Why you should stop using grocery roadmap and try GIST

Free translation of an article by Itamar Gilad, growth and strategy consultant, former Google product manager, on an approach to strategic product development planning.

Over the years I have developed a considerable number of product strategies, roadmaps and Gantt charts for projects. But I don’t do them anymore. I found a better alternative, which I’ll tell you about.

I used to do this:

Strategy - Roadmap - Project Plan - Implementation (Agile).

Such an approach to planning requires a huge amount of work, the mere need to obtain confirmation of plans from all stakeholders is not an easy task, and the ROI of such actions is very small. At the same time, any plans quickly begin to diverge from reality, the larger the planning horizon - the worse. It took me a while to realize that my roadmaps and Gantt charts are no longer relevant by the time I publish them .

In addition, this is a waterfall model (not the same as a waterfall in the design), which means almost complete lack of flexibility: changes at the top of the scheme cause rescheduling and even cancellation of projects at the bottom of the scheme.

Agile approaches in development successfully cope with waterfalls of projects, but not with waterfalls of planning. At the same time, they influence the innovation and culture of the company.

Roadmaps are designed for several large projects, which means prioritization and kills a lot of potentially successful ideas. In top-down organizations, winning ideas come from management. In bottom-up organizations, winning the idea is a difficult task, it is often necessary to create a hype around innovation and use sales skills. In my opinion, this is the last century.

So what is the alternative?

GIST

This is a planning system that I began to use, working at Google and later for several years adapted it to the principles of Lean Startup and Agile. I implemented GIST in several companies and the results are very similar - we have simple lightweight plans at the output, ready for changes and quick updates, less pressure on managers, more autonomy of the development team and increased speed of their work, simplified interaction within the company and as a result much better products and management decisions.

This system is called by me GIST by the name of the main blocks: Goals (goals), ideas (ideas), step-projects (step-projects), tasks (tasks). Each of these blocks has its own planning horizon and implies its own frequency of changes; For each block, different management tools can be used, but they all form planning for the company as a whole and determine the actions of the teams.

Goals - year 1, year 2.
Ideas.
Step projects - quarter 1, quarter 2, quarter 3, quarter 4.
Tasks.

Goals

“If you tell a person where to go, but don’t tell how to get there, you will be amazed at the results”

George S. Patton

Most strategic plans sin by defining solutions, not goals (using technology X, partnering with company Y, starting in country Z). Any general of the modern army will say that this approach should remain in the past - you need to set a goal for the soldiers and provide an opportunity to achieve it yourself ( Mission Command principle). Such an approach implies more authority, it is more meaningful and reduces the burden on management - decisions can change depending on the situation in the field, and the goals remain the same.

The goals within GIST embody these principles - they describe the company's strategy in terms of the desired result: where we want to go, when, how do we understand what came. At any given time, goals answer the question “Why are we engaged in this project?”.

I got to know more closely the process of formulating goals at Google, where every quarter we meticulously recited our goals in the form of OKR . Some people believe that OKR is one of the reasons why Google as a company is so successful.

Examples of goals in the form of an OKR record:

C: Break-even by the end of 2018.
KR: At least 5 new enterprise customers.
KR: Reduce outflow to 2.4% or less.

Q: Provide the ability to use the product on mobile platforms.
KR: MAU on mobile> 20,000.
KR:> 25% of key actions are done on mobile.

C: Pleasant user experience without bugs.
KR: Repair the top 10 bugs that users have reported.
KR: Reduce the duration of technical support response by 40%.

Ideas

“If you want to have good ideas - you need a lot of ideas. Most of them will be wrong and you need to learn to understand which ones to drop. ”
')
Linus Pauling

Ideas describe hypothetical ways to achieve goals. The key word here is hypothetical - you may have a lot of ideas on how to achieve the goal, but often no more than one out of three gives a positive result (usually even less). Ideas of the authorities, experienced product managers and designers are not much more successful than the average.

For this reason, working on the GIST, we never kill ideas prematurely, they do not collide in a deadly fight during prioritization, and management ideas, more hype / impregnated / politicized ideas have no more support than others.

Instead, we:

1. We collect all the ideas in the Bank of Ideas - basically it is a separate database or table in Excel. Any ideas are welcome, and there may be hundreds of ideas in the Bank.
2. Prioritizing, based on evidence - I like ICE-prioritization from Sean Ellis , but you can use any method (to estimate labor costs and returns - not enough !).
3. We plan to test as many ideas as possible - they are implemented as part of the step projects.

Step projects

“Think great, but start small”

One of the 8 pillars of innovation from Google

It is very tempting to choose one promising idea, turn it into a project for 9-18 months and be engaged in implementation. This is a common and very expensive mistake - you can spend quarters, sometimes even years on the still unconfirmed idea. It. burning money, because most ideas are not worth the investment at all.

Instead, we break up large projects into small step-by-step projects, step-projects, each no longer than 10 weeks, and implement them one at a time.

Example of a step project: mocap - prototype - MVP - dogfood - beta - launch (release).

In accordance with the principle of Lean Startup Build-Measure-Learn , each step-project is in fact an independent experiment on testing ideas. In the correct sequence, each step project includes an improved, augmented version of the idea, which is available to an increasing number of users at each new stage.

A real example of a project decomposed into step projects:
Initial Plan: Starting a project → mocap v.1 → mocap v. 2 → working prototype → draft version → dogfooding in the company → alpha → beta → X.
Actual actions: Starting a project → mocap v.1, preliminary research → mocap v. 2, first user research → working prototype, second user research → draft version, dogfooding in the main team → dogfooding in the company → studying the result of dogfooding → alpha → interview with alpha users → beta → beta user survey → usage statistics → Y.

As a result, the final product turns out much better than what we initially imagined. Here it is written why.

Due to the fact that each step of the project is so small, we avoid all the negative side effects of long-term projects, and therefore we have the opportunity to test many more ideas in parallel, while reducing investment and learning faster. Nothing is required to pitch or dilute the policy.

Seeing how your idea is tested in a matter of weeks and being brought to life is incredibly pleasant for all participants in the process. After that, you will never want to work on projects that are doomed to death.

Tasks

In the end, each project is broken down into “one-bite” actions, which are called tasks. This part of GIST is well implemented using Agile tools, kanban boards and other modern development project management techniques. There is no need to change anything, the only difference is that the upper layers now also correspond to the values ​​of the Agile Manifesto and are ready for changes.

Planning cycle

Planning with GIST with layered and iterative:

1. Goals usually have a planning horizon of one year or several years — the ability to think far ahead is encouraged here. They are determined at the beginning of the year and are revalued and adjusted every quarter - we do not want to pursue outdated goals.
2. Ideas are collected and prioritized all the time. We never stop the process of finding new ideas.
3. Step projects are defined at the beginning of each quarter. The team chooses what goals and ideas it wants to realize in this quarter and, accordingly, determines the step-projects on them. The list of quarterly step-projects is re-evaluated and re-prioritized every 1-2 weeks, taking into account the planning of work on tasks.

Tasks are planned in 1-2 week iterations of the preferred method for the team (for example, on planning a sprint if you have scrum ), and are adjusted daily.

In this example, the team works in parallel on four ideas for two goals. For idea 2, the first two steps of the project have already been successfully completed. And the idea №3 did not justify itself at the stage of the first step of the project and was discarded, freeing up resources to do more work for the other three ideas.

Do you still need roadmap?

I think no. Roadmaps are typically used for the following purposes:

1. Planning work - I hope, now I have convinced you that roadmaps are not necessary for this.
2. Internal communication - my experience shows that colleagues and board members readily understand and accept the GIST language - the transition is not so difficult and everyone appreciates the realism and plausibility of the tasks. Of course, the entire planning system should be available to any company employee and board member.
3. External communication - with customers and partners, expectations from a “formal roadmap” can be very high. As usual, it is our job to shift the focus of discussion from specific features to needs.
4. When using GIST, you can respond in the style of “We have a goal to engage in interaction within the product in the third quarter. I can not say exactly how this will work. We have a few ideas and we will work flexibly. Most likely, we will have MVP in the second quarter. Do you want to be among the beta testers and give feedback? ”
5. If you are lucky, and the ideas taken to work will “shoot”, then it will work better than any roadmap schedule.

Conclusion

GIST is not a fundamentally new approach - it is rather a mixture of ideas and methods that have existed separately in recent years. This is an attempt to unite all the layers of planning and create a viable plan, ready for change.

Basic principles:

• There is no separation of ideas, planning tasks and their implementation - all this happens regularly and often at the same time.
• It is necessary to formulate goals instead of decisions or vague formulations of the strategy.
• Idea banks are used instead of product backlog.
• You need to move short step-projects within the quarter, instead of long multi-quarter / multi-year projects.
• You do not need to bet on several big ideas that will take forever - we test a lot of ideas quickly and bring to release only those that work.
• Iterations are applied — revisions of each part of the plan are carried out frequently and systematically, remaining Agile at all levels.

The whole world is developing and to create cool products you need great product managers. Us in the united company “Wheels | Roof | Market ”are also needed, so we have open vacancies for product managers and project managers .

In addition, in June we launch the “ Wheels Academy ”, where we will conduct training for managers , developers, testers and UX-designers.