In our race to scale, have we forgotten that implementation needs to be messy?

In my first year as a development professional, I worked on a project in rural Bihar studying the uptake of sanitary pads and menstrual cups. The plan was clearly laid out —distribute products, monitor usage, and draw insights. But what we hadn’t planned for was  the messiness of implementation: 

• Surprise 1: Families resisted menstrual cups for unmarried girls, fearing hymen rupture which could affect marriage prospects.

• Surprise 2: Many women didn’t wear underwear due to heat, affordability, point of purchase or norms—making sanitary pads a non feasible option.

These unexpected realities revealed a sharp disconnect between tidy programme design and the complex lived experiences on the ground.

In “Big Aid is Over,” Kevin Starr critiques a common trap in development work: designing for cost-effectiveness instead of cost-efficiency. When the focus shifts to reaching millions just to make the per-unit cost look reasonable, we risk losing sight of whether a programme is truly affordable—and whether anyone will ultimately be willing (or able) to pay for it.

But there’s another concern that I fear we may be overlooking: in our drive to scale and achieve cost-effectiveness, are we losing sight of the grainy, deeply local and human realities of implementation?

Scaling a programme often begins with clearly defining it—breaking down every element into its nuts and bolts, with standard operating procedures that can be replicated across contexts. The aim is to simplify delivery: a “programme in a box” approach that teams can be trained on and efficiently deliver anywhere.

But real-world settings rarely conform to such neat designs. Programmes are implemented across diverse cultural and social landscapes, where norms, beliefs, and lived experiences vary widely. A universal playbook, however well-intentioned, will inevitably struggle to account for this complexity.

Take India’s extensive community health workforce—the ASHAs and Auxiliary Nurse Midwives (ANMs). On paper, their roles are well-defined, reporting templates are standardized, and daily schedules are prescribed. But walk into the field, and the reality unfolds differently. Here are a few examples of how ANMs navigate the diversity and complexity of their work environments: 

• ANMs are tasked with distributing Iron and Folic Acid (IFA) tablets to pregnant women to prevent anemia. Though how the distribution is planned varies by local contexts. Some ANMs choose to provide tablets in bulk—aiming to reduce the follow-up burden on women and their families, especially in cases where they may know that women are seeking private care or would have difficulty traveling to the health centre. Others avoid bulk distribution, particularly in areas where prevalent norms around receiving large quantities of tablets at once diminishes their perceived value and potentially leads to reduced consumption.. These differences complicate efforts to design uniform supply chain protocols across regions—but they often improve actual adherence by aligning delivery with local norms and constraints..

• Similar variation is seen in how ANMs support early breastfeeding. In some cases, they discourage the presence of family members immediately after childbirth to prevent ritual feeding that may delay breastfeeding. In others, they actively encourage husbands and grandmothers to be present—recognizing their role in offering emotional support and enabling informed decision-making. These adaptations are shaped by a range of factors, including the mother’s age, whether it is her first pregnancy, which also influence the level of pushback the ANM will have to encounter from the family. Together, they reflect the discretion exercised by ANMs in navigating diverse and often sensitive local contexts. 

There’s clearly no single script. No one-size-fits-all. And this can be frustrating—especially for those of us designing programmes, trying to define clean processes, or generate tidy, comparable data across sites.But this is precisely where the strength of implementation lies: in navigating the nuance of complex humans.

People’s lives are complicated and context-specific. Meaningful change happens not in spite of that messiness—but because programmes are designed to work within it.

So does this mean scale is untenable? Not at all. But it does mean we need to design for scale without flattening the realities of implementation.

Here are some ideas on how we can navigate this: 

1. Support and enable grassroots teams. Those closest to the community often understand its nuances best. Beyond training, we need to consider how to equip them with the right tools, resources, and a degree of decision-making flexibility. We need to trust that once teams are grounded in the core values and goals of a programme, they are also able to tailor delivery in ways that make sense locally.

2. Reimagine how we measure. Standard metrics are important, but they shouldn’t come at the cost of capturing the real story on the ground. When reporting frameworks don’t reflect local realities, frontline workers may understandably find workarounds, which can affect data quality and learning. We should plan for our measurement systems to be robust, but also responsive.

3. Define what must stay consistent—and what is flexible Some programme components are non-negotiable, especially those linked directly to intended outcomes. But not every detail needs to be uniform and standardized. Such standardization in the effort to be efficient can actually alienate the communities we are aiming to serve. 

Accepting the messiness of implementation isn’t a compromise—it’s a recognition of what it takes to create real, lasting change. What do you see as the biggest barriers—or enablers—to embracing this kind of messy, grounded implementation?