If your tryout automation team ’ s directive is to automate X amount of tests, and you have no strategy as to which examination they should focus on, you are wasting your clip. Before you begin writing your initiatory line of mechanisation code, make sure you have a strategy in place. Otherwise, you will have a ton of unable tests to maintain.

Don ’ t Choose a Random Goal

How many times hold you been told that the goal of the team is to get X amount of test reporting? This is an arbitrary value picked out of the sky. What is it establish on? If a UI automation team be to cover 80 % of the floor in a dash, they would never get perform in clip. We all know how flimsy UI automation is! How many times will a designer do a change that directly involve the UI and breaks the test? This is almost manageable during a sprint while you are work closely together, but how about when the product is sent to be translated to another words? The translator ineluctably arrive back with suggestions to allow for phrases more mutual and convertible. Bugs might be entered and UI changes make by a maintenance team with no psyche up to the mechanisation team, and Bam! — You have broken tests that need to be enquire.

Not but is a random goal meaningless, it ’ s hard to dog, and a pain when you have to shew you ’ ve met your finish. A test automation strategy should aim for the spirit of a end, but should not be stringently enforced to a value.

The Old ‘ 80-20 ’ Rule Strategy

Most of you have heard about thePareto Principle. In package, per Lowell Arthur, “ 20 percent of codification has 80 percent of the related errors. ” Dr. Joseph Juran, a long clip quality evangelist, recoined this to, “ the vital few and the trivial many ”, which he later changed to the “ useful many ” so as not to disregard the continue 20 percentage. A QA squad that needs to optimize their ability to discover glitch with the fewest amount of resources should take heed to this. The more you automate, the more you have to maintain. If you ’ ve ever automated, you know that is the true bear. Use your bug tracking system to show trends in bugs over X amount of time. You should be capable to name the hazardous area of your codification. Keep digging and you might find the few country that get most of your bugs. For the 80-20 Strategy, this is where you should focus firstly.

The “ Focus on Customer-Reported Bugs ” Strategy

Most companies study automation already have a manual procedure in property. It ’ s belike pretty efficacious, or the company wouldn ’ t withal be in job. This means they are likewise effectively tracking both internally and externally client-reported bugs. How many clip experience you heard something has been fix, and later it resort? Is that not frustrating? Customers can accept that bug happen, but to feature a bug arrive back really makes them question your quality standard. A scheme to focus on customer-reported bugs will prevent this gaffe. This scheme is elementary. Every customer-reported bug MUST receive automation establish around the fix. If you did a proper root effort analysis you should cognize the good spot to apply the automation. Every client-reported bug that is repair should have a process in spot to control this occurs:

For autonomous testing across multiple user personas, check out SUSATest — it explores your app like 10 different real users.

• Have a substance to identify this type of bug in the bug tag.

• Include the Root Cause Analysis resolution in the ticket, plus the proposed prevention solution.

• Based on where the automated test is enforced, include validation of this in the peer review checklist.

• Monitor for next related issues.

This scheme is simple, and should get for a good starting point for an mechanisation effort just starting out. But it is so powerful that it should be constituent of every feat.

The CRUD Strategy

This is an extremely simple scheme to implement. You don ’ t still have to investigate to start. My teams have ground this elementary strategy to be the most efficacious in get bugs early during CI. The CRUD (Create, Read, Update, Delete) scheme is basically what is suppose. Write simple tests that cover each of these basic functionalities. Applying this strategy will probably cover 80 % of all customer activities — bonus! Plus, while it might be redundant, it can be applied across all stage of testing, and is probably the easiest thing to automate. This will allow your mechanization teams to become familiar with the infrastructure while get a bombastic bang for the clam.

Always Be Reviewing

I ’ ve always loved the flick Glengarry Glen Ross, and the “ Always Be Closing ” prospect. You should be applying a similar principle to your automation strategy:

Make it a point to review the effectivity of your strategy. Have you noticed a downward trend in bugs? Have the tests found any bugs? (Make sure you identify bugs found by your automation.) What percentage? Don ’ t be afraid to discard ineffective tests. They just eat up resources. Always be critique!Joe Nolan (@ JoeSolobx) is a Mobile QA Team Manager with over 12 days of experience lead multi-nationally located QA teams, and is the beginner of the DC Software QA and Testing Meetup.