Upon further examination, he realized that he had entered the data to three decimal points — whereas the previous run had used the same data, but to six decimal points. This tiny difference in initial conditions had produced completely different results. These observations ultimately led him to formulate what we now call ‘The Butterfly Effect’, a term derived from an academic paper he presented in 1972: “Predictability: Does the Flap of a Butterfly’s Wings in Brazil Set Off a Tornado in Texas?”

The Butterfly Effect has since been used to explain numerous major changes in society. In The Tipping Point, Malcolm Gladwell describes how William Bratton — first as head of the New York Transit Authority’s police force (1990-1994) and then as head of the NYPD (1994-1996) — dramatically reduced crime in the city by implementing a few (seemingly) minor changes, such as cracking down on fare-dodging on the subway and penalizing quality-of-life crimes, like public drunkenness and public urination.

Similarly, in their best-selling book Nudge, professors Richard Thaler and Cass Sunstein describe several examples where small changes have produced radical results. For example, by painting a life-sized image of a black housefly on each urinal, authorities at Amsterdam’s Schiphol Airport reduced ‘spillage’ in men’s washrooms by 80 per cent.

Elsewhere, officials in Minnesota achieved a significant increase in tax compliance, simply by telling people that more than 90 per cent of Minnesotans already complied in full with their obligations; and in Texas, authorities achieved a 29 per cent reduction in roadside littering within a year of launching the ‘Don’t Mess with Texas’ advertising campaign.

Equally impressive examples of the Butterfly Effect have been provided by social psychologists. In one study, asking people to donate blood by having a friendly peer comply with the request was found to increase the success rate from 25 per cent to 67 per cent. More importantly, none of the subjects in the no-model experiment followed up to actually give blood, whereas 33 per cent of the subjects in the model experiment did so.

Similarly, simply commending children for already being very clean and tidy was found to increase their tidiness by a factor of four in the short term — a behaviour that remained unchanged long after the experiment was completed. Meanwhile, children who were told to be more tidy increased their tidiness in the short term by a factor of two — but reverted to their pre-test level of littering after the experiment. Children who were neither told to be more tidy nor praised for being tidy did not change their behaviour.

There are also examples from the world of business. In one study, workers in a pyjama factory were asked to take on board a seemingly small change in the way the pyjamas were sewn or boxed. One group of workers was simply informed of the change. In another group, a selected team of workers was asked to meet with management and then inform the rest of the employees about the change and help with the implementation. In yet another group, all the employees became ‘special operators’ and were asked to help implement the change. The difference in productivity gains was dramatic: the first group experienced a sharp drop in productivity and a loss of morale (17 per cent quit their jobs). In the second group, there was no morale loss (nobody quit) and the initial drop in productivity was regained within two weeks. In the third group, morale remained high and a slight drop in productivity (for one day) was soon followed by steady increase, to a level 15 per cent higher than the pre-experiment level’.

Small Changes in What, Exactly?

There is no question that small changes in initial conditions can have a big effect on outcomes. But this raises a practical question: small changes in what, exactly? Surely we do not expect that small changes in everything will produce desired effects.

The answer is to identify what academics in the system dynamics field call ‘high-leverage’ points in the system. As MIT’s Jay Forrester, founding father of this academic discipline, once said, “All [social systems] seem to have a few sensitive influence points through which the behaviour of the system can be changed.” The trick is therefore to find the high-leverage points in a given system.

---

All social systems have a few sensitive ‘influence points’ through which the behaviour of the system can be changed.

---

The problem is, different systems have different leverage points — and it is near-impossible to identify them, as they only become evident after the fact. This implies that experimentation and learning as you go along are crucial. But work in system dynamics suggests that the most likely candidates — and the areas where one can start experimenting — can be found in one of the following components
of a system:

1. THE SYSTEM’S PURPOSE

What is the system trying to achieve? This can be broken down into two subcategories:

• The paradigm or accepted philosophy on which the whole system is based. This is the set of unwritten and deeply-held assumptions and beliefs about how the world as a whole (and not just the system) works. Though unwritten, these beliefs are shared by all actors in the system in a subconscious way.
• The goal of the system, i.e. What are we trying to achieve? The goal is important because it determines how we organize the system and how we behave in it (i.e. our strategy).

2. THE MAIN ELEMENTS THAT MAKE UP THE SYSTEM

This can be divided into five subcategories:

• The rules of the system, which may be formal or informal, physical or social. They define what is accepted (or not) in a system and provide people with the parameters within which they have to operate.
• The physical structure of the system, which includes the underlying processes as well as the structure of the hierarchy.
• Information flows in the system. What information is delivered in the system, by whom and how?
• Time delays in the system. These may be information delays or delays between action and effect.
• Positive and negative feedback loops in the system.

3. THE INTERCONNECTIONS IN THE SYSTEM

This can be divided into three subcategories:

• The incentives in the system, which may be financial or non-financial, positive or negative.
• The cultural norms and values in the system. This is a set of beliefs and assumptions that acts as our moral compass, determining what is ‘right’ and ‘wrong’. It includes deeply-held beliefs as to ‘what will get you ahead in this place,’ and ‘how this place works’.
• The way the system absorbs change (also called ‘buffers’ or stabilizing stocks).

A classic example of how changing a system’s purpose or goal can produce dramatic behavioural change is provided by the redesign of Norway’s prison system. Halden Fengsel, one of Norway’s newest prisons, was 10 years and 1.5 billion Norwegian kroner (US$252m) in the making. Spread over 75 acres of sloping forest in southeastern Norway, the facility boasts amenities such as a sound studio, jogging trails and a freestanding two-bed house where inmates can host their families during overnight visits. Unlike many U.S. prisons, the air isn’t tinged with the smell of sweat and urine. Instead, the scent of orange sorbet emanates from the ‘kitchen laboratory’, where inmates take cooking courses.

Halden — Norway’s second largest prison, with a capacity of 252 inmates, opened in 2010 — embodies the guiding principles of the country’s penal system: that repressive prisons do not work and that treating prisoners humanely boosts their chances of reintegrating into society. Design plays a key role in the facility’s rehabilitation efforts. The objective is for the prison to look as much like the outside world as possible. To avoid an institutional feel, exteriors are not concrete but made of bricks, galvanized steel and larch; the buildings seem to grow organically from the woodlands. While there is one obvious symbol of incarceration — a six-metre concrete security wall along the perimeter — trees obscure it, and its top has been rounded off.

The cells rival well-appointed college dorm rooms, with flat-screen TVs and mini fridges. Designers chose long, vertical windows for the rooms because they let in more sunlight; and there are no bars on the doors. Every 10 to 12 cells share a living room and kitchen. With their stainless-steel countertops, wraparound sofas and coffee tables, they resemble IKEA showrooms. Halden’s greatest asset, though, may be the strong relationship between staff and inmates. Prison guards don’t carry guns; that creates unnecessary intimidation and social distance. Indeed, they routinely eat meals and play sports with the inmates, and half of the guards are women. It is believed this decreases aggression — and prisoners receive questionnaires asking how their prison experience can be improved.

Countries track recidivism rates differently, but even an imperfect comparison suggests that the Norwegian model works. Within two years of their release, 20 per cent of Norway’s prisoners end up back in jail. In the UK and the U.S., that figure is between 50 and 60 per cent. Of course, a low level of criminality gives Norway a massive advantage. Its prison roll lists a mere 3,300, or 69 per 100,000 people, compared with 2.3 million, or 753 per 100,000, in the U.S. — the highest rate in the world. Yet by changing the purpose of this system, its performance has been transformed.

In closing

Rather than focusing our time and resources solely on making major breakthroughs, we recommend that leaders also spend some of their time and resources thinking about the potential power of small changes. As indicated herein, small innovations are often as effective as big ones.