Human-centered government: The path forward for public sector institutions to serve their people better

Why government and public-sector institutions are necessary to truly create large-scale social equity and change, but also why and how they have to innovate to fulfill this mission.

“The more apparent it is that a system is nearing an end, the more reluctant people will be to adhere to its laws.” - The Sovereign Individual

Given the duality of my passions in the intersection of web3 and equity-centered work in education — I’ve been thinking constantly about the ideas of institutions in our society. On one hand, the very pathos of crypto tends towards the belief in a trustless society. In a trustless society, governance is decentralized to smaller self-interest groups, who engage and exchange information through algorithmic smart contracts. The idea is that these fragmented, smaller, self-interest groups are able to understand each other’s needs, interest, and behaviors in a much more targeted way and thus are able to serve their constituents more effectively. In its most grand, ideal form — all forms of corruption and collusion disappear as the system of governance and decision-making is fully transparent, immutable, and truly owned by the people. As The Sovereign Individual predicts in an outcome of the Information Age (mind you this book was written in 1997):

“The process by which the nation-state grew over the past five centuries will be put into reverse by the new logic of the information age. Local centers of power will reassert themselves as the state devolves into fragmented overlapping sovereignties.”

On the other hand, this vision of grandeur is not our current reality. While it is vital for technologists and innovators to constantly rethink about the future possibilities of the world in a long-term manner, there is an even greater sense of urgency for us to solve the inequities and injustices that are faced by our current world. This is a world of rising income inequality and access, of political polarization and racial divides. While many private sector organizations and startups tout to “change the world” — the truth is that their incentivization model is less affected by a desire for equitable, social change but rather the demands of the market and investors. And it’s no fault of their own: startups are in a struggle to survive through their run-rate, and the only means of existence is by showing metrics of rapid growth in further venture rounds, and public companies alike are incentivized by shareholders to harness greater returns. Humans and institutions alike are fueled by incentives. And this is why our public school teachers, Black and Brown citizens, the homeless — are often not served appropriately if at all. And when they are by the private sector (e.g. the desire to “serve MSMEs”) it’s often to fulfill a narrative — one that can just as easily change to the whims of the market when it’s deemed infeasible or unattractive to pursue. Through my time at Teach for America (which was a non-profit), I’ve seen firsthand how the “exit” of resources (in this case, when Corps teachers finish their two-year teaching commitment in an under-resourced community) potentially results in communities being left worse off than before. In the case of service towards the underserved, the mindset of “move fast and break things” does not work. Granted, there are private sector companies that I believe truly have created lasting social impact through innovation: Go-Jek effectively increased the minimum wage in Indonesia by introducing large-scale demand for gig workers across the nation, ServiceTitan through their SaaS platform allows small, home-services businesses to operate and scale their businesses more effectively. However, this tends to be an exception rather than the norm, and it’s not optimal to rely on the private sector for large-scale, equitable change given that this is not their main priority, but rather a side-effect of business itself.

The implication of this is that governments and public/social-sector institutions are still critical to solving the world’s hardest problems today for the most underserved constituents — especially problems that are unlikely to be served by the private sector in the short to medium term, and even in the long run. However, the way that most people (especially in the United States) perceive governments and public sector institutions in the spectrum of competence and ethical is at the bottom of the barrel. That is to say, these institutions are perceived to be the least competent, and the least ethical. This fundamental disconnect is important to bear in mind for public/social institutions, as transformational social change can often only happen through trust and belief among parties that the change is for the good of everyone involved. If we were to believe that governments and public-sector institutions are necessary and vital parties for social change — we must also put on a critical lens to understand what is necessary for them to build public trust and to be competent organizations in the creation of equitable and just infrastructures for society.

In this article, I’ll zone in on three particular aspects — taken from the strengths of the most disruptive private-sector organizations today — that can be utilized to transform governments / public-sector organizations to become inherently more human-centered: in particular these are the principles of design, data, and delivery (these are aspects that I’ve taken from the Obama-recommended book Power to the Public that I will further expand on with my own personal research and experiences). Furthermore, there are also fundamental pre-requisites that are required to institute the transformation of government / public-sector institutions to be inherently more human-centric — that is stakeholder buy-in and the fostering of talent (adapted from the book Empowered by Marty Cagan). Without these prerequisites met, these changes are doomed from the very beginning. I’ll further highlight these principles through my own experiences working in the social sector through working for Teach for America and running my own education non-profit Cornerstone Education, as well as working in IDinsight, an impact and measurement non-profit working to empower public and social-sector organizations with analytics capabilities to quantify the “impact” of their initiatives (see also: Poor Economics by Esther Duflo) — moving towards a model where the highest-impact non-profits are funded vs. the old model of “donate and hope” that used to exist.

There’s also a disclaimer that needs to be made. While human-centered innovation is often seen as the core foundation of the technology sector — it does not imply that technology is the backbone of these efforts. Human-centeredness is a mindset shift in organizational culture that can be enabled by technology, but is not centered on technology. Oftentimes, even in the private sector, we are enamored when we hear technologies like “blockchain” or “artificial intelligence” be applied to problems — in many cases without even understanding what the problems are in the first place. Oftentimes, technology is not the right solution. This article will illuminate the principles of human-centered government without necessarily relying on technology as the end-all-be-all. It’s important to have this critical lens on especially going forward in our lifetime as future leaders / problem-solvers as digital technologies will continue to gain increasing presence and hype in our lives (e.g. Hooter’s parent company Chanticleers issuing a blockchain-based rewards program in 2018 for their fast-food chains, resulting in the stock soaring 41% at the time of the announcement to $4 and now worth just under 50 cents). Whether you’re an investor, public sector or non-profit leader — you must be able to cut through this rapidly increasing noise and understand whether or not technology is needed; and if so, if it’s being used in the right way to solve the root problems.

Obviously, all of these thoughts are oversimplified, and there are tremendous nuances and obstacles involved in instituting these changes to make government and public/social-sector organizations more human-centered. The incentivization model currently existing today for government to truly be “in service of the people” is arguably also deeply flawed. However, my hope is that this can serve as a mental model for those of you working in these institutions today and those of you who are interested in pursuing a future career here. Let’s dig in.

I. Design

There are two stories (illustrated from the book Power of the Public) I’ll share that demonstrate the imperative need for human-centered design to be a core principle of government policy and programs:

1. The bailouts (CARES Act) by the US Government distributing ~$4 trillion in aid, in response to the COVID-19 pandemic, had more than half of the money being distributed towards big businesses and institutions less affected by the pandemic and others that laid off thousands of workers (https://www.washingtonpost.com/graphics/2020/business/coronavirus-bailout-spending/). Stimulus checks to individuals faced difficulty in getting to the most vulnerable populations — as the distribution involved a direct deposit to bank accounts and previously submitted tax returns (of which ~10M Americans do not have). Six months after the initial distribution of CARES stimulus checks, there were 9 million Americans who were eligible for the checks but had not yet cashed them in. Similarly, emergency funding for small businesses were quickly evaporated and claimed by institutions that did not demonstrate rigorously a need for the resources. Economists, policy advisors and legislators spent significant efforts to theorize the economic effects that would ripple from the act — but was enough effort spent to understand the physical delivery of the stimulus on ordinary and vulnerable Americans and small businesses?

2. The process residents of Michigan had to go through previously to apply for emergency assistance funding involved a 1,200 question form called the DHS-1171 and a prayer to fate for the approval of the funding. The process was long and criteria for approval was nebulous. Civilla — a design consultancy — took the form on as one of their first projects. As their first step, they tapped into their networks of social sector organizations to meet with Michigan residents who previously applied for the funding. They scheduled 90 minute to 2 hour sessions with the form’s users — deeply understanding their physical and emotional state while filling the form out. Tapping into networks again, they managed to schedule a meeting with state officers. In a rather unconventional way, they sat the officers down and made them fill the old form out within a certain time frame. None of them could finish filling it out on time. In fact, many of these state officers have never even seen this form live before. By the end of the session, they were bought in that this was a real problem to solve. By developing this radical empathy from the ground and applying on-ground design principles, Civilla took the first step in designing the new form which ended up a fraction of the time for applicants to fill out and for state workers to process.

Oftentimes, especially in hierarchical institutions like government, the top-down, hypothesis-driven approach to problem solving is what is deployed — regardless of who the end user is. In fact, in most cases, this human-centered design, “user first” mindset is not even able to be empowered because of the underlying organizational structures that are modular in nature vs. end-to-end. For example, government organizations often deploy contractual consulting teams or agencies who have siloed insight into the customers and processes — often taking days or even weeks to be able to retrieve or ask for the necessary information (I’ve seen this personally even in the private sector as a management consultant!). Or as Marty Cagan put it Empowered, teams are siloed into feature teams that handle parts of a product rather than the cohesive customer experience. Thus, the accountability to be customer-centric in the design process is more often than not wiped out. It is important to consider first whether the underlying structure of the organization and how it does things is an enabler to human-centered design thinking.

Back to the top-down, hypothesis-driven thinking of the government — many problems in the past used to be solved by economists and statisticians who would theorize (like in the example of the COVID-19 CARES Act) about the potential pros/cons to decision-making and policies. Models would be churned up, and assumptions would be made based on “educated guesses”, but there is often a disconnect to the real-life implementation of these strategies (Nassim Taleb has a great chapter in his book Skin in the Game devoted to this: https://medium.com/incerto/the-intellectual-yet-idiot-13211e2d0577). High-level problem solving has its place, but it has to be met with the same level of rigor as understanding problems from the ground and understanding what it takes to fully realize these policies. In contrast, when Civilla took up the task to redesign the DHS-1171, a considerable amount of effort was spent to connect to deeply understand stakeholder needs, to map out their user journeys and emotional states, to engage stakeholders involved in the process such as state officers, and to design a product based on that deep, on-the-ground understanding from the constituents they sought to serve.

Governments and public sector organizations must tighten the feedback loop between policymaking and policy implementation. Important details are lost when you look at things from an ivory tower perspective — which was brought to life when there were even state officers (responsible for decision-making) who did not even know what the DHS-1171 form looked like. When creating policy, the responsibility should not be abdicated from policymakers and decision-makers— they must be able to understand in their bones what it means for the policy to be delivered on the ground, the trade-offs to consider when it comes terms to design in terms of end-user impact, and should have the radical empathy to understand what the decisions entail for the users themselves.

There are four main, actionable implications that design has on the future of human-centered government:

1. Governments and public sector institutions, in designing policies or programs, should have a set of available reference customers or a pool of people that serve as their “customer research” group — based on the actual target market of their policies or programs. In product organizations, this group called the reference customers, are a small group of folks that represent actual customers who’s experiences are directly relevant to the product in design. When redesigning the form DHS-1171, we see that one of the first steps Civilla took was to tap into their connections of social sector organizations to find their reference customers. Government is often so disconnected and institutionalized that the “users” are often theoretical. This should be the first order of business.

2. Talent within the team should be transformed to include people who have two abilities: (1) to deeply understand the principles of human-centered design and how it relates to creating human-centric companies and products, and (2) people who are adept at coaching and evangelizing this process of human-centered design to the rest of the organization (many of whom are stuck in status-quo top-down, waterfall based process of building products). Oftentimes technology or design people are frustrated with this obstacle in public sector organizations. However, there are folks that realize that coaching is an important skillset and a muscle-to-build to enact deep change. It’s important to find these people.

3. Organizational structures should be designed to allow empowered, design-centric teams to be enabled to build for their users. For instance, in the example of the feature team (the norm in public sector organizations — where one team is in charge of a very specific part of the process), teams can be reorganized to have accountability over the end-to-end experience of their users. In this way, they see optimizing the user journey as a whole vs. just independent parts of the system which can quickly grow in complexity if not accounted for within the context of the whole. Teams should embed designers as a critical part of the decision-making machine in order to develop this empathy as a core part of the organization vs. outsourcing it out.

4. Policymakers should be closely integrated with the policy implementation process by creating a system that encourages the development of radical empathy. When DHS-1171 brought in the state officers to fill out the form instead of just presenting a slide deck, what this allowed was a sense of genuine appreciation for what was wrong with the process and a sense of urgency of why it needed to be changed.

You might say — are there constraints in this? For example, in the case of COVID-19 and the CARES Act, Congress was moving around the clock to understand how to deploy measures to prevent the economy from completely exploding. Was there really time to be human-centric in situations like this? The answer is a resounding “no” in the current state of the government and public institutions — where there is rarely any available infrastructure to conduct rapid design research and experimentation. What is needed then, rather than throwing around words like “Agile” and getting consultants to work on discrete projects, is to truly integrate human-centered design as a cultural and process backbone of the organization, in everything it does. Thus, when the next COVID-19 comes (god forbid) there is an ability for the government and public institutions to move fast but not compromise the principles of human-centered design. The investment in talent, organizational design, reference customers, and stakeholder buy-in is essential when it comes terms to institutionalizing design.

II. Data

Like design, there are three stories that illustrate the need for proper application of data in a public / social-sector setting to create product and programs that genuinely serve the needs of their constituents. I’ll share them below:

1. How do we get kids into the classroom? From a politician’s perspective — there’s a wide breadth of options available that all seem to be worth trying: conditional cash transfers, educating families on the benefits of education, eliminating fees, hiring extra teachers. Do you just try them all since we think intuitively that all of these “should” work? Certainly this is not how we would approach it from a business perspective. Randomized control trial experiments similar to the way modern medicine is created were run by J-PAL (an MIT-founded research lab for data-driven poverty action, who created this methodology) to measure a simple question: “What is the extra years of education created per $100 spent on specific initiatives on a population”. Through this (and obviously this is oversimplifying the details of experimental design), they causally proved that the usual suspects (scholarships, extra teachers) actually were marginally effective contributing 1-3 extra years of schooling per $100 spent, and the interventions that were the most effective — contributing 30-40 extra years of education were spending the $100 on de-worming (in communities where this was an issue) and educating families on the positive returns that could be had from education. This finding — rigorously backed by empirical evidence — led to a a non-profit Deworm the World to deworm 20-million school-aged children in 26 countries in 2009.

2. Crisis Text Line — a crisis intervention service (e.g. suicide, sexual abuse, eating disorders) — works by automating the process for seekers of help to get to talk directly to counselors via texts. Using real-time data models, CTL predicts when the structure / language of a text (NLP) suggests a higher risk of emergency, and prioritizes those messages. They are able to use this constantly updating data to prioritize increased staffing on times (e.g. 4AM) when the data seems suggests there is higher demand. There’s an element of design as well in the creation in the organization. Previously, crisis interventions happened through calls — which people feared / were uncomfortable to do. CTL, through talking to people on the ground, realized this and thus made a high-quality service available through text, making it accessible for vulnerable, low-income populations who they discovered were not apt to calling.

3. Data doesn’t have to involve sexy rigorous, experimental design or machine learning models either. In Rockford, Illinois, homelessness was one of the highest in the country — and a team was built with the mission of reducing it to zero, starting first with veteran homelessness. The biggest problem? There wasn’t any available data on homeless people. They didn’t know their names, their backgrounds, what got them there in the first place. Thus, what the team did was to first start by building a “unified list” — tapping into their networks of homeless shelters, hospitals, agencies to built a full understanding of the homeless veteran population in Rockford (which takes a lot of time!). Then, on a weekly basis, they would work through the comprehensive list of unhoused people and check up on their progress, even asking basic questions like “Who has seen this person this week?” and identifying the roadblocks for moving that person into housing. Through gathering this data and taking the time to spend time with them (again going back to design), they found out an extremely surprising and nuanced finding — that bus fares were actually a huge cause of homelessness (e.g. if transportation costs were out of reach, veterans with mental health issues would miss appointments and medication, and end up fighting addiction on the streets). This finding led to a simple policy in Rockford to permanently discount bus fares for veterans — leading to a measurable drop in veteran homelessness.

For those working in the private sector — data is already a mainstay of daily operations. “Big data”, “machine learning models”, and “continuous A/B testing” are terms that we hear every single day working in a startup or modern private sector organization. In the public sector, data that is used to make important decisions is often “point-in-time” (think of your census data that is collected only once in a while and may be subject to “drift” — or when the underlying data does not represent the current reality) or sometimes too large to be necessary in making the decisions we need to make (for example — the current state of data collection is unlikely to be as granular as the Rockford example when the team collected all of the veteran homeless population in the city by name and progress). The one great thing about COVID-19 is that we’ve seen governments and public sector institutions build infrastructure to collect real-time data in order to foster transparency and action. And for governments that do this well — it has been shown to work pretty well. For example, Singapore is able to track in real-time the spread of COVID and use that data to alert residents of certain communities in order to prevent further spread of the virus.

As pointed out in the examples above — it’s not just about sexy technology or even algorithmic decision-making. Data is not machine learning. It’s about impact, and the implications of that start with public sector institutions making the effort to institute a culture of “data first”. Like design, what this implies is that decisions are not made based on theory / global analogues but rather on ground-truths that are collected in a targeted way for the specific population in question. For starters, data collection needs to happen more regularly in order to even be able to build up the ability to extract insights. And similar to design, these insights need to be connected to policymakers and decision-makers in government and public-sector institutions. The right talent needs to be in the room (in this case data scientists or researchers) to be able to explain insights and accurately detail out the limitations of the data collected in order to not extrapolate potentially damaging insights that have bias embedded. And lastly, a culture of transparency and data-sharing needs to be instituted such that stakeholders — including the public — understand how decisions are made and other public or social sector institutions are able to utilize the data effectively to build on top of it. It’s guaranteed to be a slow, slow, slow process — given that the right infrastructure has not been built for this starting with collecting the right data — but it’s a necessary build out for governments to make the right decisions. We can’t just not do the right things because the right things are hard.

As informed citizens, we should also be expecting that accountability and transparency in data-collection and decision-making is instituted. This goes for governments and policymakers when drafting legislation, and even for the distribution of money to public / social-sector organizations. For example, now that we know that impact & measurement / experimental design methodologies are available to be able to rigorously measure “impact” — grants and CSR incentives should hold a component of measured outcomes within them. Building a data-first culture means that we’re actively questioning whether or not the right decisions are being made, and with what assumptions behind them.

However, it’s worth calling out one more fundamental prerequisite in terms of building up a proper data infrastructure. It’s that data — used wrongly — can be the source of immense damage. Existing biases that exist will only be amplified if the data collected is not representative of the population. There are two examples that highlight this:

1. In New York, an initiative was created by the Mayor’s Office of Data and Analytics in 2017 to fight the problem of rat infestation. By using data collected from a call service center — they believed — you could immediately zone in on the neighborhoods where rat occurrences were highest and focus the majority of efforts on them. However, when the head of data analytics looked into a low-income neighborhood he grew up in which had high-rates of rat problems, he noticed that the data seemed to say that there was no infestations in the area. Being confused, he looked into it further and it turned out people coming from low-income areas did not even know these call centers to report problems existed. Thus, the data that was collected was not representative of the truth because the data was selectively collected from higher-income areas where people actually knew how to utilize the call centers.

2. It’s widely known that in the 2016 campaign, publicly available data on social media (Facebook) was misused by campaign managers to promote exaggerated, often fake narratives and influence populations vulnerable to being misled. This was a private sector vulnerability that was exploited by the public sector — but a similar warning holds. Data can be misused to hurt and to deceive. It is not impossible to imagine security vulnerabilities in government / public institution data management to reveal information that should not be revealed. In Indonesia recently, an anonymous, “white hat” company suggested to have hacked a governmental database of vaccination results — obtaining sensitive information such as names and addresses from the system. It was fortunate that this was a “white hat” hacker organization that did not seek to use this data to hurt people, but you could see how this could go wrong very fast.

In instituting a data-first culture in the effort to create human-centered government — there must be efforts to ensure that the data itself is representative of the underlying constituents and that it is secure. Proper infrastructure must be built out to ensure of this, which is why investments in the right talent, and diverse talent that can hold different perspectives is extremely necessary. As Spiderman says: “With greater comes great responsibility”. Technology can be used to expose implicit biases or racism in our data collection, and perpetuate them.

It’s also important to note that there are certain things that data cannot fully measure concretely — but that still require investments in effort to. For example, it’s difficult to quantify racism, or broader discrimination as a whole. We can attempt to quantify them through proxies (e.g. hate crimes) but even these are often not representative of what we’re really trying to measure. In cases like this, it’s more important than ever that we don’t just use data as the end-all-be-all but also to strive to listen to and understand the voices on the ground.

III. Delivery

Finally, let me cap off the human-centered government principles with two stories that illustrate the importance of iterative delivery (one bad and one good example):

1. The US Immigration Services (or UCSIS for short) started to digitize the process of immigration in 2009 in an attempt to reduce the amount of time it took for applications to be approved. They hired a CIO from the private sector, Mark Schwartz, who sought to turn often monolithic government projects into “agile” — where they could quickly iterate on prototypes, learn, and develop further. However, his team faced numerous problems in the process. First, the entire concept of iterative delivery was completely foreign for the majority of staff who were used to working in a top-down, get it done in one-shot type of way, and in the evaluation process — the organization rated their efforts poorly stating why they couldn’t just get it finish it up. Having to comply with the expectations, the system was developed and released in one-go — but it turned out immigration officers hated it and it was no better than the paper process that embedded “hacks” that the officers were used to using (things like scanning and remembering certain bits of information when flipping through a paper application). The human-centered design process wasn’t followed because the expectation was just “to digitize”, and so after years and hundreds of millions of dollars of efforts, the system was just barely usable and often less efficient than the old system. Today, it is functioning although still not strongly, and contractor teams are still not always testing iteratively.

2. In contrast, when the integrated benefits office in Vermont aimed to redesign and potentially digitize the applications process for benefits, they started off with experimenting with a constrained prototype. The scale of the benefits system is massive, and so the team had a hypothesis that it would be best to start out small — by tackling a small part of the benefits system (food stamps) that did not have dependencies with other parts of the system in a 12-week phase. By the end of the 12-weeks the goal was simply to have a workable prototype that they could launch with existing users within the existing process. They eventually found that the prototype cut the time to receive a notice by more than half. With these results from the pilot, as well as understanding what they needed to change to make it better, they were able to take this and pitch to the Vermont government on a roll-out to other parts of the benefits system — instituting the stakeholder piece of “design empathy” by making legislators actually play with the prototype and understand how it worked. Through this, people were bought in and work began to start scaling this throughout the benefits system of Vermont with an understanding of the challenges that still needed to be faced.

Oftentimes, in public sector and large private sector institutions — there is a tendency to seek to build things in one fell swoop and “turn it on”. Teams are judged by how fast they can work vs. being judged on eventual outcomes. However, through these examples, we see the differences and risks undertaken between a monolithic, waterfall approach vs. a gradual, iterative, experimental approach. The monolithic approach, often driven by top-down decision making, can work but if not results in massive failure and waste of time / resources (in the UCSIS case years and hundreds of millions of dollars). The iterative, gradual approach — focused on moving softly but intentionally — focuses on testing key risks in the most minimal way. In the Vermont case, to pick a part of the system that did not have dependencies with other parts of the system, and to just test and learn what worked and what didn’t work. In this way, hypotheses are actually put into practice. In a world of available DevOps tools, and iterative experimentation infrastructure, there is absolutely no excuse not to do this.

Part of the work in order to build government and public-sector institutions that can work this in this fashion is trust (which we’ll cover in the next section). However, government leaders or public sector “entrepreneurs” who seek to create change and institute this type of infrastructure should really seek to start small in a way that minimizes disruption, to test rigorously, then to evangelize and build empathy to stakeholders and leaders that matter — just like in the Vermont case. Prototypes allow you to not only speak a vision but to be able to show it right in front of people’s eyes. That’s powerful. Oftentimes, especially countries that have had failed technology launches in the past would have leaders that are cynical and need to be proved. The plan is always to not terrify anyone, to be reassuring, and to move slowly.

IV. Fundamental Pre-Requisite: Stakeholder Buy-In and Talent

It is no secret that government institutions are often stifled in transformational change efforts due to two main reasons: (1) layers and relics of bureaucracy that are handed down due to decades of “precedent” that might not fit the current day organizational demands, and (2) the impossibility of “moving fast and breaking things”. In essence, the faults of governments are often the core strengths of disruptive startup organizations of today. Firstly, because governments don’t have an incentivization model built upon survival (public institutions will generally continue to exist even if they fail or make dramatically poor decisions), there is a natural gravitas towards the organization to maintain the status quo and to preserve the inherent structures that exist short of complete public upheaval and protest. The “old guard” of government who hold the most power and influence maintain organizational principles and rules that allow them and their “kin” to maintain their influence. This might be over-exaggerating it, but there’s definitely truth to this. The second is that government is inherently designed to be “responsible for all”. Unlike startup organizations who live and die by atomic networks and product-market fit, governments have to be harbingers of change that think of entire nations as their customers. As a result, governments must consider second/third/fourth orders of consequences in their decision-making and thus are as less able (although not completely unable to) to move fast. “Breaking things” is definitely a no-no — as consequences have the capability of inflicting large-scale damage (e.g. homelessness, bankruptcy, or worse) and completely eradicate public trust (of which we’ve observed the fallout of throughout history).

Thus, more than any other type of organization, government and public sector entities looking to institute transformational change must also know how to create champions in stakeholders from the top to the bottom who understand the importance of urgent change but also understand that a certain level of “failure” and experimentation is necessary in order for these changes to materialize. Let me share an example.

In Indonesia, Nadiem Makarim — the founder of Go-JEK (Southeast Asia’s first unicorn) — was appointed to be the country’s Minister of Education by President Joko Widodo. As seen in the conversation above, the President has given immense appreciation for the work that Nadiem is aiming to do, with creating technology products for teacher education and for creating data infrastructure to understand the individual needs of a diverse school system consisting of thousands of islands each with different needs. Because of this support, Nadiem was able to launch GovTech Education (https://cultureandtalentgovtechedu.medium.com/) — a team outside of the ministry filled with startup leaders and technologists dedicated to supporting the Ministry of Education on building top-priority, technology-enabled platforms. The jurisdiction for Nadiem to create human-centered technology products, as well as with the necessary funding to get hire strong talent, have been empowered by the President due to the trust that he has in this. It’s impossible to convince a country’s leadership or a state leadership otherwise.

However, top leadership are not the only stakeholders that matter. People throughout, who are living and breathing the effects of the policies everyday, matter as well. The most important challenge of Indonesia’s Ministry of Education to fulfill their mission is to educate and empower these stakeholders to understand the vision and the difficulties it will take to get there. These include teacher unions, the working teams of the Ministry of Education who are not used to working with technologists, and even families who are not used to this rapid change. If they are unable to do so — all of the foundational building blocks will come crashing down.

It is no secret that the work necessary to create human-centered government is difficult and represents a paradigm shift of operating for government and public institutions. The public sector is often seen as incompetent — but that has to change, and deep investments need to be put in into hiring, coaching, and promoting people who are able to drive the changes necessary. However, these talents cannot just be anyone extricated from the startup world. They need to be people that deeply care about the fundamental impact that human-centered government can bring, and people who understand that this is slow, but meaningful change — and have the long-term commitment and willingness to work with bureaucratic layers and “old-fashioned processes” to make this happen. It’s not everyone’s cup of tea. Thus far, Indonesia’s Ministry of Education has been doing a great job with this — matching private sector salaries and hiring executive talent from Indonesia’s top technology companies, including product, design, and engineering leadership from Bukalapak, Go-Jek, Kargo. This investment in talent, as you can imagine, must be met with equal belief throughout the spine of the government and public institutions. The worst case is that this breeds resentment, which can be destructive to progress.

V. Putting it all together and counterarguments

From this article, you see that creating human-centered government is not an easy task and requires not just great leadership, problem-solving, and organizational design to come together — but also is fundamentally limited or amplified by the support of stakeholders at the top who are willing to champion and fight for the principles echoed here.

Many of the aspects I pointed out in creating human-centered government is definitely an ideal state. As my friend Madeleine Setiono pointed out, policy and politics are often too intertwined — creating an incentive structure for politicians to pursue rent-seeking behavior where people end up carrying the deadweight loss. There’s not really an incentive to change this, as this precedent has often been carried out over decades. However, my belief is that in a world where people are demanding change and competence from government, there are tailwinds that exist that governments and public institutions have to adapt to in order to not only serve their constituents — but also to prevent widespread revolt. There is undoubtedly a sense of urgency that is felt for governments to have to innovate — and this will continue to get stronger and stronger.

Reiterating my fundamental belief in the top of this article: government and public institutions currently are the organizations — in the ideal state — that can truly affect large-scale social progress and equity. The work that is required for them to become organizations that can fulfill that mission is extremely difficult, but is immensely impactful. I encourage any of you reading who are passionate about building human-centered products in an impactful way to consider how you can get involved in this type of work.

For those of you who have read this article until the end, I’ll be making my move outside of Bain to pursue and to further the vision of human-centered government in Indonesia. I’ll be talking more about this in my upcoming article — so please stay tuned :)

VI. Further readings / sources cited in this article

1. Power to the Public: The Promise of Public Interest Technology by Hana Schank and Tara Dawson McGuinness

2. Poor Economics by Abhijit Banerjee and Esther Duflo

3. Handbook of Field Experiments by Abhijit Banerjee and Esther Duflo

4. Data Ethics: The New Competitive Advantage by Gry Hasselbalch and Pernille Tranberg

5. The Sovereign Individual by James Dale Davidson

6. Empowered by Marty Cagan

7. Skin in the Game by Nassim Taleb