Regression Testing of Web Applications

As the application code goes through changes, you have to ensure that the functionality offered by your application stays intact and that the functionality works equally well across all desktop browsers and mobile device browsers.

Each time a new release/version of your web application is deployed, a good best practice is running tests with a significantly higher test coverage in the regression testing of functionality, using usually a single desktop browser, and running a subset of that coverage across desktop and mobile device browsers. This best practice ensures that you are able to accomplish higher coverage with efficiency. On a single desktop browser (usually headless Chrome or Firefox), it is usually possible to run a single test with a wide variety of data inputs to cover the boundary conditions, special validation conditions, etc., but across a wide variety of desktop and mobile browsers you run the same test for a limited number of data combinations. This is because running all tests with all data combinations and conditions usually becomes time consuming and expensive but not yield a good return on such time and money investments.

An example will help this best practice be understood better. Let us take the functionality of 'product search' on an eCommerce Storefront. Users will typically search for specific product using exact product name (like "Nike Air Max Axis Men's Sneakers"), or a portion of the product name (like "Nike Air"), or on product SKU or product code (like "3111671"), or attributes of a product like size, color, etc., (like Men's Size 10 Shoes" or "Size 8 Wide Black Shoes") or search for a category/group of products (like "Athletic Shoes"). Sophisticated users may even search for products using smart/sometimes wildcard searches like "Nike Air % Men's Sneakers" or "Clearance" or "Free Shipping". It is important to cover as many variety of data combinations during regression testing, especially, if it is possible to do so economically from both money and time point of view. Automation But it is not important to run all tests for all data combinations across desktop and mobile browser platforms because problems, if any in functioning, usually are not specific to desktop and/or mobile platforms. May be it makes sense to only run a subset of the product search tests (like product name or id) and category in all browser platforms but not all.

Worksoft SaaS offers a specific Subscription Plan for running tests on Headless Chrome or Firefox called 'High Volume Testing'. This subscription plan offers a significantly higher concurrency and capacity than the 'Cross-Browser/Device Testing' Subscription Plan. This allows you to rapidly accomplish a significantly higher test coverage of regression functional tests by using the 'High Volume Testing' subscription. A subset of that high test coverage can then be run on 'Cross-Browser/Device Testing' subscription which offers lower capacity and concurrency (because it costs us a lot more provide the infrastructure to support such desktop and mobile device combinations).