Features

Flow

Over time, many companies find themselves slowing down and needing a significantly larger team to deliver the same business value they used to. This is where flow really shows its value, as an enabler for almost any other thing. The agile values go a long way to restoring flow, but it is not
AccidentalQuality introduces the following ways for you to control, understand and improve your flow:

Historic view on metrics

We visualize a multitude of different data for you, but just to name a few, as well as their impact:
  • Defect recovery times over time

    Defects (bugs) slow you down significantly, as they build up. They introduce friction for your customers, and make your codebase harder to work with. The longer it takes to fix a bug, the more they will build up to the point of no return.

  • Cycle times for pipelines

    Automation and pipelines only make a difference when they are taken seriously, and constantly maintained. A lot of engineering teams do not manage to prioritize their pipelines, letting them grow to slow and annoying roadblocks.

  • Lead times for code

    This is relevant, not only because DORA names it as one of the key metrics, but because this is the minimum time from concept to customer. It limits your ability to patch software, to introduce new strategic features and to test an idea with a customer.

  • .. and so many more, for a complete list see our Documentation


Having a historic view on any of these metrics, as well as being able to clean out outliers, see the average and understand the current trend helps you navigate where to improve.

Visualized flow

Visualizing the path software travels, helps mitigate one of the most difficult things about software - it is abstract. By visualizing the flow, a lot of observations can immediately be made about which parts of the software leads to bottlenecks, where are things stuck waiting and how long will it take for a key feature or bug fix to make it to the customer.
We visualize your lead times in the way that your software stack actually flows. Working with your lead times will bring about increases in delivery, predictability and quality as well as happier engineers.

Cultural inhibitors in flow

Software is tricky, and incredibly resistant to traditional management techniques. Anybody who has ever tried to get an engineering team to commit to a deadline, and missed that deadline later down the line, knows this to be true.
AccidentalQuality is made to mitigate and expose exactly the problematic priorities that stand in the way of delivering software blazingly fast and at high quality.
Below is a quick overview of some of the metrics that are tracked, which introduce immediate culture change when applied:
  • Pipeline repair times

    Automation only brings value when it functions as a delivery system, and when it is the highest priority to fix failures. A large chunk of companies end up ignoring broken pipelines, and end up with long queues as a result. Simply tracking the time to recovery, and making it the highest priority for the team, will immediately speed up delivery.

  • Time between automation and failure

    Automation (pipelines) are essentially feedback loops. The longer you wait for the feedback, the less efficient it becomes - you may forget what it was, be caught in a different task or simply wait for it to come. Failing fast is a DevOps mantra, and tracking how long failures take help improve engineering effectiveness across the board.

  • Blocking and critical issues

    One thing is to have a task tracking system (like Jira or Github issues) where you can mark priorities, a different thing is to actually expedite those that matter. We visualize your ability to execute on high priority tasks.

DORA

The DevOps research assessment team has had a significant impact on how management groups around the world regard software, KPIs and metrics as a whole. Backed by solid research, DORA shows companies which category of the market they are in when it comes to their execution. Since its inception in 2014, it has proven that companies in the lower echelons of performance simply cannot keep up and lose market share over time - making it essential to correct course before a competitor comes along and disrupts!
We combine all of your integrations to form the four DORA metrics for multiple areas of your sofware: - [Lead time](@alex reference to above feature) - Deployment frequency - Change failure rates - Time to recovery

Dashboards

While AccidentalQuality takes an opinionated approach to what is shown by default, we recognize that every company has different needs depending on your role in the organization, and which team you interact with. We therefore enable custom views on what graphs, visuals and data is most important and allow you to change and control the dashboard(s) to your liking.

API

Everything we process is first and foremost yours. We do not give your data to anyone else, but we open everything in our API so you can retrieve the data and reuse or change it to your liking - allowing management to construct their own PowerBI reports or what have you. To make this possible, we use the same API as we give you so that everything is always available to you.
Should you wish to learn more about our API, go to our Documentation.

Coming soon!

SAFe flow metrics

Scaled Agile Framework for enterprises has a more concrete take on what types of metrics helps Scrum teams attune to flow, and accelerate. These metrics inherit from Lean, and similar practices and give a better understanding of where queues build up, and where time is wasted in the process.
As a result, we are working on bringing these metrics to our customers in the near future - we are already in trial runs with select partners.
The metrics provided by SAFe themselves are as follows:
  • Flow Distribution

    A breakdown of how many issue types are currently in the workflow, such as features, bugs etc.

  • Flow Velocity

    Known from Scrum as 'Velocity', which is completed items in a timeframe (usually a sprint).

  • Flow Time

    How long it takes from something gets started on, until released.

  • Flow Load

    Work in progress across all steps in the workflow.

  • Flow Efficiency

    The time issues spend on the backlog vs being worked on, which helps understand where things are queued up.

  • Flow Predictability

    Planned vs delivered value.

Coming soon!

These are features we are currently collecting feedback on with our customers. If you see something you think is essential to your business, you are more than welcome to reach out to us at [email protected] - we love having a direct dialogue with both current customers and ones that are considering us!

Six Sigma metrics

Six Sigma is generally regarded as essential in mass manufactoring, and works deeply with quality. A lot of the quality overview for Six Sigma relies on per unit defects, which is not a concept that transfers to software.
Currently we are considering if there is a better way to make similar concepts available in AccidentalQuality.

Feedback

To utilize FLOW, you will need to start changing what does not work and learning from it - this requires giving and getting feedback, from both your tools but also the organization as a whole.
Tightening the feedback loop will start to manifest as better quality, security and stability across your products.

Automated reporting

Reporting is an essential component for management to understand what is under control, and what requires further action. Many companies have employees spending a significant amount of time writing and collecting data for such reports, but it does not have to be that way.
  • We automate your reporting, by collecting all the metrics and then letting you pick what to showcase.
  • Comments and reasoning for *why* the state of something is as it is, is included, which allows you to inform stakeholders of what to do.
  • Should stakeholders wish to get the raw data, you can simply extract it via the API
  • Reporting done in this manner has an additional benefit, since it encourages improvement. This is partially how organizations start moving toward a better tomorrow, because data starts driving decisions rather than gut instincts and tradition.

Process inspection

Getting feedback on process is another thing. Most processes are defined in theory, and never proven in practice - we just assume people follow the agreed upon process.
Visualising how work items actually flow through, from engineer to customer, is an eye opener for most line managers and tech leads - because it does not at all look like the agreed upon guidelines.
  • Breaking down a given issue or task to its components, and how it was actually worked on helps show whether teams follow the agreed upon process. This is useful to modify 'bad' process, or to help alleviate compliance concerns.
  • Visualising the steps an issue undergoes, or a commit, helps guide understanding of which parts can improve.

*DISCLAIMER*

Note that this feature is currently undergoing alpha testing with select partners, and may not be released to everyone in its curent format. If the above sounds alluring, please contact us at [email protected] so we can include your company!

Coming soon!

Improvement suggestions

We have worked extensively with all areas of DevOps, and after a while patterns emerge. A lot of simple changes can quite drastically change results.
A simple example that we often give, is how companies approach bugs. Let's assume that a fictional company has 20 bugs that are older than 60 days, and most bugs that are raised are solved within 60 days.
Introspecting and seeing that these 20 bugs are much older than the norm, we can quickly conclude that once a bug is pushed more than the 60 days, the chances of that bug ever being solved are slim. Rather than dragging these 20 bugs around on a backlog, never intending to fix them, they should just be closed.
The above is a simple example, but there are many others with more effect and that are much subtler. We intend to give every customer access to as much feedback as we can programmatically determine, coupled with best practices that seem to be skipped - like in the example.

*DISCLAIMER*

Note that this feature is currently in its design phase, and being iterated on. It is unclear when this will be available, but reach out to [email protected] and we can have a dialogue and provide some of these insights ahead of time for your company.

Improvement

Improvement happens when you slowly take steps forward, putting guard rails in place to ensure you do not slip backwards again. A culture of improvement starts small, but over time leads to a significant competitive advantage.

Targets and trends

One of the ways to ensure improvement, is to introduce any positive step forward as the new normal. Setting a target, and then tightening that target once achieved does something very similar. For example, having a backlog of defects or technical debt is common. Prioritising a target for what to achieve in a given time period, and then tracking that target allows discussions to be had that are hard to reject.
  • Set targets for any metric, from lead time to defect count.
  • Track those targets over a time period.
  • The trend shows if it is improving, stagnant or deterioating. When a trend is not behaving as necessary, pressure can then be put on the process.

Coming soon!

Improvement over time

We want to collect and represent what has improved over time, allowing you insights into how far you have come.
This feature is currently not prioritised, and is merely in its idea phase.