Post by account_disabled on Mar 5, 2024 11:07:42 GMT
A lean version of a product or solution in a few days, containing only the essentials. As a result, the need to estimate delivery times loses strength, because we focus our efforts on prioritizing items very well, mainly using the relationship between the difficulty of executing the work and the return value for the business . So, we know that something of great value will be delivered in a few days or weeks. Country Email List Estimates and predictability however, lean inception and other agile product development techniques such as google 's design sprint — which also aims to deliver something of great value in a short time, without having to make elaborate estimates of effort or deadline — are not a “ silver bullet ” and , therefore, may not be applicable in all cases or in all scenarios. Therefore, it is important to know the existing practices and when one may be better than the other. Lean inception and design sprint are techniques recommended for application in the construction of mvps , short-term projects, and smaller product deliverables, because with them it is possible to quickly validate and launch new solutions to solve considerably small and specific problems.
For ongoing operations and longer projects, a practice within the no estimates universe that offers good benefits is predictability using statistics. Its nature comes exclusively from looking at the past (at facts and work history), applying statistical calculations on collected data and, only then, presenting an answer to that question: when will it be ready? In this case, if there is a way to collect historical data, then it is possible to provide statistical predictability. Predictability predictability is a calculation about the future that includes a past interval and a probability of that interval occurring . In the book “when will it be done?” by author daniel vacanti , we can find two types of predictability being addressed: deterministic forecast: where 100% certainty is assumed, for example, this feature will be delivered on december 1st; probabilistic forecasting: where we balance the level of certainty and do not guarantee an outcome, for example, there is an 85% probability that we will deliver the feature on november 1st. Before we delve deeper into predictability, we need to understand that we tend to base ourselves on averages extracted from previous data.
However, there are some reasons why we should avoid this simplistic use in our calculations: averages do not communicate the percentage of confidence of completion (like rolling two six-sided dice – it is a very random process); their cycle times are not normally distributed; your average may take into account a single tragic event that defined cycle time in the past. Furthermore, we need to know the data that we are going to use in our calculation, it is easy to trust the data and even easier to make mistakes. To do this, you need to understand where this information comes from. If you use automated tools, make sure you understand their mechanism. As well, it is necessary to pay attention to the data, as for example the data from december may not be ideal for predicting the delivery speed for the month of march, given that in december we have christmas, most of the time a large part of the the team takes the opportunity to enjoy their holidays , as well as company parties. Additionally, if you are using a physical board, be sure to record all relevant information , such as date/time, types of work, etc. Getting the data concise, we need to understand the extreme outliers . The best way to deal with them is to analyze them closely and understand what caused these unusual delays, however, you cannot guarantee that this will not happen again.