Crisis Management with Kanban
The Untold Story of the IT Department Behind the Christchurch Earthquake disaster recovery effort
By David J. Anderson with Irina Dzhambazova?
People say that in the hour before an earthquake the sky turns gray, the air is heavy, and the birds fall silent. That hour was Christchurch’s lunchtime on what seemed like a regular Tuesday: gray skies, still air, and silent birds. The scene was nothing out of the ordinary during a summer’s midday in New Zealand’s second largest city. Nilgun Kulpe was on the fifth floor of his office building. John McCrone was on the phone. Sarah McCarthy was driving her car inside a multilevel parking garage. Mark Cornell was shopping. Kelly Boswell was working in the mall. Nathanael Boehm was walking down a street. Dr. Michael Ardagha was at the airport, waiting to board his flight. For all of them, at 12:50 p.m. that day, everything was pretty normal.
And then the clock hit 12:51 p.m. on February 22, 2011…
At first I thought it was another aftershock . . . but then the sound and force hit the city. . . . It felt like an incredible blast from under the earth. There was a quick swaying sensation, but then it exploded. A rumbling sound started up and within a second the entire street was flexing and heaving. Something crept up behind the domestic terminal, grabbed it by the shoulders, shook it, and then punched it in the head.
The noise was huge. Everything smashing, cracking, falling. Bricks rained past the window. It was like four strong men were on either side of the car and were shaking it up and down with murderous intent. It was swift and violent. The overhangs on all the shops immediately snapped off and dropped to the ground; and as the facades started to crumble and collapse, bricks, glass, and chunks of concrete were hurled into the street.
I was thrown around in my seat as I gripped the steering wheel. My vision narrowed, there was only the windshield and the buckling concrete ahead of me. The building kept going up and down and it was like being in a washing machine. I grabbed on to the counter to try to keep on my feet and ducked down as the tiles around me started to fall. The floor underneath me cracked, and the ventilation shaft fell from above. As soon as I stepped outside, the building directly opposite came down on top of people; the building next to that came down on people; the building to my right came down. I couldn’t run, there was nowhere to run and just trying to stay upright was like trying to stand on the back of a rodeo bull . . . I was convinced I was going to die.
For around 20 seconds, it was a disaster movie special-effects team working overtime, the end of the world was that convincing. The city was collapsing around me, people were being buried in rubble and others [were] being struck in the head by debris. It was followed by sudden silence and a rising pall of dust. The lights went out and it was dark. And then the screaming started. The screaming of people — something I have never heard. It was piercing, animal screaming. I guess it was the sound of people dying, and uncomprehending terror.
It was surreal, both horrific yet somehow impossible, and unbelievable, like a nightmare. The footage from that day is still haunting. Crumbled buildings, bodies covered in old towels, grown men crying — a city torn to pieces.
Canterbury, New Zealand, was never much of a seismically active region. No major earthquake had disrupted the 45,000 square kilometers of Canterbury’s land between August 31, 1870 and September 4, 2010. A few months prior to February 22, at exactly 4:35 a.m. on September 4, the stable, mid-eastern part of New Zealand’s South Island revealed its hidden nature with a smack that measured 7.1 on the Richter scale. It lasted less than 15 seconds. It woke up everyone from their sound sleep. It caused damage, it caused fear, and it caused disruption. But the epicenter was far enough away that there were no casualties. However, one thing that was shaken more than the Cantebrians themselves that early morning was a much closer fault, Port Hills. Unlike the distant fault that caused the September quake; Port Hills was located just under the city of Christchurch.
Many aftershocks came after the September quake. As unsettling as those were, Cantebrians believed that “the big one” was behind them, and that eventually everything would be restored to its former calmness. On February 22, they were proven wrong. What they experienced that day was measured as one of the highest ground accelerations ever recorded in an urban area. Its force at the epicenter made even gravity seem a weak force by comparison. That acceleration was four times stronger than what Japan would experience a few weeks later during the devastating March 11 earthquake that measured 9.0 on the Richter Scale and caused the Fukushima tsunami. The entire shaking episode lasted 24 seconds, more than half of which was especially intense. More than 300,000 people that day experienced shaking that was “severe” or “violent” as noted in the twelve-stage Modified Mercalli Intensity scale; the only higher classification is “extreme.”
No seismologist anywhere in New Zealand knew the Port Hills fault even existed. It had never expressed its presence (although it didn’t try that hard to hide); some of its upper layers were located just a kilometer under the streets of Christchurch. The fault’s shallow depth was a key explanation for why this 6.3 earthquake was so destructive. The Port Hills fault had been dormant for what was estimated as approximately 8,000 years — until February 22, when at 12:51 p.m. it ruptured.
In its vertical movement, the earthquake demolished lives and buildings at a speed of 3 kilometers per second. It took away 185 lives and destroyed tens of thousands of buildings, including most of Christchurch’s business district. Roads, bridges, power lines, water pipes, cell phone towers, and ordinary phone lines were not spared, either. Christchurch had been erected 160 years prior, but in 24 seconds a large part of that old town was wiped out. The price tag for those 24 seconds would total up to 13 percent of New Zealand’s GDP.
****
On February 22, 2011, at 12:51 p.m., the Beehive building[1] in Wellington also shook. This capital of New Zealand, located more than 400 kilometers away from Christchurch, is home to all of New Zealand’s ministries. Prime Minister John Key was in a meeting on the ninth floor and felt the quake distinctly. He thought it was a Wellington earthquake. It took him exactly three minutes to realize that it was instead a Christchurch earthquake. A chilling question immediately crossed his mind — how strong must the earthquake have been in Christchurch, his hometown? Upon arriving there a few hours later, the clouds of rubble he saw from his helicopter showed a picture that spoke far more than a thousand words.
Russell Healy was in a neighboring building that housed the Ministry of Social Development. The third-generation New Zealander was just finishing off his lunch break. The father of three young children felt a swaying sensation, but it was nothing too much out of the ordinary or too scary. Just as the Prime Minister did, Russell also assumed it had been a small quake in Wellington. It did not take him long to realize that he was wrong. The ordinary Tuesday had turned extraordinary for the entire nation.
Helping the injured was top priority. When the quake hit, the emergency department of one of the major hospitals was badly hit. Ceiling panels fell and trolleys migrated across corridors. The power briefly went out, but the backup power kicked in within a few seconds. When the dust settled, medical staff assumed control. Security camera videos show that as soon as the lights came on nurses and doctors scampered from cubicle to cubicle to check on their patients. Once they made sure that everyone was fine, they gathered and collectively implemented a carefully contrived and well-rehearsed plan.
While the injured were attended to, the government had another, much larger group of people to consider — the survivors who were unharmed on the outside but desperately needed a sense of normality to start healing the internal scars. Enabling them to return to work and earn their livelihoods was the most important step. Retaining them in Christchurch, making sure they did not move away, was another.
“I am lying in bed after the last 4.6 shake, I can’t get back to sleep [. . .] and I can’t stop thinking about how I am going to look after my family. I’ve missed out on a good amount of hours this week, I doubt anyone will be in a state to work in the coming weeks and if they do, will they really want to see a business mentor? I don’t have a wealth of cash reserves in the bank, primarily due to the impact of the earthquake in September, even though my business picked up from December. I needed more time to build up the cash reserves, [but] that time, along with my optimism, has now gone. I am feeling very alone. I don’t want to share my concerns with my wife, as I don’t want her to worry; I don’t want to burden my friends, they have enough on their plate. . .
These thoughts are all really selfish, I know there are still people trapped under rubble clinging on to life, the reality is, though, for me Tuesday was all about survival and the welfare of my family, Wednesday is now about how do I ensure my family survives in the weeks and months ahead. I am in a tailspin, I know it, I don’t like it, but I can’t stop myself.”
— Dave Sewell, a husband, a father of a young baby boy and a business management coach
In its immediacy, the February 22 earthquake was a huge physical event. In its aftermath, it turned into a major psychological shock. For many people, reality was suspended for an indefinite period. Many people must have lain in bed, just like Dave, in those early hours of February 23, unable to sleep. They had survived the earthquake itself but how was life meant to carry on? The Christchurch Business District (CBD), where many of them were supposed to go to work, was very badly damaged and was closed down by the authorities. Approximately 50,000 people employed by 6,000 businesses would have gone there that morning to earn their livelihood, but could not. Many more other buildings and businesses were severely damaged or destroyed in those 24 seconds.
The decision by Mayor Bob Parker to close the CBD and deny everyone access to the area came a few hours after the earthquake. Parker had been reelected a few months earlier as a direct result of his coolheaded handling of the September earthquake. The February 22 temblor had taken a toll on most buildings, and they needed to be inspected, some eventually to be demolished. What the main quake had not done, the incessant aftershocks took care of. The decision was harsh, but inevitable. Nobody was going to work there for many mornings ahead[2]. Documents, computer systems, and other business property were left behind the cordon. What the business owners had in front of the cordons were their stunned, scared, and battered employees.
The government had first implemented an earthquake support system after the September earthquake in Canterbury. Its purpose was to distribute financial aid to business owners with 20 or fewer employees who were unable to operate for a certain amount of time after the earthquake. The subsidy was meant to help them figure out how to proceed without making hasty decisions and without the worry of losing staff. It was designed such that the money would go from the employer to employees rather than as an unemployment fund; this way kept the psychological bond between an employer and the employee. To apply, employers had to fill out an online application on the website of Work and Income, a direct sub-agency of the Ministry of Social Development, and submit it in print to one of the Work and Income offices. In the week following the September earthquake, the government disbursed a total of 10.5 million New Zealand dollars to Canterbury’s small businesses.
On Thursday, February 24, Deputy Prime Minister and Finance Minister Bill English announced the creation of a much larger Earthquake Support Package to aid many of the employees and business owners who were unable to operate due to the earthquake damage. History would remember this announcement as the start of one of the fastest, most effective, and large-scale policy implementations to follow a natural disaster anywhere in the world. The Earthquake Support Subsidy was part of the package, but this time at a much bigger scale.
On February 28, Monday, at a press conference late in the afternoon, the Prime Minister, appeared in front of the cameras and microphones. The whole nation watched and listened, as they placed their entire future in his arms. Prime Minister Kay, exuding the sort of stability the Kiwis needed, announced that employers and individuals could immediately apply for the Earthquake Support Package via a new online application. This time around there was no restriction on the size of the company, and there would be an additional part especially designed for individuals who had lost their jobs. He promised that much-needed financial support would be in people’s accounts starting Wednesday, March 2, two days from his announcement and just eight days after the earthquake.
Less than 24 hours later, more than 10,000 people had already applied for the government’s Earthquake Support Package. Many more did so and benefited in the following days and weeks. By the end of June 2011, 20,000 employers and 50,000 employees received a combined total of $202 million. It was coined “phenomenal decision-making by government” and “the single most important intervention of the whole earthquake recovery.” Most businesses survived the earthquake — in fact, a larger proportion than what is usually seen after comparable events elsewhere. Few of those businesses would ever be aware of a tiny detail in the timeline, though. Upon announcing the government’s decision to help people rebuild their lives and livelihoods on February 24, the IT system that would allow for this tremendous breath of fresh air did not yet exist.
****
When the government of New Zealand decided to alleviate the suffering of so many businesses, one thing became clear — it had to be executed completely online. With the entire CBD completely cordoned, the number of business owners expected to apply was in the thousands. The seven Work and Income offices in Christchurch would be unable to handle this volume of applications. Training more people, finding the physical space for the offices, and communicating their locations would have taken time and money — assistance was needed right then, there wasn’t time to set up physical infrastructure. With thousands of people still in shock and panic, speed was of the essence. With online submission, applications could be taken quickly, the volume could scale significantly without overburdening the infrastructure, the transactional cost for each application would be a fraction of its physical alternative, and the overhead of vetting applicants would be reduced to a minimum.
In the digital domain of Work and Income, the only remotely suitable, fully online application was the system for unemployment benefits. It required a lot of information, background checks, and security against fraud. As a result, it was complicated to use, took an estimated 45 minutes for the beneficiary to complete, and involved complex back-end processing through multiple organizations. Government officials had a tough decision in front of them — should they continue to use it or go for something completely new?
There was a bigger, albeit psychological, problem: It would have been incredibly disheartening for business owners to apply for unemployment benefits. Most of the businesses that would be applying were small and medium-sized companies, full of entrepreneurial spirit and a sense of ownership and pride in accomplishing everything the hard way. They had, after all, lively businesses until less than a week before. To receive government aid in a time of need was one thing, but to have that aid labeled as unemployment benefits or social welfare was something completely different. The matter needed to be handled very delicately. An entirely new IT system would be needed to solve these problems. Could it be built in such a short time? And was it possible to do this with existing IT department workers? The government wanted to start distributing funds as soon as possible. They had to take decisive action and it had to work.
****
IT workers have a habit of behaving like order takers and delivering whatever is ordered. They don’t participate in the specification of a system as a partner and trusted advisor; instead, they simply nod their heads, accept what is asked of them, however large and complex, and get on with it.
In the public sector, this behavior of simple order taking by public servants is particularly insidious, and it tends to result in tremendous waste of taxpayers’ money and long delays in delivering valuable public services. Public-sector IT projects are notoriously slow and inefficient. They have many more stakeholders than commercial projects, which makes them more complex to begin with. Those stakeholders do not have to “pay for the work” in a public-sector setting and they tend to ask for every possible function they can think of. In the midst of the bureaucracy and the paperwork, the planning of a project alone takes several months. By the time it is kick started, the scope has most likely changed and the planning projections are off; meanwhile, the stakeholders are eager as ever to see results. And when things do not go as planned, there are a mountain of sponsors, vendors, managers, and customers happy to point the finger at each other and produce copious reports to demonstrate that it’s not their fault. None of this is helpful. None of it brings value to the taxpayer in a timely manner. Whoever was going to build the new Earthquake Support Subsidy System had to change the paradigm of public-sector IT projects. Prime Minister Kay gave the responsibility to Paula Bennett’s Ministry of Social Development (MSD).
****
The Chief Information Officer of MSD was tasked with assembling the team that would build the new Earthquake Support Subsidy System in a very short time. He handpicked 30 people during the day on Friday. They met at the Ministry of Social Development in Wellington to plan what to do. The task at hand was clear to everyone — create a system that would get fellow Kiwis through the next six weeks. While everybody understood the importance of the work, the question remained: Could they actually achieve it? Could they adjust their habits in order to do so?
The team of 30 was led by a project manager. He was spared the usual overhead that public policy entails: months of outlining specific system requirements; writing a business case; creating a system for planning, monitoring, and reporting benefits; and receiving approvals from multiple levels of government hierarchy. The little paperwork that had to be done was a mere ten percent of the overhead that was routine for a government IT project. In return, he had to make sure they delivered faster than anyone had ever done in this IT division.
The team spent the first few hours with representatives from the four stakeholder groups that were directly involved with what the system would have to do:
? The finance department would make the actual payments and had very broad requirements for the system.
? The Inland Revenue department advised on security and would do whatever minimal vetting was still needed to prevent fraud.
? The owner of the policies behind the system, the government itself, laid out how things were to happen, who was to get the money, and when.
? The call center would field all the inquiries about how the system worked and receive complaints and suggestions from its users.
In the late hours on Friday, the team already had a wall full of sketches and sticky notes describing the functionality that would be needed. Analysis and development had begun.
The Christchurch Earthquake Employment Support System comprised two parts. The first, the Earthquake Support Subsidy (ESS), was directed at employers. Similar to the way it was handled after the September earthquake, the money from the subsidy was transferred to employers who then, based on high trust, used the funds to pay wages to all of their employees. As their businesses had been disrupted in a completely disrupted city, they had an array of important decisions to make in front of them. The subsidy ensured that whether to keep their employees around in the short term was not one of them. They had to determine what to do with their businesses beyond survival — location, market, adoption of new business models, renewal of the supply chain, and so on. These employers would receive $500 a week per full-time employee and $350 per part-time employee for an initial period of six weeks.[3]
The second part of the support system, the Job Loss Cover, was dispensed to employees whose employer was forced to close the business permanently, or those who could not get in touch with their employer. Former full-time employees would receive a weekly sum of $400 for a period of six weeks. Part-time employees would receive $250 for the same period of time. The money would help pay the bills while they looked for new employment or sought new outlets for their skills. The social benefit was huge: Freshly unemployed people need not worry about finding immediate employment; instead, they could focus on sorting out their lives.
It was a dream to have a public policy that affected so many people so positively. That dream kept the team behind the system awake for the next 60 hours, during which they took almost no breaks as they created it. The project manager coordinated the whole process, making sure developers did what was asked of them. He did not sleep a single minute during that crucial weekend.
On Monday morning, the lack of rest, the immense stress, and the vastness of the project turned the dream into a nightmare for the project manager. Suffering from exhaustion, he became absolutely incoherent. He could not speak nor comprehend what others were telling him. He had to stop and take a break. The CIO sent him home. The first version of the software had yet to be released, and four levels of managers came intermittently to interrupt the developers and demand status updates. All of Christchurch was waiting. The tired and stressed out developers had created many parts of the system already, but it was difficult to juggle requirements from four separate directions and know what comprised a ready-to-release first working version of the system. The people responsible for the call centers were especially anxious to know when the system would be ready because they had to train the people who would start picking up the phones in just a few hours. When Russell Healy picked up his phone that Monday morning, he heard a trembling voice on the other end of the line: “We need your help for a few days.”
****
Russell Healy was no stranger to the team or the Ministry. He had started his career in the Ministry after graduating from Victoria University in computer science almost ten years before. In 2003 he was the first IT worker in an organization of 2,000 workers. Over the years, he had gradually created more teams. Calm and cool-headed, there had always been a bit of restlessness about Russell and the manner in which he created software and helped others create it. Were there better ways to develop software with better programming languages and more user friendliness? Russell not only looked for the answers to those questions, but also had the capability to make compelling arguments in favor of experimenting with possible solutions. In a public-sector domain, that was quite the accomplishment.
Some of the people on the ESS team had experienced time with Russell through use of innovative programming languages or processes where he had pioneered their adoption. In retrospect, the reason that a fast release of the Christchurch Earthquake Employment Support System even stood a chance was that with his help, the team had already mastered frameworks for creating very simple software products and were working with contemporary methods that he had instituted during his time there. Eventually Russell had moved to a coaching role as a contractor to his former employer and had become less active with the teams themselves. To the Ministry’s Chief Information Officer, Russell was an obvious choice to step in at such a crucial moment. It was his cool-head and avidness for processes that the developers most needed in that difficult moment. The entire government was concerned about this project and whether the team would manage to deliver. Christchurch needed help, and now everything was in the hands of software developers. They, too, needed some help.
As Russell came in the room, everyone was standing up and talking simultaneously to each other, giving status reports. Confusion reigned. No one was actually sitting at a computer and working. They were trying to come to grips with all of the requirements and the system they had to deliver: debating, pondering, arguing, wondering, but not pressing a single key on a keyboard. What was the most important thing? The pressure under which they had to make decisions and the accumulated lack of sleep made everything all the more difficult. The walls were filled with designs for prototypes and notes with requirements in no particular order. It would have been difficult for anybody to really understand what was going on without talking to half the people who were there. Russell stood in the middle of the room and tried to make sense of everything.
He had to act quickly. He had to help them get a better sense of the situation and make it easier for them to choose what to work on, help them to make the most appropriate decisions for expending their time and effort. Decision making of any scale in the public sector is a renowned problem. Responsibility for decisions is often spread across too many people, many of whom lack the capacity and time to give them sufficient attention. These people exist in a culture where willingness to take responsibility is not nurtured, and many high-ranking officials attempt to micromanage pet projects. Russell had to bypass that while remaining calm and cool-headed. This was not the time for wasteful discussions, debates, or, at the height of emotions, even quarrels. Every minute counted, as thousands of people were without means to earn their living. Whatever he did at that instant had to be acceptable to everyone in the room. It could not be invasive, it could not be revolutionary. What was required was that he bring a little more order to a complex situation.
Russell took a black marker and walked up to one of the walls. He drew one vertical line, about a meter long, walked a few steps to the right, and drew another one parallel to the first. He repeated this a few more times. Russell then drew a horizontal line across the higher end of the vertical lines. There was space enough to write a word above each column that he had formed: To-Do, Analysis, Development, Test, Ready for Release, Acceptance Test, Done. These activities were standard for the process of software development. Russell stepped back. The drawing he had just made was a simple and clear process representation.
The requirements from the team’s stakeholders had had to be visualized — at least the ones they were currently working on. Russell wrote each on a sticky note, which he placed in the appropriate column. The picture was becoming clear, but it seemed that people were working on too many things at the same time. Russell wanted to avoid that because it slowed people down while adding additional pressure. The team believed that their way would help them to do more things, but as Russell knew, the math was not so simple.
Russell took a stack of smaller sticky notes and gave two blank ones to each person. No more, no less. They each had to hand-write their name on those two sticky notes. From then on, each person had to use one of their named sticky notes as an indicator (or avatar, as these are sometimes called) of what they were working on. They would place their avatar on the bottom of the bigger post-it note that bore the name of the task they were working on. The team members got only two avatars because Russell wanted to make sure that people focused on a single task until completion. The second one was just in case the first item was blocked for some external reason. Russell was convinced that the best quality — and speed, for that matter — was achieved with full focus on a single task. It would be better to focus on completing and delivering a few things at a time than to fail to deliver many things.
A sense of order quickly crept in. People could easily see what was needed and could easily pull a task without needing to discuss what to work on next. The noise subsided as the developers took their seats and began working on the tickets they had all agreed were most important. Later that day, the first version of the system was released. As soon as word spread in the Ministry that the system was ready, Prime Minister Kay scheduled his press briefing for the afternoon. The aid campaign was now open for business. Citizens of Canterbury could begin applying immediately and expect the money, as promised, by Wednesday.
****
This was only the beginning. For the new system to be truly and fully operational, many more features needed to be added. But Russell knew how unhealthy it was to have all of them lying around. Those as yet unstarted features hung like a cloud bearing down on the developers. They were continually reminded of the burden they had yet to carry. That weekend’s mode of working, while fruitful, could not continue or soon everyone in the room was going to be incoherent. The welfare of their fellow countrymen depended on their ability to continue working. They had to find a sustainable pace. Russell knew that it was crucial to assess the importance of each requirement and to prioritize them. He wanted to sequence the requirements in smaller batches and remove the amorphous black cloud that hung over them, gnawing at their minds, distracting their thought processes, and affecting the quality of work already in progress. He would hold the conversation with all the stakeholders privately. He would take personal responsibility for whatever decisions were made regarding priorities, scheduling, and sequencing of work. This wasn’t a time for endless debate. It was a time to keep the team focused on delivery. He would relieve them of the burden of worrying about the work that was yet to come.
Russell arranged to have a meeting with representatives of each stakeholder group every day at ten o’clock for exactly one hour. To keep the meeting short and to avoid overwhelming the developers with too many things to do, he made one pivotal decision and request from the stakeholders. Russell and the project manager (who had returned from his rest) would come out of the meeting with only small tasks that were the most important to work on. Whatever they held in their hands as ideas on paper had to be live in the system within a maximum of two days from that day’s meeting. All requirements that were discussed had to conform to that constraint. Russell was very strict about that. He saw it as a waste of valuable time to discuss other ideas that were either too big or involved too many small functions to be implemented within two days’ time.
Russell understood how vital it was for people to experience the sensation of having delivered value each and every day; twice, in fact, as the team decided to release their written and tested code in the morning and in the evening. The frustration of working on things that do not see the eyes of end consumers for months, a practice that was the status quo in the public sector, was the last thing those 30 people needed in those days of early March 2011. The pressure of the emergency situation would have burnt out the whole team. The feeling from pushing a piece of actual, working functionality live for users, every day, was a new and fascinating experience for them.
Each morning at 11 a.m., Russell entered the developers’ room carrying a fresh stack of highly important tasks to work on that day. They quickly discussed them, making sure everybody understood everything. Then they began working silently and calmly, just as they had done the day that Russell arrived. Nobody questioned Russell’s judgment; all they wanted was to have something to work on at every moment. Sticky notes for each new task were placed on the wall in the To-Do column. Team members started picking them, and as the work progressed, pulling them into the appropriate columns for Analysis, Development, and Test, attaching their avatar to indicate their ownership of the work. Russell monitored their card wall closely, making sure that each task had a designated owner and that if someone was blocked for some reason, he could assist immediately.
As soon as a work item was developed, it was quickly sent for testing and then completed for delivery. Russell wanted to figure out the quickest way to get a work item to move from the very left side of the wall to the very right side of it. He wanted to make sure not only that everyone had something to do, but that every task was worked on continuously. If the code had been written, someone immediately needed to test it. His goal was to shrink the time between commitment at the meeting with the stakeholders to delivery to the live production system. The value of early delivery meant more functionality for more distressed people needing relief from the hardship the earthquake had brought them.
****
Organizing work in this way enabled a sense of calm to take over, and it spread beyond the IT team. Whenever someone from any of the four levels of management wanted to see the system’s progress and which features would be ready for delivery, all they had to do was check the board or get an update from Russell. A small blue note on each task told whether it would be delivered in the first or second release of the day. That was especially important for the call center stakeholders so they could quickly update the call centers with that information. Those telephone lines had become very busy and were a reliable source of information for many Cantebrians in need. And no one ever bothered the developers by interrupting them to ask for status updates.
In this manner, every day the developers came to work they added yet more vital features to the system. By the time they left the office in the evening, something they had personally developed that very day (or the previous day) was already in use by the people of Christchurch. Everyone was happy, and in the end, almost no businesses left Christchurch or were closed, thanks to the pivotal financial help they received in the first six weeks after the earthquake.
****
The reminders of the earthquake are with Cantebrians every single day, sometimes in a good sense, such as when they see everything to which it gave birth: The mall made from cargo containers; the Cardboard Cathedral; the bar that occupies a bus; or all the creative outbursts that have arisen from Gap Filler,[4] such as the Dance-o-mat and the Pallet Pavilion. Sometimes the reminders are not as positive:
I can hear an earthquake coming, and confuse it with trucks. You freeze and hold your breath. You look at what’s around you, what can fall on you. You don’t relax till it passes. The feeling of readiness will probably stay with me my whole life. Most of all I miss the complacency — we took for granted that each day would not bring us abrupt terror.
— Rachael and Eddie
People’s lives have changed since the shock of the earthquake, probably beyond recognition. But they have stood up and they have carried on with new zeal. And that is what humans are uncannily good at — surviving, prevailing, and moving on.
****
When Russell Healy walked into the Ministry of Social Development building in Wellington on Monday morning, February 28, 2011, he was walking in with a plan. Russell wasn’t a lone ranger with an innate brilliance for turning around struggling IT projects; he was a trained manager working from a playbook. There was a method to his actions. His were not the off-the-cuff situational reactions of a cowboy genius. He had a good idea of what he was walking into: a team of people filled with passion and highly motivated by the task they’d been handed only the Friday evening prior. He knew many of them personally. He’d worked with them. He knew they’d been working all weekend. He knew they’d thrown themselves into their project with gusto. He also knew they were almost certainly overwhelmed and buried in the middle of the work.
Russell was going to employ the Kanban Method to help the Earthquake Employment Support System project focus on service delivery and get more done faster by focusing acutely on working on only a few items at a time. He knew the team was technically competent. He didn’t doubt their technical abilities to create the software required. He, himself, had introduced many of the technologies and programming techniques they were using. He’d trained some of the team members personally. They didn’t need coaching on development technology; they needed help in an area they’d never been trained in: organizing themselves for rapid, focused service delivery. They needed direct management assistance and coaching.
First, Russell wanted to see what was truly going on in the software development emergency room. He wanted to see the work, the workflow, and who was working on what. He wanted to visualize what was invisible. By visualizing the development process on the wall as a series of columns, and the work as sticky notes that flowed left to right across those columns as the activities involved changed, and by having the workers attach their avatars to the work tickets, he could see what was going on. The whole team could see what was in progress at which stage of completion, who was working on what, and they could observe the movement of the tickets to determine how long things were taking to flow from start to finish. In doing so, Russell was able to engage all of his senses to the task of management. He enabled the whole team to share the same sensory experience. They could all see the board. They all had the same information and were able to adjust their behavior accordingly.
Secondly, Russell wanted to relieve the team members of over-burdening, reduce multi-tasking, and focus on high-quality delivery of just a few things at any given time. By setting a policy of only two items per person and enforcing this by issuing only two avatar tickets for each individual, he greatly reduced the multi-tasking and increased the level of focus. To truly lift the burden of the massed cloud of requirements hanging over them and let them clear their minds, he had to insert himself as a middleman. He took on the personal responsibility of facilitating prioritization with stakeholders. He used his knowledge of the “cost of delay” concept to facilitate conversations that solicited input on “What do we really need most for delivery two days from now?” It’s a simple question with a time-based driver. He wasn’t asking, “What will give us the most return on investment?” He wasn’t asking, “What will generate the most revenue?” or “What will be cheapest to implement?” He was asking a simple question, “What functions do most of our user population need most two days from now?”
Next, he taught the team to manage the flow of work by trying to commit to just one work item each, two if necessary because something was blocked for a reason external to the immediate project team. Blockers were visualized so that he could escalate them to management. He arranged to meet with stakeholders once per day. He fed new committed work to the team every day. They already had the technical capability and technology to deploy completed code quickly and frequently. So he arranged to do this twice per day. By doing so, it was possible that features committed to in the morning could be deployed to production the following day: a one-day lead time for new functionality from stakeholder request to live production deployment on a public-sector service project. Taxpayers have never seen online services created so efficiently and cost-effectively.
Russell made sure important policies were explicitly declared, understood by everyone, and made visible and transparent: A maximum of two items per person was enforced with the sticky avatar system; stakeholder meetings were at 10 a.m. every day; team meetings were at 11 a.m. every day; deployments were twice per day at the end of the morning and the evening sessions; after the initial push, the team would work a sustainable pace — they would go home and sleep at night; he would take sole responsibility for facilitating prioritization and communicating new committed work to the team.
Russell employed the Kanban Method to guide his actions. He followed the principles of “start with what you do now” and “respect existing roles, responsibilities, and job titles.” He implicitly followed the third principle of “agreeing to pursue evolutionary change” by not shaking things up and revolutionizing the working practices and processes. He’d simply augmented them with some visualization, some clarity and focus, and some explicit policies to enable clarity and focus.
He then explicitly used the first four management practices of the Kanban Method. He made the invisible work and working processes visible. He limited the quantity of work-in-progress. He focused the team on managing the flow of work from commitment with stakeholders to deployment to the live production environment. And he made explicit all the policies that would ensure that these things happened and gained agreement and consensus around the implementation of the new policies.
Trained, experienced management that can keep a calm head and take decisive, effective action in a crisis is vital for success. Harnessing the passion and commitment of a skilled and highly educated workforce can make the difference between an effective desirable outcome and yet another IT project failure. In this case, failure wasn’t an option. New Zealand was facing a national disaster. No one was in the mood for excuses and another story of a well-meaning IT project gone awry.
There were doubtless many heroes on the weekend of February 24, 2011, in Christchurch: Lives were saved and rescues performed. Few people know the story of the Earthquake Employment Support System and its successful delivery. Only a few like Russell Healy[5] have the satisfaction of knowing that their contribution enabled the Prime Minister to look good on TV and deliver on his promise of a week earlier. The value of the EESS can be measured in the effectiveness of the Earthquake Employment Subsidy: the smallest number of business failures after any comparable natural disaster anywhere in the world.[6]
[1] The Beehive building houses New Zealand’s Parliament and other government offices, so called because of its unusual, beehive-shaped architecture.
[2] Due to the thousands of persistent, damaging aftershocks, some parts of the CBD would remain cordoned for 859 days (the last cordon was removed on June 30, 2013), a worldwide precedent for any urban area.
[3] In retrospect, some of the business owners that benefited from the ESS openly admitted that this package was what prevented them from closing down their businesses.
[4] According to its website, gapfiller.org.nz, “Gap Filler aims to temporarily activate vacant sites within Christchurch with creative projects, to make for a more interesting, dynamic and vibrant city.”
[5] Russell Healy can be contacted through his website https://getkanban.com/
[6] To learn more about Kanban visit https://leankanban.com/
For more writing by David J. Anderson visit his author page on Amazon.com