Today I give you a mental tool that programmers can use to evaluate whether a product requirement is sound or not.

This can be used not only for the daily communication between programmers and product managers, but also for the evaluation of requirements in the process of the company, and even to establish a scoring system to do the data analysis of a large team.

Assess the three dimensions of requirements

R&d personnel’s assessment of product demand can be divided into three dimensions:

  • Value identification – is it valuable to users and what is the input/output ratio

  • Requirements quality – are requirements easy to understand, are details clear, and are logic valid

  • Technical feasibility – can do it, how much cost, how much risk

When the evaluation of a requirement is divided into three separate and mutually exclusive dimensions, development and product communication can be more comprehensive. When we say that the needs of product managers are reliable or unreliable, we tend to refer to one or two of three dimensions.

For example, if the product manager requests to add baby to all users’ nicknames on Children’s Day, the developer can reply:

“I understand the demand (demand quality – high), but arbitrarily changing users’ nicknames is not in the interests of users, there may be legal risks (value recognition – low), it is difficult to change all users’ nicknames, I need to investigate with other developers (technical feasibility – low)”

Dimensional scoring system

When we evaluate the needs, we can quantitatively score the three dimensions. The purpose of quantitative scoring is twofold. On the one hand, it can make the evaluation conclusion of R&D more rigorous and clear. On the other hand, with quantitative scoring, we can do data analysis, such as horizontal comparison of requirements and quality distribution of each product manager.

Please refer to the following scoring criteria, ranging from 0 to 5 with a median of 3.

Value identification:

  • 0 – Completely disagrees with the value of the requirement and cannot be developed

  • 1 – Limited value, disproportionate input and output, do not agree to put into development at this stage

  • 2 – Reservations, need to be discussed

  • 3 – Basic recognition, can be put into development

  • 4 – Agree

  • 5 – Strongly agree

Quality of demand:

  • 0 – Unacceptable

  • 1 – The quality of requirements is very poor. Review after re-sorting

  • 2 – There are product bugs/illogical/important omissions/difficult to understand and need to be modified

  • 3 – Basically understandable, details need to be added

  • 4 – Good demand quality

  • 5 – Good quality

Technical feasibility:

  • 0 – Cannot be implemented

  • 1 – The difficulty is high, and some functions may not be realized, which requires research

  • 2 – Some difficulty and high cost

  • 3 – Regular items

  • 4 – Easy to implement, or do not need to publish

  • 5 – No development required

For each requirement that may go into development, there is a score by the r&d team, which can be recorded as an appendix to the requirements document, and the score is dynamically updated as the situation changes.

Or give the children’s Day to the user nickname plus the baby example, if the scoring system to evaluate, may be the following results:

  • Value identification: 0 – Modify user information without authorization, violate user interests, may have legal risks, do not agree to enter the development

  • Requirements quality: 4 – Requirements have been understood

  • Technical feasibility: 1 – Change the nickname of all users, it is difficult to achieve, need to discuss the scheme

This way, the development team’s attitude towards the requirement is clear.

conclusion

Having established three dimensions for evaluating requirements, developers can communicate more fully with product managers. On this basis, we can establish a scoring system to make the results more rigorous and clear. Team data analysis can also be done after the scoring results are recorded and summarized.

The experience in data analysis will be elaborated in a subsequent article. Hopefully the three dimensions and the grading system will work for you.