How do A/B tests work?

**In a nutshell: **A/B testing is all about studying causality by creating believable clones — two identical items (or, more typically, two statistically identical groups) — and then seeing the effects of treating them differently.

Image for post

When I say two identical items, I mean even more identical than this. The key is to find “believable clones” … or let randomization plus large sample sizes create them for you. Image: SOURCE.

Scientific, controlled experiments are incredible tools; they give you permission to talk about what causes what. Without them, all you have is correlation, which is often unhelpful for decision-making.

Experiments are your license to use the word “because” in polite conversation.

Unfortunately, it’s fairly common to see folks deluding themselves about the quality of their inferences, claiming the benefits of scientific experimentation without having done a proper experiment. When there’s uncertainty, what you’re doing doesn’t count as an experiment unless all three of these components are present:

  • Different treatments applied
  • Treatments randomly assigned
  • **Scientific hypothesis tested **(see my explanation here)

If you need a refresher on this topic and the logic around it, check out my article Are you guilty of using the word “experiment” incorrectly?

Why do experiments work?

To understand why experiments work as tools for making inferences about cause-and-effect, take a look at the logic behind one of the simplest experiments you can do: the A/B test.

Short explanation

If you don’t feel like reading a detailed example, take a look at this GIF and then skip to the final section (“The secret sauce is randomization”):

Image for post

Long explanation

If you prefer a thorough example, I’ve got you covered.

Imagine that your company has had a grey logo for a few years. Now that all your competitors have grey logos too (imitation is the sincerest form of flattery), your execs insist on rebranding to a brighter color… but which one?

Image for post

The logo your users see is grey, but that’s about to change.

After a careful assessment of what’s practical given your company’s website color scheme, your design team identifies the only two feasible candidates: blue and orange.

The CEO’s favorite color is blue, so she picks approving blue as the default action. In other words, she’s saying that if there’s no further information, she’s happy to err on the side of blue. Luckily for you, she’s a strong data-driven leader who is willing to allow data to change her mind to orange.

In order to switch to the alternative action of approving an orange logo, the CEO requires evidence that an orange logo causes your current user population to click more (relative to blue) on specific parts of your website.

You’re the senior data scientist at your company, so your ears prick up. You immediately identify that your CEO’s approach to decision-making fits the framework from frequentist statistics_. _After listening to her carefully, you confirm that her null and alternative hypotheses have to do with matters of cause-and-effect. That means you need to do an experiment! Summarizing what she tells you:

Default action: Approve blue logo.

Alternative action: Approve orange logo.

Null hypothesis: Orange logo does not cause at least 10% more clicking than blue logo.

Alternative hypothesis: Orange logo does cause at least 10% more clicking than blue logo.

#statistics #technology #towards-data-science #data-science #analytics #data analytic

What is GEEK

Buddha Community

How do A/B tests work?

The Ultimate Guide to Multiclass A/B Testing

One essential skill that certainly useful for any data analytics professional to comprehend is the ability to perform an A/B testing and gather conclusions accordingly.

Before we proceed further, it might be useful to have a quick refresher on the definition of A/B testing in the first place. As the name suggests, we can think of A/B testing as the act of testing two alternatives, A and B, and use the test result to choose which alternative is superior to the other. For convenience, let’s call this type of A/B testing as the binary A/B testing.

Despite its name, A/B testing in fact can be made more general, i.e. to include more than two alternatives/classes to be tested. To name a few, analyzing click-through rate (CTR) from a multisegment digital campaign and redemption rate of various tiers of promos are two nice examples of such multiclass A/B testing.

The difference in the number of classes involved between binary and multiclass A / B testing also results in a slight difference in the statistical methods used to draw conclusions from them. While in binary testings one would straightforwardly use a simple t-test, it turns out that an additional (preliminary) step is needed for their multiclass counterparts.

In this post, I will give one possible strategy to deal with (gather conclusions from) multiclass A/B testings. I will demonstrate the step-by-step process through a concrete example so you can follow along. Are you ready?

#hypothesis-testing #a-b-testing #click-through-rate #t-test #chi-square-test #testing

Tamia  Walter

Tamia Walter

1596754901

Testing Microservices Applications

The shift towards microservices and modular applications makes testing more important and more challenging at the same time. You have to make sure that the microservices running in containers perform well and as intended, but you can no longer rely on conventional testing strategies to get the job done.

This is where new testing approaches are needed. Testing your microservices applications require the right approach, a suitable set of tools, and immense attention to details. This article will guide you through the process of testing your microservices and talk about the challenges you will have to overcome along the way. Let’s get started, shall we?

A Brave New World

Traditionally, testing a monolith application meant configuring a test environment and setting up all of the application components in a way that matched the production environment. It took time to set up the testing environment, and there were a lot of complexities around the process.

Testing also requires the application to run in full. It is not possible to test monolith apps on a per-component basis, mainly because there is usually a base code that ties everything together, and the app is designed to run as a complete app to work properly.

Microservices running in containers offer one particular advantage: universal compatibility. You don’t have to match the testing environment with the deployment architecture exactly, and you can get away with testing individual components rather than the full app in some situations.

Of course, you will have to embrace the new cloud-native approach across the pipeline. Rather than creating critical dependencies between microservices, you need to treat each one as a semi-independent module.

The only monolith or centralized portion of the application is the database, but this too is an easy challenge to overcome. As long as you have a persistent database running on your test environment, you can perform tests at any time.

Keep in mind that there are additional things to focus on when testing microservices.

  • Microservices rely on network communications to talk to each other, so network reliability and requirements must be part of the testing.
  • Automation and infrastructure elements are now added as codes, and you have to make sure that they also run properly when microservices are pushed through the pipeline
  • While containerization is universal, you still have to pay attention to specific dependencies and create a testing strategy that allows for those dependencies to be included

Test containers are the method of choice for many developers. Unlike monolith apps, which lets you use stubs and mocks for testing, microservices need to be tested in test containers. Many CI/CD pipelines actually integrate production microservices as part of the testing process.

Contract Testing as an Approach

As mentioned before, there are many ways to test microservices effectively, but the one approach that developers now use reliably is contract testing. Loosely coupled microservices can be tested in an effective and efficient way using contract testing, mainly because this testing approach focuses on contracts; in other words, it focuses on how components or microservices communicate with each other.

Syntax and semantics construct how components communicate with each other. By defining syntax and semantics in a standardized way and testing microservices based on their ability to generate the right message formats and meet behavioral expectations, you can rest assured knowing that the microservices will behave as intended when deployed.

Ways to Test Microservices

It is easy to fall into the trap of making testing microservices complicated, but there are ways to avoid this problem. Testing microservices doesn’t have to be complicated at all when you have the right strategy in place.

There are several ways to test microservices too, including:

  • Unit testing: Which allows developers to test microservices in a granular way. It doesn’t limit testing to individual microservices, but rather allows developers to take a more granular approach such as testing individual features or runtimes.
  • Integration testing: Which handles the testing of microservices in an interactive way. Microservices still need to work with each other when they are deployed, and integration testing is a key process in making sure that they do.
  • End-to-end testing: Which⁠—as the name suggests⁠—tests microservices as a complete app. This type of testing enables the testing of features, UI, communications, and other components that construct the app.

What’s important to note is the fact that these testing approaches allow for asynchronous testing. After all, asynchronous development is what makes developing microservices very appealing in the first place. By allowing for asynchronous testing, you can also make sure that components or microservices can be updated independently to one another.

#blog #microservices #testing #caylent #contract testing #end-to-end testing #hoverfly #integration testing #microservices #microservices architecture #pact #testing #unit testing #vagrant #vcr

Software Testing 101: Regression Tests, Unit Tests, Integration Tests

Automation and segregation can help you build better software
If you write automated tests and deliver them to the customer, he can make sure the software is working properly. And, at the end of the day, he paid for it.

Ok. We can segregate or separate the tests according to some criteria. For example, “white box” tests are used to measure the internal quality of the software, in addition to the expected results. They are very useful to know the percentage of lines of code executed, the cyclomatic complexity and several other software metrics. Unit tests are white box tests.

#testing #software testing #regression tests #unit tests #integration tests

Dejah  Reinger

Dejah Reinger

1599859380

How to Do API Testing?

Nowadays API testing is an integral part of testing. There are a lot of tools like postman, insomnia, etc. There are many articles that ask what is API, What is API testing, but the problem is How to do API testing? What I need to validate.

Note: In this article, I am going to use postman assertions for all the examples since it is the most popular tool. But this article is not intended only for the postman tool.

Let’s directly jump to the topic.

Let’s consider you have an API endpoint example http://dzone.com/getuserDetails/{{username}} when you send the get request to that URL it returns the JSON response.

My API endpoint is http://dzone.com/getuserDetails/{{username}}

The response is in JSON format like below

JSON

{
  "jobTitle": "string",
  "userid": "string",
  "phoneNumber": "string",
  "password": "string",
  "email": "user@example.com",
  "firstName": "string",
  "lastName": "string",
  "userName": "string",
  "country": "string",
  "region": "string",
  "city": "string",
  "department": "string",
  "userType": 0
}

In the JSON we can see there are properties and associated values.

Now, For example, if we need details of the user with the username ‘ganeshhegde’ we need to send a **GET **request to **http://dzone.com/getuserDetails/ganeshhegde **

dzone.com

Now there are two scenarios.

1. Valid Usecase: User is available in the database and it returns user details with status code 200

2. Invalid Usecase: User is Unavailable/Invalid user in this case it returns status with code 404 with not found message.

#tutorial #performance #api #test automation #api testing #testing and qa #application programming interface #testing as a service #testing tutorial #api test

Aurelie  Block

Aurelie Block

1598916060

Top 10 Automation Testing Tools: 2020 Edition

The demand for delivering quality software faster — or “Quality at Speed” — requires organizations to search for solutions in Agile, continuous integration (CI), and DevOps methodologies. Test automation is an essential part of these aspects. The latest World Quality Report 2018–2019 suggests that test automation is the biggest bottleneck to deliver “Quality at Speed,” as it is an enabler of successful Agile and DevOps adoption.

Test automation cannot be realized without good tools; as they determine how automation is performed and whether the benefits of automation can be delivered. Test automation tools is a crucial component in the DevOps toolchain. The current test automation trends have increased in applying artificial intelligence and machine learning (AI/ML) to offer advanced capabilities for test optimization, intelligent test generation, execution, and reporting. It will be worthwhile to understand which tools are best poised to take advantage of these trends.****

#automation-testing #automation-testing-tools #testing #testing-tools #selenium #open-source #test-automation #automated-testing