Disciplined Agile

I was recently asked how is technical debt addressed in Disciplined Agile Delivery (DAD), a very important question.  Because DAD promotes a full, explicit delivery lifecycle there are many opportunities to first avoid creating new technical debt in the first place and second to address existing technical debt appropriately.

Here are some of the strategies that DAD promotes when it pertains to technical debt:

1. Do a bit of up front thinking.  One of the process goals of DAD is Identify Architecture Strategy.  By thinking through critical technical issues before you implement your solution you have the opportunity to avoid a technical strategy which needs to be reworked at a future date.  The most effective way to deal with technical debt is to avoid it in the first place.
2. Have an explicit architecture owner.  The Architecture Owner (AO) on a disciplined agile team is responsible for guiding the team through technical decisions, particularly those at the architecture level.  AOs often mentor other team members in design skills, skills that should help them to avoid injecting new technical debt into the environment.  They should also be on the lookout for existing technical debt and where appropriate motivate the team to address that technical debt when appropriate.
3. Be enterprise aware.  Disciplined agile teams are enterprise aware, realizing that what they do should leverage and enhance the overall organizational ecosystem.  They will work close with your enterprise architecture and reuse/asset teams, if you have such, so that they can take advantage of existing assets.   Assets could include code, patterns, services, templates, guidelines, or anything else worthy of being reused.
   An important strategy for avoiding technical debt is to reuse existing assets and not rebuild or rebuy something that you already have.
4. Refactor technical debt away.  DAD provides guidance for when to apply several forms of refactoring, including code refactoring, database refactoring, and user interface (UI) refactoring.  Refactorings are typically very small, such as renaming an operation or splitting a database column, so should just be part of everyday development.  Rework, on the other hand, is more substantive and should be explicitly planned.  The Architecture owner will often negotiate rework-oriented work items with the Product Owner (the person on the team who is responsible for prioritizing the work).
5. Regression test continuously.  One of the easiest ways to find problems in your work is to have a comprehensive regression test suite that is run regularly.  This test suite will help you detect when defects are injected into your code, enabling you to fix them, or back out the changes, right away.
6. Automate code/schema analysis.  There are many tools available for assessing the quality of your code and even your database schema.  Disciplined agile teams will include the use of these tools in their continuous integration (CI) strategy.  Knowing where your technical debt exists is the first step in removing it.
7. Measure technical debt.  Organizations that are serious about technical debt measure it, something that code/schema analysis tools help with, and more importantly keep an eye on the trends (which should be going down over time).  You may choose to track code quality metrics, data quality metrics, usability metrics, time to address defects, time to add features, and many other things.
8. Explicitly govern technical debt.  Several of the previous strategies require investment that some organizations wouldn’t normally consider to be part of the mandate of a delivery team.  For your organization to succeed at reducing technical debt it must be governed, albeit in an agile fashion.  This means it needs to be understood by senior management, measured (see previous point), and funded.  The DA tool kit includes explicit guidance around how to govern agile teams effectively.
9. Reducing technical debt should be part of your culture.  Technical debt isn’t going to fix itself, and worse yet will accrue “interest” over time in the form of slower and more expensive evolution of your existing assets.
10. Address technical debt before handing over an asset.  Passing systems with high technical debt to other teams, such as a sustainment team or maintenance group is generally a bad practice. It should be ingrained in your culture that each team is responsible for keeping the quality of their solutions high. It is reasonable to expect maintenance groups to resist accepting systems that have high technical debt.
11. Accept some technical debt.  Sometimes you will decide to explicitly accept some short term technical debt for tactical reasons.  Perhaps you need to get something developed quickly because you are running a market experiment (a la Lean Startup).  Perhaps there is a new component or framework about to be delivered by another group in your organization, so you’re writing a small portion of what you need for now until you can replace it with the more robust asset.  Regardless of the reason, part of the decision to accept technical debt is to also accept the need to pay it down at some point in the future.  Having good regression testing assets in place assures that refactoring accepted technical debt in the future can be done with low risk.

There are many good online resources regarding technical debt, and the best single one that we have found is Israel Gat’s blog.   Technical debt is real and you need a viable strategy to manage it.  Otherwise you run the risk of slowly choking the life out of your organization’s IT infrastructure.  The DA too lkit can help you to understand how the strategies described above fit into your overall agile delivery process.

Related Resources

• The Spilled Juice Analogy for Technical Debt
• Data Technical Debt: How to Address Quality Problems in Data Sources
• Do Agile Teams Pay Down Technical Debt in Practice?
• Crushed by Technical Debt? (Webinar, earn 1 PDU)

A common question we get regarding Disciplined Agile (DA) is “What makes DA more disciplined than other approaches to agile?”  It’s a fair question, particularly from someone who is new to DA.  This blog posting explores this question, explicitly summarizing the critical strategies that exhibit the greater levels of discipline in DA as compared with what we see in many agile teams today.

It is clear that many common agile practices require discipline.  For example, agile teams it takes discipline to hold concise, streamlined coordination/Scrum meetings; to consistently deliver business value every iteration; to test continuously throughout the lifecycle; to improve your process “in flight”; to work closely with stakeholders and many more things.  Discipline is very important to the success of agile teams, see The Discipline of Agile for a detailed discussion, and DA takes it to the next level in the following ways:

1. Reducing the feedback cycle. Techniques that shorten the time between doing something and getting feedback about it are generally lower risk and result in lower cost to address any changes than techniques with longer feedback cycles.  Many of these techniques require agile team members to have new skills and to take a more disciplined approach to their work than they may have in less-than-agile situations.  There are several common ways to shorten the feedback cycle that are common to agile software development that are adopted by DA.  These techniques, listed in order of immediacy, include non-solo development (e.g. pair programming), active stakeholder participation, continuous integration (CI), continuous deployment (CD), short iterations/sprints, and short release cycles.
2. Continual learning. Continual learning is an important aspect of agile software development in general, not just DA.  However, DA explicitly addresses the need for three levels of learning: individual, team, and organizational/enterprise.  It also addresses the need for three categories of learning: domain, technical, and process.  Continual learning strategies include active stakeholder participation, coaching, mentoring, individual learning, non-solo development, experimenting via minimum viable products (MVPs), spikes, retrospectives/reflections, sharing lessons learned between teams, and stakeholder demonstrations.
3. Incremental delivery of consumable solutions.  Being able to deliver potentially shippable software increments at the end of each iteration is a good start that clearly requires discipline.  DA goes one step further and advises you to explicitly produce a potentially consumable solution every iteration, something that requires even greater discipline.  Every construction iteration your team requires the discipline to create working software that is “done”, to write deliverable documentation such as operations manuals and user documentation, to address consumability (usability), to consider organizational change issues pertaining to your solution, and operations and support issues (an aspect of DevOps).
4. Being process goal-driven.  The DA tool kit promotes a process goal-driven approach. For each goal we describe the issues pertaining to that the goal. For example, with initial planning you need to consider issues such as the amount of initial detail you intend to capture, the amount of ongoing detail throughout the project, the length of iterations, how you will communicate the schedule (if at all), and how you will produce an initial cost estimate (if at all). Each issue can be addressed by several strategies, each of which has trade-offs. Our experience is that this goals-driven, suggestive approach provides just enough guidance for solution delivery teams while being sufficiently flexible so that teams can tailor the process to address the context of the situation in which they find themselves in. The challenge is that it requires significant discipline by agile teams to consider the issues around each goal and then choose the strategy which that is most appropriate for them.
5. Enterprise awareness.  Whether you like it or not, as you adopt agile you will constrained by the organizational ecosystem, and you need to act accordingly.  It takes discipline to be enterprise aware and to work with enterprise folks who may not be completely agile yet, and have the patience to help them.  It takes discipline to work with your operations and support staff in a “DevOps” manner throughout the lifecycle, particularly when they may not be motivated to do so.  Despite the fact that governing bodies such as project management offices (PMOs), architecture and database authorities, and operations may indeed be a source of impediments to your DA adoption, or they may be your greatest allies, these authorities serve important functions in any large enterprise.  Therefore a disciplined approach to proactively working with them and being a positive change agent to make collaboration with them more effective is required.
6. Adopting a full delivery lifecycle.  Despite some agilists reluctance to admit that projects go through phases, DA explicitly recognizes that they do.  Building serious solutions requires a lot more than just doing the cool construction stuff.  It takes discipline to ignore this rhetoric and frame your project within the scope of a full delivery lifecycle.  The two project-based life cycles explicitly depict pre-delivery activities, a three-phase delivery lifecycle, and post-delivery activities (operations and support).
7. Streamlining inception activities.  The project life cycles include an Inception phase, known as Sprint 0 in Scrum, where you perform fundamental project initiation work.  Unfortunately in our experience we have seen many organizations treat this phase as an opportunity to do massive amounts of upfront documentation in the form of project plans, charters, and requirements specifications.  Some people have referred to the practice of doing too much transitory documentation up front on an agile project as Water-Scrum-Fall.  We cannot stress enough that this is NOT the intent of the Inception phase.  While we provide many alternatives for documenting your vision in Inception, from very heavy to very light, you should take a minimalist approach to the Inception phase and strive to reach the stakeholder consensus milestone as quickly as possible.  If you are spending more than a few weeks on this phase, you may be regressing to a Water-Scrum-Fall approach.  It takes discipline to be aware of this trap and to streamline your approach as much as possible.
8. Streamlining transition activities.  In most mid-to-large sized organizations the deployment of solutions is carefully controlled, particularly when the solutions share architectures and have project interdependencies.  For these reasons release cycles to your stakeholders are less frequent that you would like because of existing complexities within the environment.  However, the ability to frequently deploy value to your stakeholders is a competitive advantage; therefore you should reduce the release cycle as much as possible.  This requires a great degree of discipline in areas such as pre-production integration and deployment testing; regular coordination between project teams and with operations and support staff; Change management around both technology and requirements; and adoption of continuous deployment practices to such a degree that very frequent deployments are the norm and the Transition “phase” becomes an automated transition activity.
9.  Adopting agile governance. It is easier to avoid your traditional governance and tell management that “agile is different” than it is to work with your governors to adapt your governance to properly guide the delivery of your agile teams.  Every organization has a necessary degree of governance and there are ways to make it especially effective on agile initiatives.  It takes discipline to work with your governors to help them understand how disciplined agile teams operate and then discipline to accept and conform to the resulting governance process.
10. Moving to lean. For all of the process goals we describe a range of options to address the issues pertaining to that goal.  These options ranged from traditional/heavier approaches that we generally advised against except in very specific situations to agile strategies to very lean strategies.  Generally, the leaner the strategy the greater the discipline it requires.

Adopting a disciplined approach to agile delivery requires the courage to rethink some of the agile rhetoric and make compromises where necessary for the benefit of the “whole enterprise” and not just the whole team.  In our experience most agile projects make certain compromises that are not classically agile in order to get the job done.  Rather than hiding this and fearing reprisals from those who would accuse you of regressing to a traditional approach, it is better to have the courage to take a pragmatic approach to using agile in your situation.

Effective application of DA certainly requires discipline and skill, but in our experience the key determinant of success is the ability and willingness of the team to work well together and with  stakeholders, both within and external to the team.  

A common question that we get is whether it’s possible for a team to take an agile approach in a regulatory environment.  The answer of course is a resounding yes, although your approach will need to be tailored to reflect the constraints of the applicable regulation(s).

Let’s explore issues pertaining to compliance:

1. The regulations vary.  Not all regulations are created equal.  For example, financial regulations such as Sarbanes Oxley (SoX) are typically less stringent than life-critical things such as the various Federal Drug Administration (FDA) regulations.  So, one regulatory compliancy strategy does not fit all and your team will instead need to tailor their agile strategy to reflect the applicable regulations that you face.
2. Agile teams are working in a regulatory compliance scenarios.  The quick answer is yes.  As you can see in the chart above, the 2016 Agility at Scale study found that two-thirds of agile teams face either regulatory, organizational, or both forms of compliance.
3. Organizations are succeeding at applying agile within a regulatory regime.  The 2012 Agility at Scale study found that some respondents indicated that their organizations had successfully applied agile strategies with regulatory situations. As you can see in the chart below they are applying agile in all types of regulatory environments, including but not limited to life-critical and financial.  If other organizations are succeeding at doing so perhaps yours can as well.
4. Organizations are failing at this too.  The 2012 Agility at Scale study also asked if organizations had agile project teams that failed within regulatory situations and respondents indicated that they had.  If other organizations are struggling with agile and regulatory compliance then yours might too, so please consider the advice provided below.
5. The regulations rarely tell you how to work.  Regulations typically provide criteria that your process needs to meet.  For example they may call out the need to have independent testing, but they won’t say that you need to have an onerous testing phase nor that all testing needs to be done this way.  There you could adopt parallel independent testing in addition to your whole team testing efforts to conform to this requirement.  The implication is that you can tailor your solution delivery process to be as agile as you can while still being compliant – you don’t need to take a waterfall/V-model style approach.
6. Sometimes compliancy is self imposed.   Some compliancy requirements are not legislated, such as FDA and SoX, but are instead willingly adopted by your organization.  Examples of this include compliancy regimes such as ISO-900X and CMMI, strategies which may have been adopted for marketing reasons (typically by IT service providers) or perhaps process improvement reasons.  As you can see in the chart organizations are both succeeding and failing at applying agile in these situations.
7. You need to read the regulations.  Our experience is that many organizations will let their more bureaucratic-leaning staff members interpret how to conform to regulations.  Not surprisingly their strategy often involves a lot more paperwork, activities, and checkpoints than is actually needed.  When pragmatic people are asked to interpret regulations you often end up with a more pragramatic response.  So, if you’re in a regulatory environment we’ve found that it behooves you to take the time to read the regulations so that you can streamline how your agile team addresses them.  Fair warning: Most regulations are incredibly dry reading.

Disciplined Agile Delivery (DAD) addresses regulatory compliance issues via several key strategies:

1. Adopt a hybrid process.  DAD is a hybrid tool kit that adopts strategies from a range of sources including Scrum, XP, Agile Modeling, Kanban, Unified Process, and many more.  Regulations typically cover a wide range of issues and as a result you need to adopt supporting practices from numerous sources.  This may include management practices from Scrum, agile development practices from XP, agile documentation practices from Agile Modeling, data quality practices from Agile Data, and so on.  The DA toolkit has already done the heavy lifting for you by showing how these practices fit together, unlike methods such as Scrum which leave this work up to you.
2. Adopt a full delivery lifecycle.  Most regulations address the full delivery lifecycle, not just construction.  DAD supports a full delivery lifecyle, in fact it supports several such lifecycles (a Scrum-based lifecycle, a lean lifecycle, a continuous delivery lifecycle, and so on) to reflect the differing contexts faced by teams in typical enterprise environments.
3. Focus on solutions, not just software.  Disciplined agile teams produce consumable solutions, not just “shippable software”.  DAD recognizes that delivery teams are working on solutions that have a software component, that run on hardware, that are supported by documentation, and that the team may even change the business process around the usage of a system and even the organization structure of the people using it.
4. Take a goal-driven approach. Recognizing that solution delivery teams find themselves in unique situations, DAD doesn’t prescribe how they should work.  Instead, it focuses on providing advice for how teams can tailor their strategy to reflect that context of the situation that they find themselves in.  DAD does this by promoting a process goal driven approach.  This strategy guides teams through the process decisions that they’re making, some of which will be driven by regulatory compliance.  The DA tool kit has already done a lot of the heavy lifting regarding how to tailor your agile process to meeting scaling concerns such as regulatory compliance, large teams, geographically distributed teams, and other issues.  
5. Adopt an explicit governance strategy.  DAD has agile governance strategies built right in, including explicit light-weight milestones, metrics, named phases, and many other aspects of governance expected by many regulations.  Once again, DAD has done a lot of the heavy lifting for you.
6. Be enterprise aware.  DAD promotes the concept of enterprise awareness, the recognition that agile teams do not work in a vacuum.  This includes strategies for engaging with enterprise architects, how to deal with enhancement requests and defect reports coming in from operations, and how to work with other enterprise professionals.  These can be key issues to understand when tailoring agile to be compliant within an existing organizational ecosystem – your entire process needs to comply to the regulations, not just the development portion of it.

In short, yes it is possible to successfully follow a disciplined agile strategy given the constraints of regulatory compliance.  

A few people have commented that Disciplined Agile Delivery (DAD) promotes a wide range of practices, which they like because it makes their options explicit but which they also potentially dislike because there’s so many practices to choose from.  This then leads to the question of why do we need so many practices?  First, there are a lot of practices out there to begin with and our philosophy is to help people know that they have options and help them to select the right ones.  Second, our experience is that for a practice to be easily consumable it should be:

1. Small.  Small practices are generally more straightforward than larger practices.  As a result they’re easier to understand and to adopt.
2. Cohesive.  A good practice addresses one issue and one issue only.
3. Loosely coupled.  Good practices, like good software modules, have minimal dependencies on other practices.

Practices that are small, cohesive, and loosely coupled are easier to configure into more interesting solutions.  For example, the practice of test-first programming (TFP) is combined with refactoring to form test-driven development (TDD).  The practice of (writing) executable specifications can be combined with TDD, or TFP for that matter, to give you behavior driven development (BDD) or acceptance test-driven development (ATDD).  The combination of iteration modeling, model storming, look-ahead modeling, and BDD can give you a strategy for addressing emergent requirements and design during an iteration.

Of these three aspects, we’ve found that coupling has the greatest impact on your ability to tailor your approach to meet the unique situation you find yourself in.  Just like highly coupled software is difficult to maintain and enhance, processes built from highly-coupled practices are too.  For example, consider the way that Scrum describes product backlogs.  A product backlog is one of several strategies that agile teams may use to manage their work.  In the case of Scrum, the strategy is to prioritize requirements by business value and then focus on implementing the highest priority work at all times.   Unfortunately Scrum has coupled many important practices to the product backlog concept.  For example, initial requirements modeling is often referred to as populating the backlog.  Prioritization of new requirements and exploring upcoming requirements is referred to as grooming the backlog.   There are several potential problems to consider:

1. The term “populating the backlog” masks the fact that not only are you writing initial functional requirements (for example user stories or features) as part of your initial requirements modeling efforts you’re also sketching things (processes, screens, data structures, …), identifying non-functional requirements, holding modeling sessions, and many other things.
2. It makes it harder for people new to agile to understand how it works.  Think of it like this, what’s a more descriptive term, “populating the backlog” or “initial requirements modeling”?
3. It makes it harder to combine practices.  If you wanted to swap out the product backlog for something a bit more sophisticated, such as a work item list, or something leaner like a work item pool, what do you do with the practices coupled to product backlog?  Do you  rework “backlog grooming” to be “work item list grooming”?  Do you rework “populating the backlog” to be “populating the work item pool”?  Even if these things are easy to do, seems like needless effort to me.

In conclusion, we have found that adopting small, cohesive, and loosely coupled practices enables you to adopt and tailor a process strategy that better reflects the context of the situation that you face.    Not only is high cohesion and loose coupling great strategies for software, their great strategies for software process too.

Recently on the Disciplined Agile Delivery LinkedIn discussion group we talked about how people in an existing traditional role fit into a disciplined agile team.  The challenge is that in organizations new to agile will have people who take on traditional roles such as business analyst, programmer, solution architecture, tester, project manager and so on.  Yet DAD has roles such as team member, team lead, architecture owner, and so on that are different from traditional roles.  Clearly this is a “DAD adoption” challenge that we need to overcome.

At first you will build cross functional agile teams with the people you have available to you.  Agile teams are whole teams, which means that the team has sufficient skills to get the job done.  The implication is that you’ll initially build a team from the analysts, programmers, testers, … that are available and are willing to join the team.  As they say, you go to war with the army that you have.  At first an analyst is likely to focus on analysis activities, a programmer on programming activities, and so on because that’s what they know how to do.  Because they are working closely with one another in a iterative, incremental, and collaborative manner they quickly start to pick up skills from one another.  Over time these people who were originally specialized in one aspect of solution delivery become T-skilled generalizing specialists with a wider range of abilities.  This enables them to collaborate more easily and be more effective as IT professionals.

An interesting side effect of this is that as your team becomes more experienced in working in an agile way, and as team members gain a wider range of skills, the conversation shifts from “I’m a tester, so I’m responsible for testing” to “as a team how are we going to approach testing X?”  In other words, you start to focus on performing valuable activities instead of the roles responsible for doing so.  This is an important shift in mindset on your agile transformation journey.