Participatory Budgeting is a process through which a group of stakeholders decides on, or contributes to, decisions made on the use of shared resources, such as portfolio budgets or development team capacity.
By creating forums in which participants explore, shape, and negotiate over the use of resources, Participatory Budgeting helps ensure that the right decision-makers are making the best possible decisions with the best available information.
Lean Portfolio Fiduciaries use Participatory Budgeting to provide Business Owners, Epic Owners, Product and Solution Managers, Solution-System Architects and other stakeholders with the forum needed to effectively and efficiently implement Lean Budgeting. Participatory Budgeting is often referred to as Collaborative Budgeting when used internally.
Business Owners, Product and Solution Managers use Participatory Budgeting to engage with customers, partners and other external (and internal) stakeholders to determine preferences and priorities associated with Features and Capabilities. In this application, the results feed Weighted Shortest Job First (WSJF) calculations by providing clear insight into stakeholder preferences and increasing the confidence of the user-business value. This application of Participatory Budgeting is also known as Buy a Feature.
Participatory Budgeting in Action for Portfolio Planning
To better understand how a SAFe portfolio might use Participatory Budgeting (PB), let's consider an example of a for a bank inspired by this article. The timing of the PB forum would likely be associated with the dynamic budgeting events, typically on PI boundaries, with budgets adjusted twice annually (see Lean Portfolio Management).
Here is a sequence that we've found helpful:
Preparing for the Participatory Budgeting Event:
2) Understand current Value Stream budgets.
3) Develop Lean Business Cases for Epics, ensuring that a rough order of magnitude is used to provide an estimate of projected costs is included for each epic (see further guidance on this, below). These Epics are contained within your Portfolio Kanban and are ready for "Go/No-Go" Funding decision.
4) Develop budgets for Enabler Epics.
5) Prepare the Participatory Budgeting Content.
Conducting your Participatory Budgeting Event:
1) Gather the right people.
2) Conduct your forums.
3) Act on the results.
Let's review each of these.
Confirming Strategic Themes
The bank has established three strategic themes for this portfolio:
1) Reduce costs by retiring obsolete systems.
2) Improve ability to release on demand by improving DevOps.
3) Increase usability of applications.
An expected result of Participatory Budgeting are funding of value streams and epics that align with these strategic themes.
Understand Value Stream Budgets
The SAFe Portfolio Canvas provides a convenient, single-page overview of the portfolios development value streams. Our example portfolio comprises three development value streams with a total portfolio budget of $9,500,000:
Loan Origination: The loan origination value stream supports the systems by which a borrower applies for a new loan and the internal processing of that loan application, including the disbursal of funds (or declining the application). Current budget: $3,800,000.
Fraud Management: The fraud management value stream develops the systems that manage fraud. It is separated from loan origination to comply with audit and security requirements. Current budget: $2,500,000
Credit Scoring: The credit scoring value stream uses multiple sources of data and various strategic partners to develop a credit score of loan applicants. Current Budget: $3,200,000.
Note that while we're not explicitly identifying the time frame of this budget, it can be reasonably assumed to be a time frame of approximately 6 to 12 months, consistent with SAFe Lean Portfolio Management.
To prepare for the Participatory Budgeting session the Lean Portfolio Team should identify a list of the highest priority Epics (Business and Enabler) that are in the Portfolio Kanban. Some of these may be Epics in-flight that are included in the current budgets of the value stream, and are included in this list for the next PB because they are projected to exceed guardrails established by the portfolio for the value streams. Others are new ideas that have been added to the portfolio Kanban through Continuous Exploration and have completed all of the activities under Analysis, with the exception of the GO/NO GO decision (this decision is incorporated into the PB process). All of them are of a magnitude that "trips the threshold" of portfolio spending for a single epic, affect multiple development value streams, and/or take longer than a single PI, which means they are subject to review by the LPM team.
Let's return to the "cost estimate" or "projected cost" of these Epics and Enablers. If we force the team to create a detailed cost estimate we're inadvertently forcing the team to adopt traditional practices because of the level of detailed planning required for a detailed estimate. A better alternative is to provide a rough order of magnitude that is used for prioritizing initiatives, with input from the technical community - the only people who truly know how much something will cost.
Here are the Epics and Enablers that will be included in this PB forum:
Explore New Fraud Service: This is a Value Stream Epic associated with the Fraud value stream to develop a prototype integration of a new fraud service from a SaaS vendor. Because it is new and uncertain work, and because we are not yet fully committed to this new solution, and because it exceeds the budget threshold for an epic that has been established, we are electing to place this into Horizon 3. Projected Cost: $600,000.
Decommission Legacy Credit System: This Horizon 0 Epic associated with the Credit value stream completes the decommissioning of our credit system by migrating the last of the historical data from the old legacy system into our data lake. Projected Cost: $1,000,000
Move to Angular: This Enabler is designed to align ALL of the value streams to Angular, starting with the loan origination system. Projected Cost: $800,000
Move Configuration Management to Puppet: This Enabler is designed to improve the ability for the ARTs to manage infrastructure. This includes training for all teams in the Portfolio across all value streams. Projected Cost: $400,000.
The total proposed cost for all Epics and Enablers is therefore $2,800,000.
Before going further, let's consider some key aspects of potential tension and how SAFe LPM practices can resolve them.
Q: If the Value Streams have their own budget, and value stream owners are expected to manage that budget through decentralized decision-making, why are Epics within a value stream subject to LPM review?
A: Healthy organizations based on transparency also have checks and balances that ensure that spending is aligned with the organization. By establishing budget guardrails for large initiatives, the organization ensures it is balancing autonomy with governance.
Q: Why are Epics or Enablers that affect multiple value streams allocated to a single value stream for implementation?
A: If you place the locus of ownership for an Epic or an Enabler outside of a value stream you inadvertently create a project-centric culture - with all of the challenges that value streams are designed to avoid. A far better choice is to place the responsibility for the Epic or Enabler within a value stream, even when that Epic or Enabler affects multiple value streams.
Preparing the Participatory Budgeting Content
Let's summarize our content:
The portfolio has a total budget of $9,500,000 allocated against three value streams. We assume the value stream owners understand the details of how they are planning to manage their budgets.
The portfolio is considering 2 Epics and 2 Enablers with a total projected cost of $2,800,00.
Our goal is to make the funding choices explicit. Accordingly, our content for the PB forums will list all options for consideration, for it is during the forums where the magic of PB takes place. Here is the content at the start of the forum:
Note that the total project budget if all value streams are funded at their current level AND all Epics and Enablers are funded is $11,300,000. This exceeds the current budget of $9,500,000 and reflects the total of the Epics and Enablers.
The reality, of course, is that the proposed Epics and Enablers will be funded through the existing budgets of the value streams. Of course, the Portfolio team might be able to secure additional funding, but in this example the budget is fixed at $9,500,00. We'll see how the LPM fiduciaries and value stream owners allocate these funds.
Gathering the Right People
The most common participants in SAFe Participatory Budgeting forums are: Business Owners, Enterprise Architects, Product Managers, Solution Owners and other executives. Once assembled, you'll find that you often have as many as 20 to 30 people who should be included in the process. Because this is too many to structure healthy collaboration, you will organize multiple forums of 5 to 8 participants and then analyze the results across the forums to identify actionable patterns.
For our example, let's start with a single forum focused on Business and Solution Owners, one each from each of the value streams, plus two technical leaders who are championing the Enablers. These 8 people represent a core set of constituents for the allocation of funds. Because the team is distributed in Europe, India and the United States, they use the Conteneo Weave platform to manage their forums.
Conducting the Forum
The forum starts quickly. The Credit Value Stream team of Ming (Business Owner), Kerstin (Solution Owner) and Franklin (one of the Enterprise Architects), quickly fund the Epic to decommission the legacy system. Because they intend to fund this epic from the projected budget of the Credit Scoring value stream, the credit scoring value stream budget is reduced by $1,000,000.
A similar coalition quickly forms around the Epic Explore New Fraud Service, as every participant is excited about the potential of the new technology, which as a unique set of machine learning capabilities that could help drive innovation in all of the value streams. This is reflected as a uniform purchase, with the Fraud Management budget being reduced by $600,000 to reflect the agreements.
Having funded the most important Epics, the team turns its attention to the value streams. They agree to fund the value streams based on their natural affiliations. Accordingly Ming and Kerstin allocate their budget to the Credit Scoring value stream; while Joe (Business Owner) and Sallee (Solution Owner) focus on Loan Origination and Satish (Business Owner) and Lauren (Solution Owner) focus on Fraud Management.
The forum is now more complex, and the negotiations will become harder. The remaining budget of the participants is $2,050,000 while the remaining needed budget is still higher.
The participants with the greatest budget are John and Franklin, architects representing technical concerns. Lauren and Satish also have budgets and are comfortable allocating these budgets to either the value streams or the remaining enablers.
Through a challenging set of discussions the team realizes the following:
The loan origination team had some challenges during the last PI and cannot take the lead on the move to Angular. Furthermore, the loan origination team needs agreement from the other participants to secure their remaining $575,000 in requested funding.
The credit scoring team does not want to tackle moving to Angular while decommissioning a legacy system. However, because the team is focused on infrastructure, the credit scoring team would like to work on Puppet. Accordingly, funding is reduced for credit scoring to reflect that adopting Puppet is associated with this value stream and the architects manage the difference.
The focus on infrastructure in the credit scoring team created a small amount of excess capacity in the form of 1 Agile team. These capacity will be allocated to the loan origination team.
The fraud management team decides to adopt angular, but because their solutions are not used by external customers the cost of moving to angular is perceived as lower. This lowers the initial investment, which will be funded through the fraud management value stream.
The final set of changes are made and the allocations are adjusted between the team members. The resulting forum still represents the total allocated budget of $9,500,000, with empowered business leaders making the best decisions available.
Different forums would produce different patterns of results, but in the end the allocations would match the costs and budgets of the participants.
Note: The Portfolio Canvas would continue to capture the budgets of the three value streams:
Loan Origination: $3,000,000, a slight increase, but designed to solely focus on their core deliverables.
Fraud Management: $3,100,000 = $2,000,000 for Run the Business, $600,000 for exploring a new fraud service, and $500,000 for adopting Angular.
Credit Scoring: $3,400,000 = $2,000,000 for Run the Business, $1,000,000 for Decommissioning the legacy system and $400,000 for Adopting Puppet.
This result is common in Participatory Budgeting forums.