ensure up time
- Reduced detection and fix times for predictable environment in testing and production
WHEN AUTOMATION FAILED
Case Study: Automated Everything for Retailer
- The retailer wanted a Big Bang approach which took nine months and resulted in a three to one cost ratio to convert all scripts and build new ones
- The retailer spent at one-to-one cost ratio to maintain scripts post delivery
- Overcome by a change in the system after the second year, the cost became two-to-one that of manual delivery
- The need to rework the project to a specified scope added a three-to-one cost
- Overall cost for “fatal decision” and revision was close to five-to-one the original spend with a negatively impacted time to market
NOT A SILVER BULLET
Despite the many advantages automation can provide, it can also create a significant backslide when deployed improperly. From the retail example above, we can see blindly automating in all directions is a poor choice—resulting in higher costs and slower time to market.
As an enterprise preparing to deploy automation take note—begin with your redundant, repeatable tasks that require minimal intellectual capital. Take the time to diligently examine your balance of cost, quality and time to market—allowing your SMEs to determine when and where the proper execution of automation will deliver optimum results.
Just because automation may be an option—doesn’t mean it’s the right one. In the words of Mark Evans, “Leave silver bullets for shooting werewolves and automate where it makes sense when it makes sense, and the ROI is either faster time to market, reduction in cost, or some combination of both.”