Showing results for 
Show  only  | Search instead for 
Did you mean: 

How to Design for Long-Term Behavior Change—Part 1: New Habit Formation

Lithium Alumni (Retired)

Congats to Germany for winning the FIFA World Cup. Now, the game is over, it's time to get back to some serious gamification and talk about designing for sustained behavior changes.


Last time I told you about my biggest takeaway from GSummit 2014 after I gave a quick recap of the conference. I referenced Prof. BJ Fogg’s session specifically and I am devoting this (and the next) blog post to talk about some of my own work on long-term behavior changes. This is part 1, new habit formation.


The subject of long-term sustainable behavior change has been the center of my work for quite some time. In fact, my recent book—the science of social 2—is entirely devoted to sustainable (long-term) social strategies. But my interest for this subject began when I was studying enterprises’ effort to engage employees internally and customers externally. Despite the abundance of technologies and strategies that drive engagement, many companies’ efforts to motivate employee engagement (or customer engagement) seem at best transient. They either don’t work, or if they did work, their effect didn’t last long. Meaning the “change” was not sustainable. So companies have to constantly repeat their engagement program with their employees (or engagement campaign with their customers). And when they do repeat, the result is usually a diminishing return.


Motivation is Ineffective at Driving Long-Term Behavior

carrot & stick mgmt.pngDespite its ineffectiveness, companies still like to resort to their most familiar tool to drive behaviors—the carrot and the stick. If you want people (employees or customers) to behave a certain way, motivate them with rewards and incentives—a pay raise, a job promotion, a good deal (a discount), a gift, etc. Companies like to leverage these extrinsic motivators because they are easier to control, execute, understand, and their cost can be more precisely quantified. The problem is when the motivation is gone—after someone gets his pay raise, promotion, discount, gift, etc.—people’s behaviors often revert back to their old habit.


As you can see, motivation is only effective at driving temporary behavior changes, because people’s motivations are volatile—they change quickly—and they are easily influence by the environment and the people around them. So changes that are purely driven by motivation are usually not sustainable. To drive long-term behaviors, we need more determination than motivation, but that’s not easy to come by either. So what’s the single most important factor in driving long-term behavior?


The Key Driver for Long-Term Behavior

Two years ago, I was invited by Prof. Fogg himself to speak at a Health Habit conference he organized at Stanford. Since most health related habits are long-lasting, the conference explored possible ways that we can achieve long-term behavior changes. As it turns out, none of them involve any extrinsic motivation, and there are really only 3 paths that people develop long-lasting habits:

  1. If people have an epiphany (a self-realization): This path is intrinsically motivated, so we have little influence over it.
  2. If the context changed: This path involves changing people’s physical and social environment that is often not easily achieved. But when it’s feasible, it’s very effective.
  3. If the behavior is developed through a sequence of baby steps (tiny habits): This path involves designing a sequence of tiny habits leading to the desired behavior. The key is that every tiny habit is a baby step that is very simple. Although designing the sequence of tiny habits could be challenging, it’s at least systematic.


Tiny Habit Behavior Change px300.pngWe will focus on the 3rd path today because it reveals the key factor that drives habit formation. The power of baby steps in driving long-term behavior is in its simplicity. Prof. Fogg’s key insight is that if a behavior is simple enough, then people will do it regardless of their motivation. Clearly, from the Fogg’s behavior model, when the behavior is so simple that it’s all the way to on the right, then you will remain above the activation threshold whether you are motivated or not. Since you are always above the activation threshold, you will perform the behavior when triggered to do so. Therefore, if we coupled this baby step with a positive reinforcement, we can repeat the trigger as many times as we need until you learn the behavior—when it becomes a habit that lasts indefinitely.


So as it turns out, although the motivation factor is important for driving temporary behaviors, the ability factor (the simplicity of the behavior) matters much more in creating long-term habits. Furthermore, when the behavior is simple enough you don’t even need too much determination (which typically fades over time).


Creating Long-Term Behavior Change ≠ Creating Long-Term Behavior

Although we now know the key to creating a long-term habit, that’s only half of the picture. It is important to recognize that creating a habit is not the same thing as changing a habit. Changing a habit is more challenging, because it involves 2 steps:

  1. create a new habit
  2. get rid of the old habit


Change Ahead px250.pngSo now we understand step #1—how to create new habits, in part 2 of this topic (my next blog) I’ll address getting rid of the old habit and replacing it with the new habit. 



Creating new habits (i.e. long-term behaviors) is challenging, and most companies are not very successful with that part. Said another way, they suck at it because they focus on the wrong factor (i.e. motivation) of the behavior model (i.e. motivation, ability, and trigger). Contrary to what most people believe, motivation has little effect in driving long-term behavior. Motivation typically works well only to drive short-term, temporary behaviors.


The crucial factor for driving long-term behavior is ability. The insight is that if you can make a behavior simple enough, then regardless of how people’s motivation might change, they would still be above the activation threshold for that behavior. Once we have that, all we need is a routine trigger to drive long-term behavior; a simple task for a simple behavior change.


While this simple formula helps us create new habits, long-term behavior change requires one more step—replacing the old habit. Stay tuned for part 2 and more insights on designing for long-term behavior changes.



Michael Wu, Ph.D.mwu_whiteKangolHat_blog.jpg is CRM2010MKTAWRD_influentials.pngLithium's Chief Scientist. His research includes: deriving insights from big data, understanding the behavioral economics of gamification, engaging + finding true social media influencers, developing predictive + actionable social analytics algorithms, social CRM, and using cyber anthropology + social network analysis to unravel the collective dynamics of communities + social networks.


Michael was voted a 2010 Influential Leader by CRM Magazine for his work on predictive social analytics + its application to Social CRM. He's a blogger on Lithosphere, and you can follow him @mich8elwu or Google+.


Great article by Michael. Most initiatives fail because training or change management is often an afterthought at best. But you don't need  a large organizational change management infrastructure. Just a systematic approach to creating the right baby steps that are easy to do and drive behavior change. The good news is that most of the baby steps people already do outside of work as they social engage in Facebook, Twitter, etc can be leveraged internally.


Thanks. Wayne Tarken

Lithium Alumni (Retired)

Hello Wayne,


Thank you for the comment. 


Although many of the baby steps people already do outside of work as they engage on social media, these are usually not sufficient to drive behavior change, b/c there are already existing old habits. Next time we will discuss how to get rid of those old habit and replace them with the new habit for the long run, so people behavior can actually be changed.


Glad to see you here. Hope to see you again next time.


Great article, as always. I totally agree with the thought that motivation is ineffective at driving long-term behavior.

But isn't it the case that this is true only as long as we are talking about a behavior that is always the same because the context /environment where it is supposed to happen stays always the same? 

While it can be a great way to 'design' processes and environments that are stable and so enhance the ability to create foster particular behaviors by making everything easier to stay over the activation threshold, I'm wondering if, at the same time, we are designing the fun out of something?

The second most common reason of people that quit their job is: "I don't feel to progress at this job. I have to go somewhere else to be challenged again." Most often, these are people that experience their jobs as too easy. Their ability to perform good is always above the challenge. As long as they don't see an enormous meaning by doing what they are doing, they will probably miss the challenge and start looking for soemthing new. 

So, by taking this into account, I think that if we want to raise someone over the activity threshold regulary to be able to do a particular behavior and to even make it a habit, we should find a point over the threshold line that is not too far on the right side but still needs a particular level of motivation. Otherwise, without these regulary peaks of motivation (no matter how the motivation is triggered, hopefully not just extrisically) I'm concerned that we sacrifice the fun for creating a (boring) habit. Even if it works and we get the habit we aimed for, we increased the chance of loosing the employee because of boreout 😉


Just a thought...cheers, Roman 

Lithium Alumni (Retired)

Hello @RomanRackwitz,


Thanks again for the comment here.


I sincerely apologize that I couldn’t reply earlier, because it’s a great question you raised, and it’s probably one that many have been wondering about. My travel has just been pretty crazy and I simply don’t have the ability (i.e. time) to get to this, despite that I’m totally motivated to address your question. Part of the reason, as you can see from the length of my reply, is that there are some subtleties that require more detail explanation. But I hope the wait is worth it.


Alright, let me get into this…


First, when I talk about the ineffectiveness of motivation in this post, I meant most of the extrinsic motivations that people, companies, and gamification tools apply externally. Remember, if the reason a person wants to do a certain activity is not inherent to the activity itself, they are extrinsically motivated (see my earlier post on this subject). If a person is intrinsically motivated to do a certain activity (e.g. singing, reading, solving math puzzles, etc.), these motivation can sustain over long term. But this person will probably be doing those activity already. The fact that they are not already doing the active you want to drive means they are probably not intrinsically motivated in the first place. Hence, the need for motivation to get them to do it.


Second, the notion design of baby steps doesn’t meant that you don’t make the activity more challenging over time. You just need to START with a very simple baby step that doesn’t require any motivation for people to perform it. As people start doing the behavior, you do want to design the next step a little bit more challenging. Hence the “steps” (i.e. levels, progression, etc.) in the design of baby steps. You do want to give people the intrinsic motivation of mastery—that they can get better. And starting with a baby step give more people the ability and opportunity to get better over time.


If you remember my talk at this year’s GSummit, I briefly talked about how to design the baby step strategy. It doesn’t mean that every step is a baby step. It just means that every step is a baby step relative to the skill they’ve acquired from taking all the previous steps. The “relative to the skill they’ve acquire is the crucial point.” If we look at any one step, that step can be huge and seem very challenging (e.g. Angry Bird level 100). But if the person have completely level 1—99, the skills they’ve acquired would make level 100 look like a baby step. It is still a new next step, so it would still look challenging, but once they’ve gone through level 1—99, then all of a sudden level 100 seem much more achievable.


In terms of designing the behavior for habit formation. You do want to start as far right as possible. But you just want to make sure that the people with more ability can move quickly up the ladder of baby steps, eventually they will slow down somewhere on the ladder b/c the steps do get bigger and more challenging. If there is a user who just zip through all the steps, then we are not designing it right.


I can understand the confusion here, because a habit is really a different behavior than the one-time behavior. What I meant is that the habit of flossing—the behavior of flossing every night—is not the same as flossing one-time. So if you want to create habit, you do want to start with something so simple that people almost feel bored when doing it once. But remember, we are not trying to get people to do it once, we are trying to get people do it regularly, which is really a different behavior. Once they get into the habit of doing it REGULARLY, then we can slowly increase the difficulty through the baby step design I describe above—increase the step size, but still a baby step relative to the skill they’ve acquired.


Alright, I hope this help clarify the confusion.


Thank you again for commenting and always asking great and challenging questions. It shows who's thinking about this seriously.  😉

See you again next time.



Very nice article, It is very informative!


I would like to add two things here though,


For one, I totally agree that new habits can be formed by taking many baby steps toward changing it, but habits are something formed over long periods of time, and it would probably take awful lots of effort and time to amend them. For that reason, It is probably worth time to investigate the investment horizon before attempting such effort


Secondly, I think short term behavior changes can be just as valuable to businesses. In the case of a gamified BPM ecosystem, businesses can leverage these short-term behavior changes to buy time to react to situations that mandate a bigger change (eg. Redesign the business process).


Just my 2c 🙂




Lithium Alumni (Retired)

Hello @franklu 


Thank you for taking the time to comment and your thoughtfulness.


First, our conventional wisdom often tells us that habits are something that requires a lot of time and effort, and that is why creating habits is challenging. However, the point of the baby-step design strategy is to make it so simple that it's almost effortless. It might still take you time, because it may take many baby steps to get to the desired habit. If it still require a lot of effort so that we have to think about it when we act, then we didn't design the baby step right--the steps are not small enough. The baby steps has to start with very tiny habit (see my reply to @RomanRackwitz above).


I totally agree with your second point. I don't believe I've ever said that short-term change is bad or not valuable to business. It is just that business already know how to drive these short-term behavior changes. They are just not very good at driving the long-term behavior yet. The real danger is that businesses often think that they can drive long-term behavior change the same way they drive the short-term changes. That's the point of this post. Certainly there are value to both short-term and long-term changes, and this applies to business as well as pretty much everything else in our lives.


Thanks again for the comment, and I hope to see you again next time.


Lithium Alumni (Retired)

Hm, read everything and all the comments.

Short term motivation may work in many employee situations which is why many companies use extrinsic motivators. We all agree there.

However, customers, well, we want lifetime customers if possible. We may be willing to invest more to make them lifetime customers. To make it so that our products and services are bought and used habitually. We go to the gym habitually, church habitually, and buy coca cola habitually. When I think of changing the habits of a group of people (my customers, prospects, and employees) I think about religion. I know it is probably not the most politically correct place to go with illustrations. But religion has taught politics and business how to change habits, even the lives, of individuals and groups. Religion has been changing the habits of people longer than all of us. Why do you light incense every morning in the little shrine in your house? Why do you go to church every week? Why do you read the bible every day? Why do you believe what you believe and act how you act within the “border” of your religious or spiritual beliefs? I think Apple is successful because they built a "religion" more than a business. Apple customers have been loyal to the extreme since the 80s. Apple customers are not "mindless consumers" but part of a "movement"... a "rebellion." The big commercial in the 80s. Throw the big hammer at the screen and break the chains. You are free. Now join the revolution. Apple did a great job. Sorry if my response is too out there. For the record I only represent the views or opinions of myself. But I enjoy topics like this. Plenty to ponder. I will read article 2 now.

Lithium Alumni (Retired)

Hello @JustinFr,


Thx for the comment. And BTW, no comment is too far out there. We are here to learn and explore, so any comments are welcome.


There are certainly research being done on changing behavior of groups. They are much harder and less suitable for design purpose, but they do exist. I might discuss some of that in a later post. Although religion is one way to change group behavior. That is not the only way. Cultural norms, societal norms, and even familial norms, exerts forces that shape our habits, such as eating 3 meals a day, slurping while eating noodles, brushing teeth, etc. Religion is only an example of sociocultural norms that shape our habits.


It is easy to conclude that Apple or Coca Cola follow the same paradigm as religion, because religion is older, but there are many cultural and social norms that are even older and more fundamental in terms of human psychology and behavior economics, like social acceptance and social rejections. Religion have used these fundamental behavior drivers to get people to identify with the group (i.e. the religion), and then leverage other social facilitations to change behavior. From a behavior perspective, I see religion, aristocracy, and government as institutions that have tried many ways to manage and shape the behavior of their "user." Through trial and error, they've basically stumble upon some fundamental principles that governs human behavior. Then subsequently, they've been use as tactics for managing their follower or citizens alike, because they worked. 


However, the fundamental principles that governs human behavior exist way before religion. Some of them even exist in social animals, such as wolf packs, apes colonies, etc, before human walk the earth.


Thank you for the comment and discussion. 

See you again next time.




Dear MikeW,


I have been reading up on Fogg's behaviour model and got fascinated. However I do have some questions. If one where to actually use the FBM, how would you quantify the three factors. For example if I want to evaluate different possible technologies to get people to drive more economical and I want to quantify the simplicity factors.


1 - An email to remind people to drive economical because it is good for the environment and tell them to not exceed 80 km/h(they would still have to actively do this themselfs)

2 - The same email however now they just have to push a fuelsaving button instead, and can drive any speed instead of the 80 km/h.

3 - The same email however now one would have to get an upgrade and switch cars for 3 days.

If I want to evaluate the simplicity factors. Where would I start?


With Kind regards,




master student behavior & neursocience


Lithium Alumni (Retired)

Hello @Rsocial


Thx for commenting and apologize for the late reply. I've been out in Australia (Sydney & Melbourne) all last week and have just return to SF.


FBM is a design model, not a quantitative model. As such, it is not suitable for accurate measurement purpose. For one thing you can't really measure the motivation factor accurately b/c we don't have a way to effectively and efficiently measure what's going on inside people's brain. One might argue that you can use response output as a proxy for people's motivation, but that is misleading, b/c response output is not affected by motivation alone. It could simply b/c you have an effective trigger or an simple task.


You can try to quantify simplicity by measuring how much resources is required to perform the task. But you won't really know how simple the behavior is, because you will also need to know how much resources do the subject have access to in order to answer that.


There is no rigorous way to measure any of the 3 factors effectively and accurately. A scientist can always poke holes in it. However, you can heuristically determine whethere there is motivation, is there a trigger, and comparatively how simple the behavior is compare to another reference behavior based on how much resources are needed to carry out the task. Keep in mind that these are not accurate measurements though. And trying to treat them as accurate measurements could be misleading. 


For your example, the resource needed for #1 is just time to read the email. The resources needed for #2 is to read, but also to push a button. So it's less simple (more difficult). But how much more difficult? That is not easily quantifiable.


Finally, you can always survey people. For example, ask people to rank how difficult each task is. But I'm not sure you will get anything consistent out unless your sample is very very large.


Alright, not sure if this help, and sorry if this didn't help. But that is not what the FBM is designed to do.


See you around next time.