Rewriting our National O/S Re-igniting Singapore's e-gov tech mojo.

I recently met a fellow geek for lunch. He was bothered by the Singapore government agencies’ heavy reliance on outsourcing and was unsure whether outsourcing can continue to deliver sustained value in the face of a rapidly-changing IT and consumer-device landscape. He wanted my thoughts on the best way to go about rewriting Singapore’s “national operating system”, on the assumption that funding wasn’t a constraint. I gave him an off-the-cuff response, but feel it deserves more thought and a proper analysis and response. Here goes nothing.

It is probably fair for me to opine that Singapore’s current e-gov initiatives, while functional, are not up to the mark on the user interface and user experience side when measured against other global-standard internet services that Singapore’s internet denizens have gotten used to. Our e-gov systems work but feel as if they were frankenstein-ed together; a patchwork of systems cobbled together by different vendors at different times, and defined by product teams with little emphasis on UI/UX. SingPass remains stoically a Java app interfacing with web and mobile browsers through a javascript layer; if SSO wasn’t secure enough, why hasn’t RSO (reduced sign-on) been considered and implemented? I have yet to notice any e-gov sites that are truly mobile-ready; our official e-gov website looks just like its PC browser equivalent when viewed through smartphones, while Gov.SG is only barely mobile-compatible, and feels like a WAP mini-site.

SingPass - Inspect Element

Screenshot of Inspect Element on a SingPass form field, showing the underlying Java app and JS layer.

It’s such a yucky experience that I try to avoid interacting with e-gov services unless I absolutely need to. The only e-gov services that (still) matter to me as a citizen and entrepreneur are NS, CPF, IRAS, OneMotoring.sg, MOM and ACRA. When I do need to access those e-gov services, I am forced to set my phone aside and reach for my PC browser to avoid tearing my hair out. Sometimes, Chrome doesn’t work for some e-gov services, so I whip out Safari or even Firefox to see if any of them work better than the dreaded IE Explorer.

Somebody, kill me please.

1. Cause

As always, Google has the answers. I found a comprehensive account of Singapore’s computerisation and e-government history at an NUS wiki page. I stumbled across a recent report published by the Centre for Liveable Cities, titled “Singapore’s Evolution into a Smart Nation“. The 3 panelists are familiar figures to me; Khoong Hock Yun was and still is a senior management figure at IDA when I started on my first job. David Lim was still an Acting Minister then, and Dr. Tan Chin Nam was Permanent Secretary of MICA, all back in 2006. The report brings us down memory lane to Singapore’s 1980 National Computerisation Plan, and the 5 IT masterplans that followed. What stood out for me in the context of this blog post however, was the following comment by Dr. Tan:

Dr Tan responded to Mr Lim’s earlier comment on building internal
technical capacity, and explained that for a long time the Singapore civil
service “had capacity and know-how to build systems.” Today, with
cheaper resources elsewhere, Singapore—and not just its civil service—
has a stronger focus on being IT “architects” rather than “builders” of
technology. Possibly, there is need for a new balance, in order to have
capability within to drive consultants more efficiently, Dr Tan said. He
believes the key challenge is in figuring out how to perform analytics and
use data to make life better.

– Builders vs Architects of Technology

I think it’s a great perspective, but Dr. Tan is being really nice. As a consequence of spinning out NCS and the shift towards complete outsourcing, the government has lost nearly all of its technical talent and retreated to pure vendor management. Hindsight is always 20/20, but I’d like to posit that one pivotal decision and the Civil Service’s delay in reacting to its effects cascaded to present us with the Singapore e-gov predicament that we face today.

The decision by NCB to sell NCS to Singtel in 1997 for S$65M paved the way for complete outsourcing of the technical execution and delivery of e-gov systems and services. Its spinout was justified on grounds of (1) spurring market development for government IT projects to the private sector, (2) enabling the Civil Service to use a larger pool of resources from the IT industry (beyond just NCS) and (3) preventing any conflict of interest in allocating Civil Service IT projects. Having said that, I don’t think NCS has ever ceded its market leader position in e-gov projects to anyone else in Singapore. In 2013, NCS’ revenue contribution to SingTel’s topline was S$1.407B. Even with a conservative 3-times revenue multiple for NCS’ enterprise valuation with an approximate market cap of S$4.221B, we’re looking at Singtel pocketing a cool IRR of nearly 30% over 16 years, or nearly 65-times return on their initial cost of capital. But hang on, the largest shareholder is Temasek Holdings with 51.88% of Singtel’s share capital. That means, with some clever balance sheet accounting, that Singapore has privatised its key e-gov tech partner, awarded countless e-gov projects to them based on merit (it has to be!) and in the process recognised substantial investor drool-worthy “homerun-like” capital gains.

2. Effect

Everybody wins and no one loses…or so we think. As tech companies get larger, history has shown that they tend to lose agility, innovativeness and competitiveness. Kodak, IBM, Yahoo and Dell are great examples. I think NCS is no different, evidenced by the quality of their work, at least on the e-gov front. Talk to NCS and they’d blame the product owners in Civil Service. Talk to civil servants, and well…let’s not get there. In letting go of technical execution and delivery, Singapore’s Civil Service has not continued to develop its inter-disciplinary ‘muscles’ to both define and build user-friendly, modern e-gov products. It makes me think of the following scene in Wall-E.

Fat humans and Wall-E

Fat humans and Wall-E; will we still be able to walk on our own again?

 3. Full Stack Approach for Singapore e-Gov

Does it mean NCS has failed? No, I don’t think so. I don’t think anyone still believes in big bang IT implementations or wholesale rewrites. Take away NCS and other e-gov tech providers such as CrimsonLogic and all our existing e-gov systems and citizen services will grind to a halt. A more sensible iterative course of action is needed.

Over the past year in the Valley, startups have been observed to be undertaking a new strategy in their perpetual fight against incumbents, termed the “full stack” approach. Chris Dixon says it best. It’s apparent that our e-gov journey has forced us into a partial stack outcome that is suboptimal on 3 fronts:

  1. Bad product experience. Nest is great because of deep, Apple-like integration between software, hardware, design, services, etc, something they couldn’t have achieved licensing to Honeywell etc.

  2. Cultural resistance to new technologies. The media industry is notoriously slow to adopt new technologies, so Buzzfeed and Netflix are (mostly) bypassing them.

  3. Unfavorable economics. Your slice of the stack might be quite valuable but without control of the end customer it’s very hard to get paid accordingly.

Gasp! It’s a perfect analysis of the current state of Singapore e-gov! Poor UI/UX, lack of innovation and expensive IT outsourcing budgets due to the need for partners like NCS to make enough margins to justify their corporate existence to shareholders.

A full-stack approach for Singapore e-government could mean the following steps:

  1. Create a Private Limited company in Singapore that is guaranteed by the Singapore Government, akin to the quasi-private legal status of entities such as Infocomm Investments (wholly owned by IDA) or EDB Investments (presumably wholly owned by EDB). Attach the entity to a sufficiently influential and relevant government “body”, such as the recently launched Smart Nation Programme Office. Capitalise the entity with sufficient operating budget to…
  2. Assemble a full-stack team comprising of core management, product design, product management, web/mobile/backend software engineers and devops. Grow the team to 30+ pax by the end of year 1, 50 – 60 pax by the end of year 2, and hopefully to 80+ pax by the end of year 3. Start with at least 2 engineering teams, each led by a Director of Engineering, initially tasked to build…
    • A bleeding-edge national authentication system based on Reduced Sign-On (RSO) principles, with multi-screen multi-factor authentication, that is initially backwards compatible with SingPass (for easier adoption and less mainstream user confusion), and actively deployed and marketed to citizens who are issued their NRIC at the age of 15, and re-issued their NRIC at the age of 30. +1 to Isaac Tan for his suggestion to dovetail deployment with NRIC issuance.
    • An all-of-government micro-services backend as the foundation for a plethora of existing and yet-to-be-imagined g2g, g2b and g2c interactions.
    • An open gov data “middleware” stack, visualization and API, with selected reference adoptions by government agencies that experience high-frequency inbound interest from citizens and businesses.
  3. Embrace open standards and open source (where possible), with an emphasis on sharing of best practices and development and adoption progress with the community, like any startup would do for community management.
  4. More software, less hardware, although the team shouldn’t shirk away from creating custom hardware-software fusion that needs to remain proprietary, for security and architectural purposes.
  5. Be a people developer. Too many of Singapore’s best and brightest technology talent don’t find sufficiently challenging technical roles in our city state, and end up bleeding out of our ecosystem into Silicon Valley. It’s time we brought some of them back and retain some of our future youth.
  6. Maintain the culture of “Deploy or Die” as the e-gov skunkworks, and once sufficient deployment and adoption of engineered products has been achieved, its organisational structure and interfacing with the rest of Civil Service needs to be rebalanced.
  7. Avoid complete commercialisation, but find a way to quantify and distribute upside for the team, so as to be in a position to always attract the best talent that would otherwise be forced to pick working for this entity versus an equity-lucrative fully private tech startup.

All that remains is whether the powers that be agrees with this approach, and whether enough can be rallied to the monumental journey to rewrite Singapore’s “national operating system”. Beyond having an e-gov infrastructure and service layer that Singapore can be proud of, we should also consider whether such an initiative can catalyse Singapore to transition from a city state exporting its statecraft of urban planning and renewal, industrial parks and special economic zone development, to a less real-estate based and more Mozilla-inspired software powerhouse that facilitates stronger conversations and equitable commercial transfers in a triumvirate with hotshot ‘unicorn‘ startups and ‘siren server‘ MNCs about citizenry and people-object data.

Singapore can never have enough land, people and high GDP y-o-y growth, but we should certainly be trying to punch above our weight on the international scale by becoming more relevant digitally, not just to our citizens, but also to other cities in the world.

NB 1: Mozilla has >1K employees and US$314M in revenues in 2013. Contrast that with NCS and its >8K employees and S$1.407B in revenues in the same year. In terms of revenue per employee, rounding down to the nearest thousand, Mozilla wins this match-up at US$314K per employee while NCS is at ~US$137.5K per employee. In terms of reach, Mozilla and its suite of products is also a more recognised brand than than NCS ever was, and can be.

NB 2: It may have taken America 200+ years to give the rest of us Silicon Valley and its tirade of innovations, and a mere 50 years (4-odd times acceleration) for Singapore to get to where it is today, but Singapore and the American civilisation is relatively young by Father History’s norms. In comparison, it took humanity 85 million years for the human brain to evolve into the most amazing “computer” in this world, only for the very same humans to design and build computers 192+ years on (a whopping 442,708X faster) from the first computing device to get us to computing and software tech that can nearly mimic the human brain. Don’t believe me? Try taking this quiz yourself.

NB 3: In my opinion, politics cannot be the solution to Singapore’s future and may very well destroy us if left unchecked. Technology is the only way. The sooner we get started, the better our odds are to evolve Singapore past First Foundation and shoot for some form of Asimov’s imagined Second Foundation.

NB 4: special thanks to Chang Sau Sheong and Sam Hon for proofreading my article.

About James Chan

James Chan is an entrepreneur, investor, geek, photographer and husband/father based out of Singapore. Apart from frequent travels to Vietnam, Myanmar and Indonesia for work, James can also be found online via his trusty 15" Retina MacBook Pro or iPhone 6+.