GIST model
GIST is a hierarchical planning model developed by Itamar Gilad while working at Google, and documented in his book Evidence Guided.
In this model, tasks are grouped into a four-layer hierarchy that assists with prioritisation, planning and guiding development. The levels are Goals, Ideas, Steps and Tasks, and the model takes its name from the initial letters of each level.
- Goals are the desired outcomes - measurable benefits for the customers and for the business
- Ideas are potential ways to achieve the goals
- Steps are activities that help to develop an idea (everything from creating models to a beta launch)
- Tasks are the individual work items that help to perform a step
With a hierarchical breakdown from goals to ideas, and from ideas to steps, this model provides a way to both track progress and keep in mind alternative options.
GIST boards
To prevent the delivery teams from focusing too much on tasks, and to keep focused on the goals, teams should visualise the current plan using a GIST board. A GIST board is a visual representation of the progress towards milestone goals, acting as a big-picture view of the current plan and helping with progress reporting and review. Gilad recommends making the trio of an engineering team lead, product manager and designer responsible for maintaining the board. In this sense, the GIST board is an example of a team scoreboard from The 4 Disciplines of Execution.
The board usually lists goals on the left side. Gilad recommends limiting the board to 4 goals per a quarter of a year (effectively 4 goals per milestone). The middle of the board contains the ideas that are expected to contribute to a goal. The steps to validate and deliver an idea are listed on the right. A GIST board usually does not contain tasks, and teams are expected to manage them in some other planning tool (but they should be referenced back to the steps on the GIST board).
Selecting good goals
Gilad proposes using Objectives and Key Results to select “outcome goals centered on measurably improving” the product experience. In that way, the Goal of the GIST hierarchy sits within the OKR structure at the product manager’s (or product delivery team’s) level of responsibility. Gilad makes a distinction between a “North Star Metric” (how much value is delivered) and “Top Business Metric” (how much value is captured), but they are both effectively tracked at the same level.
By documenting the assumptions about both types of value, GIST method helps to avoid Air Sandwich plans.
Selecting good ideas
In GIST, ideas are options to progress towards the goal. Gilad stresses that the ideas are hypothetical, and that most ideas won’t work. That is why it’s important to try out and test multiple ideas, and invest in those that produce results. Different alternatives need to be evaluated quickly and prioritised (Scoring models such as ICE can be very helpful for that).
The difference between Steps and Tasks
Steps are the “discovery engine of GIST”. They are mini-projects designed to develop an idea further. Gilad suggests that they can be as simple as generating projections in a spreadsheet, or running a beta test with users. Steps are designed to increase confidence in an idea - and they will usually progress through some initial analysis to interviews, usability tests, internal and external beta testing to the full launch. Steps describe the workflow of how to develop an idea, not the individual work items in that flow. Tasks capture the individual work items.
How GIST differs from impact maps?
Both the GIST model and Impact Maps connect work items to goals through several hierarchical levels, promoting the idea that the same goals can be achieved in different ways, so their purpose is very similar. However, they differ in granularity of the detail captured.
Impact maps capture more details around goals: At the top level, GIST captures “goals”, which are benefits both for the customers and the business. There is no big difference between the two levels of value in this hierarchy. Impact maps make a significant distinction between the two types of value. The “Goal” in impact mapping terminology is the business value (corresponding to the top business metric in GIST terminology). It is then broken into potential impacts (user value), providing potentially alternative options for reaching the goal. In that way, impact maps promote more optionality around achieving goals. Impact maps additionally group impacts under different groups of actors, capturing more easily situations where a business goal involves delivering value to multiple customers or user segments.
GIST captures more information about deliverables: When capturing deliverables and work items, GIST is a lot more detailed than impact maps. It breaks down work into steps and tasks, which can then be directly assigned to delivery team members and tracked on the same hierarchy. Impact maps contain deliverables at the lowest level, but these are typically relatively coarse, and intended to be taken into a different tool for managing work. In that regard, the “deliverable” part of impact maps most closely corresponds to the “idea” part of GIST, without even trying to capture the workflow to test an idea and the tasks required to implement an idea. From that perspective, impact maps can be more useful for a collaborative brainstorming session, where GIST hierarchies can be more useful for driving actual work forward.
GIST slices by confidence, impact maps slice by user segment and impact: While both impact maps and GIST promote the idea of optionality, where the goals can be achieved in many different ways, GIST is more flexible and more detailed in that regard. Impact maps are fairly strict about the middle of the hierarchy, preferring measurable user or customer behaviour changes. GIST is more open, allowing different types of ideas that could contribute to the goal. As a big difference, GIST also introduces the component of confidence for different ideas (typically through the Confidence Meter. Impact maps have no specific way of capturing confidence. On the other hand, because impact maps capture actors and impacts explicitly, they make it easier to slice work by impact (e.g. by reducing the user segment or delivering a smaller impact). GIST makes it easier to slice work by confidence, showing individual steps to increase confidence for an idea clearly.
Learn more about the GIST model
- Evidence-Guided: Creating High Impact Products in the Face of Uncertainty, ISBN 978-8409536399, by Itamar Gilad