Whether it’s lowering costs, improving quality, accelerating time-to-market, or increasing test capacity, lots of people think test automation is the answer to all of their problems. And while it most certainly can be, it’s not a magic bullet, either. But provided you’re smart, you’ll see an ROI in next to no time. Here are some tips to help you get there.
Photo courtesy of Krstnn Hrmnsn (Creative Commons)
| If you’re just starting to look into automation of any kind, it’s worth bearing this in mind: you can’t automate everything. Nor should you—there’s just too much nuance involved in many types of testing for automation to ever be a viable option. |
You probably don’t want to hear this, but if you’re expecting Oracle Application Testing Suite (Oracle OATS) or other automated testing tools like HP QTP or Selenium to be the magic-bullet answer to all your ERP testing woes, you’re going to be disappointed.
While the value that automation can bring enterprise-level companies is undisputed (especially if they have multiple releases per year), it’s not as simple as buying a license, hiring a vendor, and raking in the savings.
Granted, we all wish it was; but that just isn’t true.
If you’re just starting to look into automation of any kind, it’s worth bearing this in mind: you can’t automate everything. Nor should you—there’s just too much nuance involved in many types of testing for automation to ever be a viable option. In those cases, it’s far better to leave that kind of testing to your more-than-capable SMEs.
But when it comes to other, more repetitive forms of testing, though, the sky is the limit—even though that’s not without its own drawbacks.
Once you start isolating the kinds of tests you want to automate (static and repetitive tests and tests that need to be run for every release/upgrade are your best bets), you’ll probably realize another fundamental truth of automation: people never have a hard time finding good candidates for test automation. In fact, most tend to find so many that they can’t figure out where to even start! For a lot of companies, this can lead to “paralysis by analysis” as managers struggle to define the scope of their project.
When that happens, it’s best to talk to the folks on the front lines—the SMEs that know your systems best. As masters of their own domains, SMEs can pinpoint testing pains and illustrate exactly where automation can be most effective. Remember: It’s not uncommon for companies to spend several weeks or months on acceptance test cycles while still failing achieve a 95% Defect Removal Efficiency (DRE); automation can reduce that time considerably and enhace DRE at the same time.
At the same time, SMEs don’t merely help define your project’s scope; they help ensure that the most-executed and often-repeated tests are automated first. These “low hanging fruit” test cases typically tend to have the largest time/effort disparity between manual and automated executions, ensuring the highest possible return and making them an invaluable asset in the pilot-sized, “proof-of-concept” engagements that so often precede full-scale investments. Succeed there and you’re home free.
Sure, it might not be a magic bullet; but, with a little strategy on your part, it can most certainly be a profitable one. And when it comes to any investments, isn’t that the most important thing?
Cheers,
Mike Hodge
Lighthouse Technologies, Inc
Software Testing | Quality Assurance Consulting | Oracle EBS Consulting
