Exclusive: Inside Hong Kong's Efficiency Unit

By Joshua Chambers

Interview with Kim Salkeld on using data to help an ageing population.

Government communication is a lot like jazz. As Miles Davis said: “It isn’t the notes you play, it’s the ones you don't”.

That’s the experience of Hong Kong’s Efficiency Unit. This team cut citizen complaints by 50% in one year just by taking out jargon and poor communications.

GovInsider caught up with Kim Salkeld, Head of the Hong Kong Efficiency Unit, to discuss his organisation’s use of data to improve communications.

Using data to cut costs

The Efficiency Unit runs a central government call centre, called 1823. “Initially, we began it because we thought we could deliver a more cost-effective service than if each department runs its own,” Salkeld explains. The call data has proved to be the greatest benefit.

Salkeld’s team used call centre data to find the most common complaints to government. It wanted to cut costs, and found patterns in the data. In particular, two communications campaigns were causing big problems for government.

First was an SMS message sent to students alerting them about financial assistance. This “was poorly worded” and badly timed, he explains, arriving before a full letter had been posted with details of the offering. “People were calling up because they thought they should have done something”. The team rewrote this message and “all of those categories of complaints fell away” - cutting complaints by 50% from that agency.

A second problem was a confusing, jargon-filled government letter, which “buried key information within the document”. The team used data and behavioral economics to revise the letter, simplifying the wording and pulling key messages up to the top of the page. Now “the person picks up the letter and they know what it’s about” - cutting complaints to that agency by 47%.

Salkeld’s Efficiency Unit technically benefits from increased complaints because “departments have to pay us for the service we provide them,” he notes. But the bigger objective is more important: reduced call times are “a huge benefit for the public,” he says, and “a financial saving for the department to spend elsewhere.”

Keeping communications clear

The learnings from 1823 are applicable to all digital services, Salkeld believes. For example: a great number of agencies release apps or portals that either don’t receive much use or result in “high frustration”, he says. Good user-experience design is fundamental for digital services.

The Efficiency Unit has been focusing on payment portals, which Salkeld thinks must be simpler and clearer to understand. “Take the burden off the citizen and give them something that behaves in the way they expect it to,” he says. He has a series of steps that will improve these systems.

First, officials must resist the temptation to show their knowledge. Citizens don’t want a complex discussion of the policy when they are using a service, and there are better forums for this information. Users are “feeling a bit nervous and stressed already… anything that gets in the way of their understanding is likely to provoke complaints.”

Second, the site should use numerical steps. “Rather than expecting individuals to read through lots of words, make the steps as simple as possible and the words as simple as possible,” he advises.

Third, “get away from words as far as possible”. Common graphics and symbols can help, ensuring people with different language abilities understand what is required of them.

Fourth, FAQs are helpful to citizens, and encourage civil servants to consider the users. It can be tricky for knowledgeable officials to “step out of that background and put themselves in the shoes of someone who is trying to approach government and simply get it to work,” he says.

Ageing demographic

Image: Daniel Lee, published under a Creative Commons License.

So far, we have discussed smaller reforms, but the big issue affecting Hong Kong is its ageing population. Today, there are 4 million people working and 1 million retired, Salkeld says, but in ten years’ time, 3.5 million will be working, 2.7m will be retired and over a million people will be over 80.

These changes will put pressure on the city, particularly in healthcare and elderly services. “Public services are going to need to be transformed. Since the overall workforce will be coming down, the public sector can’t really afford to grow without taking productive capacity out of the economy,” he says. “We’re going to have to think about how all existing services can be done in different ways.

Three broad changes are required, Salkeld says. First, government should “move as many people out of low-value backend jobs, and try to get as much of that automated as possible to free up people who can then be in frontline services.” Technology can help with elderly care, but it’s inevitable that people will need to be involved.

Automation can be done by using artificial intelligence and machine learning. Services can also be made more efficient by using design thinking that reduces the number of steps people must take.

Second, government must “free up the entrepreneurial capacity in the community to create new services that help the city work better,” he says. Government must share more data across silos, work together to tackle increasingly complex problems, and use analytics to measure the impact of interventions.

Third, the city itself must change. “It’s a city designed around people getting to work, where fewer people are going to work and want to enjoy the city to live and retire in,” he says. Housing estates will need re-designing, as will transport systems, and elderly care must not be confined to hospitals.

Technology can play a huge role, he believes, but only if there is a high level of trust. Officials must “engage with citizens directly rather than through the barriers or opinion polling,” he says, even when services are delivered digitally. “People want the trust that comes from public servants being willing to come into the community and meet them.”

Sensors can make an impact, but “there’s a real antipathy towards technology,” he warns. Citizens don’t want to feel like a product, with companies gathering their data and selling it on. “There is great potential to do great work with technology, but there’s huge potential to undermine autonomy and privacy and people worry about this - I think rightly.”

A changing role

The Efficiency Unit was founded in 1992 as a “small think tank” to think strategically and put together a “comprehensive plan of reform”, Salkeld says. Meanwhile, a separate Management Services Agency was founded to provide consultancy on big projects.

Over a decade, there were a number of huge reforms. Some government agencies were being spun-off into profit-making trading funds. Meanwhile, the entire public sector was undergoing computerisation, with efforts led from the very centre.

The think tank merged with the consultants in 2002, combining the strategic vision with practical delivery skills. But change never stops, and now Salkeld is seeing a shift away from “big bang” changes and towards “what we can do quickly, simply and well to build up confidence in larger changes.”

His unit is taking on the role of an innovation lab, he says, but it’s important innovation isn’t confined to central units. “If you put your innovation space wholly outside your delivery services, you get a gap where’s knowledge and then there’s new ideas. Lots can get lost at the interface.”

Salkeld’s unit is changing its approach to drive greater innovation across government. In particular, it’s using data to demonstrate the need for change, and building internal capacity to deliver quickly.

Throughout the ‘80s and ‘90s, the city was run like a large orchestra playing Romantic symphonies - hefty units would play big roles, conducted with a light touch at the centre.

Now, government needs to become more freestyle. There is improvisation and experimentation - with the Efficiency Unit as the rhythm section banging the drum for change.