Este documento descreve o processo "Shape Up" usado pela Basecamp e adotado pela AskNicely. O processo inclui ciclos de 6 semanas de trabalho de engenharia na Basecamp (4 semanas na AskNicely) seguidos por um período de "cooldown" de 2 semanas (1 semana na AskNicely). O processo começa com a "Formação" onde a estratégia do produto identifica potenciais tarefas. Uma vez definido o apetite (o tempo, esforço e recursos que a empresa está disposta a gastar), uma solução é formada para caber dentro do apetite. O resultado é um documento de "Pitch" que articula o problema e propõe uma solução. Depois, os líderes da estratégia do produto realizam uma "Mesa de Apostas" onde os pitches são considerados para o próximo ciclo. Finalmente, as apostas são atribuídas a pequenas equipes que trabalham durante o ciclo de construção.

Shape Up Process

Basecamp uses a 6-week cycle during which engineering is building work, and product are shaping upcoming work. After each cycle is a 2-week cooldown period during which engineering is free to fix bugs and technical debt, and product stakeholders are selecting (betting) for the next cycle.

At AskNicely we run a slightly shorter 4-week cycle and a 1-week cooldown.

Shaping

Those in charge of product strategy identify pieces of work as candidates for shaping.

First the Appetite the company has for undertaking a piece of work must be set — i.e. how much time, effort, and resources they are willing to spend:

• Large batch (a full build cycle, 4 weeks)
• Small batch (part of a build cycle, 1-2 weeks)

Once the appetite is set, a solution is Shaped that fits realistically within the prescribed appetite. Shaping is a focused collaborative exercise between a small group such as:

• The executive who proposed the work as strategically important
• A PM responsible for developing the piece of work
• A technical lead who can validate approaches or point out issues

The outcome of the shaping process is a Pitch document that:

• Articulates the problem, explains why the status quo doesn't work, or where there is a functionality gap
• Proposes a solution that can be realistically achieved with the set appetite
• Highlights any potential "rabbit holes" that have been identified so far
• Explicitly state any "no-gos" that are intentionally out of bounds and excluded

Betting

For each cycle, product strategy leaders and stakeholders come together to hold a Betting Table, where the list of shaped pitches is considered for betting on. This small group consists primarily of executive-level product strategy stakeholders.

Pitches that are considered well-shaped and strategically important enough are betted on for the next cycle. Any concerns or requirements that arise from discussion are added to the pitch documents.

It's important to understand that pitches only contain high-level concepts with no high-fidelity mockups or wireframes.

For this reason, it is critical for all betting stakeholders to review the pitch documents up for discussion in detail before hand so that the betting session can run efficiently instead of being dragged down by having to achieve basic understanding of the work being discusses.

Build

Bets are assigned to small teams who self-organise to design and implement a solution during the 4-week build cycle. Teams work on either a single large batch bet, or a set of 2-3 small batch bets. A team usually consists of:

• 1-2 engineers
• 1 designer
• 1 product manager

The first thing the team does is identify the smallest complete vertical slice that they can complete and deploy to demonstrate progress to stakeholders within 3 days of starting the project.

During the cycle engineering does only focused feature work and is not interrupted to fix bugs or support customers. Any issues that arise during the cycle can wait to be fixed during the cooldown period.

In a very small number of extreme scenarios engineers may be diverted to address critical issues. In 2019, the only incident that was urgent enough to interrupt a cycle was the massive performance issues with Chime that rendered the product unusable for them. Everything else could wait.