This post serves two purposes. It is yet another installation in The Series That Ate My Blog; and it’s a kind of personal exploration of work in progress on the Rapid Software Testing Guide to Test Reporting. Your feedback and questions on this post will help to inform the second project, so I welcome your comments.
As a tester, your mission is to evaluate the product and report on its status, typically with a special emphasis on finding problems that matter. We’ve discussed bug reporting in the Rapid Testing Guide to Making Good Bug Reports. In this installment of Breaking the Test Case Addiction, I’m describing test reporting as something that responsible testers do.
Sounds straightforward, right? But right away, I want to address the risk of misunderstanding, so let me clear up what I mean by certain terms here.
Responsible testers are people who assume the role of tester on a project, and who commit themselves to doing that job well over time. Supporting testers (which we used to call “helpers”) help the test effort temporarily or intermittently, but are not committed to the testing role. Supporting testers are generally not required to report on their testing work to the same degree as responsible testers are.
In this post, when I say test project, I’m referring to any set of activities focused on testing of any product or service, or any part of it: a low-level unit, a function, a component, a feature, a story, a service, an entire system… A test project can contain lots of little test projects. Accordingly, depending on the level of granularity we’re referring to, a test project might happen over moments or minutes, days, weeks, or months. A report on a test project might cover similar spans of time—instants, episodes, sprints, releases…
“Test project” here could refer to something that happens outside of development. More typically, it refers to testing activity that happens inside a development project, in parallel with the other aspects of development, like design, programming, or other testing.
When I say product here, I mean anything that anyone has produced that might be subject to testing. While that includes running code, “product” could include code that is not running yet; prototypes and mockups; specifications and other requirement documents; flowcharts, diagrams, or state models; user documentation; sales and marketing material; or ideas about any of those things. When we refer to testing activity pointed at things that are static, like most of the items in the preceding list, we usually call it “review”; we might also call it “performing a thought experiment”. Review is a kind of testing activity that may be closely or distantly associated with performing a test—which brings us to what we mean by “testing”.
Testing, Test Activities, and Review
When I say testing here, I am using the Rapid Software Testing definition. To us, testing is the process of evaluating a product by learning about it through experiencing, exploring, and experimenting.
Testing includes many activities: questioning, studying, modeling, operating the product, manipulating it, making inferences, analyzing risk, thinking critically, recording the process, reporting on it, etc. Testing activities also include investigating and analyzing bugs and suspicious behaviour. Testing typically includes applying tools to help with any testing activities.
A test is an instance of testing, and to perform a test means to explore, experiment with, and gain experience of a product. In general, to perform a test implies that we will operate and observe a product or its output by some means.
In review, operation of the product as such typically isn’t available. In review, though, we engage in other testing activities as mentioned above. We can’t perform experiments on the running product but, as I mentioned above, we might perform thought experiments on it, imagining interactions between the product and the people using it. Of course, a thought experiment isn’t the same as a real-world experiment; that’s a key difference between review and performing a test.
Why go on about all this? Because reporting is central to our role as testers. We test; we learn; and we report on what we’ve learned.
Are you doing testing work of any kind, or even thinking about doing testing? Then you’ve got a test project on the go, and you can report on its status, even if your report starts with “I haven’t started testing the product yet, but here are some ideas about how we might go about it.”
Next, let’s unpack the idea of a report. A report is a description, explanation, or justification of something. A report is a communication, but a report is not necessarily a document.
Communicating a report might happen as conversation in a hallway, or beside a coffee machine or a water cooler; as a couple of sentences uttered at a stand-up meeting; as a quick mention of a bug in passing to a developer; as a lengthy description of the status of the product and the status of testing at a go-live meeting. A report might be conveyed in writing as a paragraph, a page, or several pages of text; as (heaven help us) a PowerPoint presentation; or as hundreds of pages in bound books, formally presented to a government or regulatory body.
We might include or refer to artifacts collected or produced during the activity that led to the report—the reporter’s raw notes, data sets, program code, design notes for the activity itself. A report might be supplemented with illustrations, charts, graphs, or diagrams, sketched on a whiteboard or formally rendered on glossy paper. Or a report might be accompanied by photographs, audio, video, mind maps, tables, and references to other artifacts.
A test report is any description, explanation, or justification of the status of a test project.
A comprehensive test report is all of those things together.
A professional test report is one that is competently, thoughtfully, and ethically designed to serve your clients in their context. A professional test report need not be a comprehensive test report, nor vice versa.
Some might say that a test report is “just the facts”, but it isn’t; it cannot be. A test report is based on facts, but it’s a story about facts—a story framed for the person or people receiving it. Stories always emphasize some things and leave other things out. We never have all the facts, and facts are sometimes in dispute. Stories are always, to some degree, biased by the storyteller and focused by what the storyteller wants the audience to hear, to learn, and to know. Those biases can seen be as problems in the report, features of it, or both.
The audience for your test report might include insiders who are directly involved in the testing and development work; other insiders (who might be overseeing that work, or affected by it without being directly involved); or outsiders.
For now, I’m going to assume your audience is in the first two categories. On that basis, it helps to consider what the audiences for a test report probably wants to know above all else.
They almost certainly don’t want to know about test case counts (although they might think they do).
They almost certainly don’t want to know about pass-fail ratios (although they might think they do).
They almost certainly don’t want to know about when the testing is going to be done (although they might think they do).
(I realize that these claims may sound strange to you. I will address these (non-)desires in a future post.)
Having been a program manager, a developer, and having worked with lots of them, I can tell you what those people almost certainly do want to know:
What is the actual status of the product? Are there problems that threaten the value of the product? Do these problems threaten the on-time, successful completion of our work?
A test report addresses those questions.
Three Aspects of Test Reporting
A good test report braids three strands of story together:
- a story about the product and its status; what the product is, what it does, how it works, how it doesn’t work, and how it might not work in ways that matter to our various clients. This is a story about bugs, problems, and risks about the product.
- a story about how the testing was done—how the product story in was obtained; how we configured, operated, observed, and evaluated the product. A thread in this second strand of the testing story involves describing the ways in which we recognized problems; our oracles. Another thread in this strand involves where we looked for problems; our coverage. Yet another thread includes what we haven’t covered yet, or won’t cover at all unless something changes.
- a story about the quality of the testing work—why the testing that was done can be trusted, or to the degree that it is untrustworthy, issues that present obstacles to the fastest, least expensive, most powerful testing we can do. In this strand, we also identify what we might need or recommend to the testing better, and we may also provide a context and and evaluation of the quality of the report itself.
Most of the time, the client of the testing will be most interested in that first strand. Sometimes the client might be more interested in one of the other two. Nonetheless, whatever form the report might take, the reporter should at least be prepared to address all three strands.
If you’re not credible, your reports won’t be taken seriously. In your reporting, you may be delivering surprising or uncomfortable information. Your clients, unconsciously or deliberately, may assume that you’re mistaken or that you’re exaggerating risks, and they may try to micro-manage your reporting. Credibility is an antidote to all this.
To build and maintain credibility, it’s important to actually care about the project and the people on it. It’s important to take your work and your skills seriously, and to demonstrate that seriousness in your attitude, commitments, and behaviour. There will be more to say about this later, but for now…
- Actually know how to do your job.
- Gain experience with the product.
- Study the technology in and around your project.
- Read all of the relevant requirement, specification, and standards documents carefully, especially when you’re in a regulated environment.
- Take notes diligently on your own work to inform your reporting.
- Sweat the details in your own work.
- Find things to appreciate about the work of others.
- Acknowledge mistakes, correct them and learn from them.
- Do not tell lies or exaggerate.
Examples of Test Report Documents
Note that Part 7 of this series included a number of test reports delivered verbally. Here I’m providing examples of test report documents.
As you survey them, you might want to consider the context for which they’re intended; the reporting levels that they focus on (product, testing, or quality-of-testing); the evidence or references included to support the report; and what the report might need or could leave out.
Note that while a couple of reports refer to specific things to be checked, there is rarely even a mention of test cases. The focus, instead, is usually on bugs or potential problems in the product that represent risk to the value of the product, and therefore risk to the business.
Spot Check Test Report
Here is an example of a real, comprehensive, professional test report, prepared by James Bach and edited by me. Over five pages, it describes a paired exploratory testing session that found problems in a real medical device. (The names, nouns and verbs have been changed to shield the identity of the company and the product.)
Cheese Grater Incident Report
This is two reports in one: a whimsical yet serious report on repairing a broken Parmesan cheese dispenser; and a much longer, detailed set of notes on how to perform an investigation and report on it. Indeed, the latter section is a really worthwhile complement to this blog post.
OEW Case Tool
An example of a two-page summary report (from 1994!) about a computer-aided software engineering (CASE) tool at Borland.
Y2K Compliance Report
An eight-page report prepared for compliance with Y2K requirements, including notes on strategy; the test approaches that were applied (and risks that prompted those approaches); the results; and a list of specific items that needed to be checked.
OWL Quality Plan
This is a report on proposed plans for testing another Borland product, the Object Windows Library. The report includes a table linking product risks to testing work necessary to investigate those risks. It also includes a listing of components and sub-components in the product.
An Exploratory Tester’s Notebook
This paper on recording and reporting includes a report on my spontaneous investigation of an in-flight entertainment system, and a couple of session-based test management session sheets.
A Sticky Situation
This is an example of a form of reporting that’s sometimes called an “information radiator”. It visualizes the status of a test project (and some degree of test coverage) using sticky notes.
The Low-Tech Testing Dashboard
Of this, James Bach says “Back in 1997, I was challenged by top management to create a way to convey testing status at a glance. Thus was born the “low-tech testing dashboard” which has since been rendered in various electronic, distributed forms. The important thing about the dashboard is that there are no “measurements.” We don’t count anything. Instead there are assessments. These are subjective, yes, but always grounded in evidence.“
Who Killed My Battery?
A splendid research paper on what drains mobile phone batteries… and why. Also a presentation on YouTube: https://www.youtube.com/watch?v=_uv057DP2Vs
Once again, these reports don’t focus test cases, but on testing. They’re examples of powerful and reasonable test reports that offer an alternative to management that is fixated on test cases.
Managers are more likely to relax their obsession with test cases when we provide them with reports that tell the product and testing stories.
Two more posts to go. Next!