A feature, within the scaled Agile definition (SAfe), requires a benefits hypothesis and acceptance criteria. These establish what and why you are testing and how you will determine success or failure. Each feature will usually have three key components that form the minimum requirements: Beneficiaries. These are needed upfront to establish the hypothesis and the acceptance criteria. Benefit hypothesis. Acceptance criteria.
The minimum requirements already discussed can be made more detailed by covering a few related areas:
The benefit hypothesis.
The feature’s business value.
A clear feature description.
The two fundamental elements of the benefits hypothesis and the acceptance criteria can be unpacked in a little more detail to illustrate their individual and collective roles for features.
The benefit hypothesis is the business value that the feature is expected to deliver. Similar to a scientific hypothesis, this is a statement that will ultimately be tested to see if it is correct. A good formula to use is:
If (proposition), then (benefit)
The proposition is what your team plans to deliver, while the benefit is the value that this will deliver. Benefits can be business-side and include:
Improved data streams.
On the client-side, benefits can include:
Increase customer satisfaction.
Greater simplicity for better customer experiences (CX).
As a Product Manager writing a feature canvas, it is a good idea to ask a few qualifying questions upfront. A few things to consider include:
How likely is this proposition able to deliver this benefit?
Is this feature’s success rate quantifiable?
You must be able to validate your hypothesis to measure the relative success or lack of success of the related feature. Ongoing optimization or even a decision to pivot will not be possible without the ability to quantify how well the proposition succeeded in delivering the benefit.
In Scaled Agile Frameworks (SAFe) a feature’s acceptance criteria are usually written by the stakeholder or the product owner. The acceptance criteria should provide a framework to measure whether the benefit is being delivered by the proposition. In other words, has the feature shown the benefit hypothesis to be correct? If not, is it possible to optimize or would it be better to pivot?
The main functions of acceptance criteria are:
Determine if the feature has been implemented correctly.
Establish whether the business benefits are being delivered.
Mitigate implementation risks.
Facilitate early validation testing to prevent unnecessary costs and effort.
Inform user stories and functional tests.