A doctoral project should not spend years in private development only to reach a handful of people at the end. The work can begin helping practitioners much earlier, and their feedback can make the final project better.
That is the idea behind iterative research and distribution. You research a problem, share a useful piece of what you are learning, talk with people doing the work, test a change where appropriate and revise. Tools such as NotebookLM, Substack and YouTube can support that process. But the tools are not the point. The point is to learn faster and contribute something useful to a community of practice.
Key Takeaways
- Publish small, useful pieces of research to invite feedback before completing a final project.
- Combine action research, lean testing and practitioner relationships to improve each iteration.
- Share different versions of your work with different audiences to protect sensitive organizational information.
- Use AI to find credible sources and peers, then learn directly from practitioners.
Table of Contents
- Why Publish Before the Research Is Finished?
- Combine Action Research, Lean Startup and a Community of Practice
- Build the Community by Serving It
- Choose the Right Audience for Each Version of Your Work
- Balance Research Effort With Distribution Effort
- Use AI to Find Peers, Practices and People to Contact
- Use NotebookLM, Substack and YouTube for Different Jobs
- Make Every Course an Iteration
Why Publish Before the Research Is Finished?
Traditional academic research often resembles a waterfall project: conduct four to eight years of work, release the result and then receive feedback. An agile approach may produce smaller releases more quickly, but speed alone does not guarantee meaningful feedback from practitioners.
The approach I want doctoral students to consider is closer to lean research. Publish something small, learn from the response and produce a better iteration. That first piece might grow out of a course research project. It does not have to be a book, or even a polished account of everything you know.
Publishing for its own sake misses the point. It reminds me of someone who said he had ironed his pants even though the wrinkles were still there. The task was technically complete, but its purpose was not. Similarly, the goal is not to check a box marked “published.” It is to invite useful feedback and serve people who can apply what you are learning.
Combine Action Research, Lean Startup and a Community of Practice
Three ideas work particularly well together. Each answers a different question about how research can lead to change.
Action research gives you the cycle
Action research moves through planning, acting, observing and reflecting. You develop a plan, try something, pay attention to what happens and use that learning to plan again. A doctoral course can help you develop one part of the plan; your work in an organization may give you an opportunity to test a component of it.
That does not mean implementing every recommendation from every course. It means treating your plans as ideas that can be examined in practice, rather than documents that must remain untouched until graduation.
Lean startup helps you act effectively
The lean startup approach adds detail to the acting stage. Start with the minimum amount of work needed to test an idea. Build something, measure the response, learn and then decide whether to adjust course.
In an organizational setting, your “minimum viable product” might be a modest change to a program rather than a new product. Form a hypothesis about what that change could improve, make the change where you have the opportunity to do so and observe the result. The advantage is not simply doing things quickly. It is shortening the time between an idea and learning whether that idea holds up.
A community of practice helps you learn with the right people
A community of inquiry might consist of fellow students examining questions together. A community of practice brings together people engaged in similar work. If you direct programs at a rescue mission, other program directors at rescue missions are a natural group to learn with.
Such a community has a shared domain of learning, practitioners who care about it and expertise built through their experience. These are not merely potential readers for your eventual project. They are people who can challenge an assumption, identify a practical obstacle or describe an approach that has worked in their setting.
Put the three ideas together and the cycle becomes collaborative:
- Frame and plan together: Develop a question and a hypothesis with input from people close to the problem.
- Test and act together: Share a small resource or try a manageable change.
- Measure and observe together: Ask what happened and what practitioners noticed.
- Synthesize and reflect together: Revise the idea before beginning another cycle.
Not everyone in your community has to run your experiment. You might share an article with ten people and have deeper conversations with three of them. Those conversations can still substantially improve your next iteration.
Build the Community by Serving It
How do you begin a community of practice when no formal group exists? Start by being helpful. Find useful research, identify promising approaches and give practitioners resources that help them do their jobs. At first, they may think of you almost as a free consultant. That is a good place to start.
Relationships can grow through individual conversations, perhaps by Zoom. Later, when people share an interest and see value in learning from one another, a group conversation may make sense. There is no need to launch a large network before you have built trust.
Our action research pyramid begins with AI-assisted exploration, then moves toward highly cited peer-reviewed work, authoritative sources, model organizations and experts. Each layer can help you ask better questions of the next. The aim is not to stop at an AI-generated answer. It is to discover credible material and people worth engaging directly.
Choose the Right Audience for Each Version of Your Work
One of the largest risks in distributing doctoral work is sharing the right information with the wrong audience. A candid internal conversation can be essential to a strong project and inappropriate for a public article.
I find it helpful to think about the different circles in Jesus’s ministry: a close group of three, the twelve, a wider group of 120 and the crowds. The lesson for research distribution is to recognize that communication changes as the audience expands. Greater reach calls for greater care with boundaries and language.
Your closest circle might include people who review an assignment or doctoral project before submission. A somewhat wider circle could include a doctoral committee, an executive leadership team or trusted peers in similar organizations. Staff and leaders at peer organizations may be the best audience for a practical article. A public newsletter, podcast, YouTube channel or book reaches farther still.
Most of your effort may belong with the people nearest the work. I would put perhaps 80 to 90 percent of the effort into engaging these closer circles rather than rushing toward a broad publication. That is not a retreat from writing a book. It is how you give a future book the benefit of repeated feedback.
Do not force every course project to be suitable for public release. An internal project may need specific organizational details to be useful. Before publishing anything derived from it, decide what belongs in private discussion and what can responsibly be shared more broadly.
Balance Research Effort With Distribution Effort
Think of the work as a two-sided funnel. On one side, you bring in AI-assisted exploration where permitted, authoritative sources, peer-reviewed research, examples from model organizations and conversations with experts. Some projects may also involve formal quantitative or qualitative studies, though that is not required in this doctoral context.
In the middle are your course projects, literature review and doctoral project. On the other side are the ways you put the learning to use: resources for your organization, conversations with a community of practice, workshops, webinars, classes, articles and eventually broader channels such as Substack, podcasts, YouTube or a book.
It is possible to publish a book chapter as an article while a book is taking shape. But think about the book’s argument as you do so. Collecting unrelated blog posts into a manuscript does not automatically produce a coherent book.
I sometimes frame impact as research quality multiplied by distribution. That is not a formal measurement, just a useful reminder. Excellent research that reaches almost nobody has limited opportunity to change practice. Wide distribution cannot rescue weak research. You need to work on both sides.
Distribution also has a role in organizational change. Drawing on Everett Rogers’s work on the diffusion of innovations, a change agent helps people recognize a need for change, establish an exchange of information, diagnose problems and develop an intention to act. The work continues after an initial action: adoption needs to become stable enough that the change does not simply disappear.
Use AI to Find Peers, Practices and People to Contact
AI can be particularly useful when you are trying to understand what peer organizations are doing. A practical process begins before you write a prompt: define the kinds of organizations relevant to your question.
- Build a peer list. Start with a defined set of organizations or a clear category, such as rescue missions or Christian community development organizations. Organizations with substantial public information are easier to research this way.
- Scan broadly. Explore the domain across many organizations to see which approaches and questions emerge.
- Narrow the field. Identify the organizations with the most relevant examples, then examine their practices more closely.
- Identify people to contact. Use what you found to locate experts who can help you move beyond publicly available information.
For example, I have worked from a list of about 200 organizations to identify 15 with significant social enterprises. A follow-up question examined practices among those 15. The same approach can help investigate thrift stores, aftercare programs, work therapy or educational and vocational resources.
A predefined list makes the scope of the request clearer. You can provide organization websites and ask an AI research tool to examine those organizations, rather than assuming it will choose the peers most relevant to your context. Alternatively, you can start with a carefully defined category and narrow it after the broad scan.
The final step matters most: contact people. An AI-assisted overview can help you find promising examples, but conversations with practitioners are where you can ask how an approach works, what has been difficult and what you have misunderstood.
Use NotebookLM, Substack and YouTube for Different Jobs
The tools in an iterative workflow do not all do the same thing. NotebookLM can help you work with a collection of source materials, subject to your program’s rules. Substack gives you a place for articles and a newsletter. YouTube and a podcast provide additional ways to share ideas and invite engagement.
In our doctoral program, we want students to become capable of setting up and using these channels, even if they later decide not to keep every experiment public. A student should know how to establish a blog and newsletter, post a video and work with the research tools available to them. Competence comes from trying each one, not merely deciding in advance which sounds useful.
City Vision has developed a NotebookLM for each doctoral course and plans to provide it after students complete that course. Part of the reason is educational: students need to engage the assigned books themselves. Part is ethical: giving access to a notebook containing course materials before students obtain those books could undermine that expectation. Once available, the notebooks should remain useful beyond the doctoral program in ministry work.
There is an important distinction between researching with AI and researching without AI. Students are encouraged to follow the applicable course guidelines and may use course materials in their own NotebookLM notebooks where permitted. But NotebookLM and other AI tools should not become a way to skim assigned reading or write papers and assignments, apart from narrow exceptions in those guidelines.
Think of the difference between climbing a mountain and taking an electric bike up it. If the purpose of the work is to build your ability to read, reason and synthesize, outsourcing that effort leaves you less prepared for the next climb. AI may help with a permitted research task; it should not replace the learning the assignment is designed to develop.
Make Every Course an Iteration
An iterative approach does not require you to start a new public project every semester. It requires you to connect what you are learning over time.
Customize course research to your organizational context where you can. Ask which component of a plan you might realistically implement after the course. Observe what happens. Revisit earlier work when a later course, an expert conversation or an experiment changes your understanding.
Self-publishing may be appropriate for some resources, including certain AI-assisted research outputs when permitted. In most cases, do not publish course papers or final projects unchanged. They may contain details about your organization that were suitable for an internal assignment but not for public distribution. Make a deliberate decision about the audience and create a public-facing version when needed.
Finally, build a feedback mechanism that extends across the doctoral program. Aim for at least 50 conversations about your developing ideas with internal stakeholders, peers and outside experts. Try components of your plans when circumstances allow. Revise earlier material based on what you learn. Begin with individual meetings and let an informal community of practice develop gradually.
The result is more than a better final document. It is research that has been tested against real questions, improved by people who know the work and shared in forms that can help others use it.