I own this book and took these notes to further my own learning. Taking notes, publishing them and re-reading them allow me to flatten my forgetting curve. If you enjoy these notes, I highly encourage you to do the same, buy this book here and take your own notes.

Designing for behavioral change

Introduction

• You don't “break a habit” (language can mislead), you ‘untangle’ (it requires persistence)
• The behavior model: B = MAT
• Fogg Method: 1) select the right target behavior, 2) Make the target behavior easy to do, 3) ensure a trigger will prompt the behavior
• Designing for behavior change in 4 steps: 1) understanding of the mind, 2) discover the right behavior, 3) design the initial product, 4) iterate and refine
• Behavioral plan: a detailed story of how the user progress from being a neophyte to accomplishing the action while using the product
• Designing for behavior change is not persuading because: 1) it relies on the fact that people already want to act, 2) it's about action and not about beliefs or intentions. Pushing someone to buy a product that he would not otherwise buy is a very different task than helping a recovering alcoholic avoid cues in the environment to drink

I] Understanding the mind and behavior change

• Simply putting bottles of water at an eye-level in Google's kitchen increased water uptake by 47% (Kuang 2012)
• In our daily life, we are often in ‘autopilot’ and we aren't making conscious ‘decisions’ at all. Researchers estimate that 1/2 of our daily lives are spent executing daily habits and other intuitive behaviors, and not consciously thinking about what we are doing
• Brain 1: intuitive and emotional vs. brain 2: deliberative, conscious, slow, self-aware
• Ambiguity effect: we are intuitively uncomfortable with actions in which the potential effects have unknown probabilities
• Anchoring: we automatically use an initial reference point (anchor) as a basis for estimates, even if the estimates are wrong
• Attentional bias: we pay attention to particular cues in our environment based on our internal states. (ex addicted to drug see more cues related to their addiction)
• Availability cascade: incorrect and correct ideas can become increasingly believed and widespread because of repetition by well-meaning people who don't want to appear wrong
• Availability heuristic: we estimate the likelihood of events based on how easy they are to remember (for example easily believe that name of famous people are very common)
• LESSON 1: The 1st time a user will try your APP, he will immediately judge it based on their prior experiences and associations. You don't have time to convince them logically, instead, you must gain insights into their prior associations to avoid land mines and find positives hook that help people change their own behavior
• With Pepsodent the ads men Claude Hopkins helped the American to create the habit of brushing their teeth: 1) he taught American the cue of feeling for tooth film, 2) tooth film became a routine, 3) reward was a minty tingle in their mouths
• LESSON 2: Human likes to have habits, it saves our mind work. Create or change existing habits requires 1) a specific cue, 2) a stable routine, 3) a reward (ideally that occurs immediately)
• LESSON 3: Our peers provide answers. People use shortcuts, that's why its important to show pictures of peoples that looks like your users. People are more likely to take action if they think others are taking the same
• LESSON 4: Users have limited available attention, limited times, limited memories, limited willpower… build you interface in consequences and try to take into account the other demands on the users’ brain
• Easier is better (less mental effort), familiar is better (‘mere exposure effects), beauty is better (halo effect), rewarding make the user wants to come back… routines is build with the expectations of reward
• Don’t make user failed (frequently), a human does not like to fail
• We do urgent things first

Why we take certain actions and not others

• Cue > Reaction > Evaluation > ability > timing
• Bj Fogg model for intentional action: 1) motivation (pleasure, pain, fear, hope, acceptance, rejection), 2) ability (easiness), 3) trigger (cue to act now)
• A cue can be external or internal and will need to fight against inattentional blindness (every day our brain filter out a lot of information to avoid the overload, think about the gorilla experiment during an NBA game)
• For new actions: external cues are vital (placing the product in the daily environment, using different cue each time to avoid fatigue and being ignored, building an association with part's of a person's existing routines
• Users evaluate the product in a blink of an eye, hence the importance to focus on TRUST and first-time user experience that will create the association and emotions linked to the action
• Highlights benefits, minimize costs, downplaying alternatives
• Making the benefits of the application clear, removing frictions (costs) and frustrations are important but when designing for behavioral change the most important is to remember that the value ascribed to the actions or product by the user could be totally misaligned with what we personally think. Taking stairs vs. elevator, for instance, the user may not believe that it might lead to long-term health…. Making any actions useless here
• User needs to have: Ability, action needed, resources, skills, belief in success
• ADD TIME to an action to make it more real “I should set-up my retirement account Thursday night at 8 PM”…. This helps the user to answer the question “when to act”
• Pre-commit to a specific time = way to help us to decide when to act
• What is important with the time of action is to understand: what the product actively does to make the timing ripe for action and what it does to align with the time when a person is naturally inclined to take action
• CREATE action funnel: cue > reaction > evaluation > ability > timing (people can drop out at every step)
• Fogg model shows that in order to increase an action to occur you can either influence the motivation or increase the ability to act (decreasing ‘costs’). What is important in this model is also that if it's already easy to perform the action, making it even easier won't help (diminish marginal returns)
• To summarize barriers to cue: 1) user forgets to act (limited attention because of his environment), 2) intuitive reaction (user does not trust), 3) conscious evaluation (user is not motivated, cost too high), 4) ability (user does not know how to perform, fears failure, …) 5) insufficient urgency (procrastination)

3 strategies for behavior change

• 1.Cheat Strategy
• Default-it: example the 401k auto-enrollment retirement saving plan in the UK, example good camera default setting that is simple and provide good pictures in all scenarios
• Make it incidental: make the action happen when the user does something else. For example how to make people take a vitamin? Show the benefits, pay them…. Or just add them with their consent in their daily food
• 2.Make or Change habits
• Before the habit is created the user still need to choose to act
• Create a habit TACTIC: (cue-routine-reward) > CRR model => 1) identify a routine, 2) identify a reward (random is better + need to be systematic), 3) identify an unambiguous and single-purpose cue in a person's daily life, 4) make sure user knows what is the cue, action, and reward associated 5) make sure user wants and can easily undertake the routine, 6)deploy the cue, 7)facilitate the routine + track it, 8) find immediate reward from the product, 9) repeat 6-8 and iterate
• Changing a habit TACTIC
• Changing a habit is very hard…. Because the habit is unconscious and automatic + never truly disappears… it can remain latent and unused
• 5 tactics to change a habit: 1) avoid the cue, 2) replace the routine, 3) cleverly use consciousness to interfere, 4) use mindfulness to avoid acting on the cue
• Replace the routine: 1) identify the trigger and the reward, 2) when triggers occur, consciously engage in a different routine that provides a similar reward, 3) continue until the new routine is installed
• Conscious interference: the Prius effect
• Use mindfulness: mindfulness in Buddhism = being aware of the present moment and its experiences. Bring the cue to a conscious awareness
• 3.Support conscious action
• Help the user to think about the action and take the necessary steps to make it happens. Educating people… use this strategy when the 2 first ones are not possible

II] Discovering the right actor, action, and outcome

• “Mindless eating” (Wansink 2010). Free pop-corn before a movie, people eat them anyway. + soup bowls re-fill experiment…. Ccl: our eating behaviors are often on autopilot mode… that's why he started the “small plate Movement”, to counter our auto-cue of “eating everything on my plate”
• The outcome of the Discovery process is 3 things: the outcome > the actor > the action
• Five stages of behavior discovery: 1)clarify the product vision, 2) identify product outcome, 3) generate a list of actions, 4) get to know your users and what is feasible with them, 5)Evaluate actions and prioritize the best one
• 1.Product vision: find the concrete impact that the product should have (nike+ “help people exercise more”)
• 2.Product outcome: clarify the outcome the product will achieve (which type (env or people), where, what, when) > clear and measurable outcomes: California will have an average BMI of 24, Employees at x won't smoke…. Avoid unclear outcome: users will understand the danger of smoking…. The target outcome is the definition of the success of the product… outcome does not contain the ‘how’ and not necessarily contain ‘how much’ at this stage
• Discovery process of the company-centric: product vision > company objectives > user outcome > action > actor
• Discovery process of the user-centric company: product vision > user outcome > action > actor
• 2 types of companies: 1)focus on how the product will benefit the user, which in turn helps the company bottom line. 2)focus on how the product will benefit the company, by way of providing value to the users
• 3.generate clear actions: imagine that the outcome is “help users have money for saving”, actions could be: finding a higher paying job, selling the unused asset, getting a lower mortgage rate, spending less money on daily purchases…. For each action, define: who is doing the action, what is specifically and physically doing the person, how does it cause the target outcome. Always start with small and easy actions, favorise one-time action, think about minimum viable action of your actions
• 4.research the target users: observation on the field, prior experience with the product, existing motivations to act, relationship with the company (trust), barriers to action, generate targeted group of users (personas)
• 5.Select the target actions: ease for users, the cost for the company. Define success and failure.
• NB: Handle diverse population, most behavior won't apply to everyone with the same results, one-size fit all product does not exist > adapt and personalize to your target audience

III] Developing the conceptual design

• Structure the action (feasibility) > design the environment (motivation & Ressources) > prepare the user (motivate)
• Structure the action: most of the time there is way more on the plate than what we think, for instance running: is not just about waking-up and run… we need shoes, we need to plan the route, finding when doing it, etc, etc. Write the sequences of steps in the real world before doing the action, label each of the steps: things to do with the product, things that the product should do in response, things that need to be done in the real world (outside of the product)… look for missing steps, one-time steps vs. several steps. especially for new users. SIMPLIFY IT, cut repeated actions to one-time actions, cut big repetitive actions to simpler ones, drop nicely to have steps. Cheat and default if you can, identify potential habits if needed…. MAKE THINGS EASY: avoid cognitive overload, easy to understand, and easy to accomplish… it's all about appearance, if you think you won't be able to achieve it, you won't even try… careful with the easiness of a task, in the other hand great effort build commitment to a product (sunk cost effect)
• Design the environment: help users to take action with one of these tactics: 1) increase motivation, 2) cue the user to act, 3)generate a feedback loop, 4)remove competition, 5)remove obstacles
• 1.Increase motivation: leverage existing motivation, before adding new ones. Understand intrinsic (inherent enjoyment of the activity itself) vs extrinsic motivation (desire to achieve a particular outcome, reward [money, winning a competition]). Fitbit = enjoyment of the effort + enjoyment of achieving a goal and being rewarded by the congratulations in the APP…. Careful with extrinsic motivation… too many could override intrinsic motivation and demotivate users…. It's better to use extrinsic to create intrinsic motivation. Finally avoid pushing the users… or best: the trick is to use the threat of punishment (ideally a self-imposed one) to motivate action without actually punishing people. Pull future motivation into the present (dan Ariely medication+movie example). Type of motivation: monetary, social, intrinsic: exploring….
• 2.Cue the user to act: immediacy + urgency. “Act Now!” cta… recent researchers have shown the impactor of the principle of “implementations intentions” in which people make a plan for actions in the future. It’s a way to tell the mind that when Y happens, do X… environment cues action in the future
• 3.Generate feedback loop: feedback users on their exercises and sleep habits (fitbit). FL should be timely, clear, and actionable
• Knock out the competition: search what in the environment is demotivating for the users, what in the environment got the attention of the user, what actions are already done…. Rules of thumb are to always drag user attention to 1 thing at a time. You can also use competing factors to your advantage by connecting it to your goals (for instance application built-in Facebook)… The third approach = shout louder, be more motivating, easier to accomplish
• 5.Remove and avoid obstacles: remove friction
• Preparing the user: Wilsons experiment: changing the student's perceptions of the past, they were more likely to succeed in the future… Prepare users to take action.
• Narrate and change how the user sees themselves: you want the user to move, narrate, and help him to see himself as someone who has already been exercising and just needs to do more. Tactic: Ask users about the past and congratulate them for this
• Associate and change how the user sees the action. Associate with positive + familiar
• Educate and change how users see the world. The spectrum of thinking intervention: when in a familiar situation with very little thinking, products can use automation/cheat/build habits vs. in unfamiliar situation (intensive thinking) where the product should make actions easier and simpler, create several steps, educate users

IV] Designing the interface

• structuring the action to make it feasible and inviting for the user, 2) constructing the env to support the action, 3) preparing the user to take the action
• Designing phase = from a to-do list to something that people want to interact with
• User stories are useful because they oblige the product team to think about the specific action from the users’ perspective (not from a business top-down approach)
• Provide the behavioral plan but avoid to specify the interface design. Focus on what the product needs to do, and not on how the product should do it.
• Let the designer team do its magic and coming up with a great experience
• Very likely the designer team is gonna use: a familiar design pattern for behavioral change (trackers, reminders, social sharing, planners, gamification,…)
• This pattern should fit the culture of the company
• When reviewing the mock-up/wireframe think about 2 things: 1. looking for gaps in the overall users experience across the product (create action funnel: cue, reaction, evaluation, ability, timing. + Action: structuring the action, constructing the environment, preparing the user to take the action). 2. looking for opportunities to apply specific tactics (mechanisms that affect decisions (100 things designers need to know, seductive interactive design…)
• QUESTION for STRUCTURE THE ACTION: do the screens provide the minimum viable action? What parts can be dropped, simplified, or automatized? Are there repeated actions that can be turned into habits? Is each action clearly communicated to the user? Does it appear feasible up-front?
• QUESTION for CONSTRUCT: does the user have a clear motivation to continue on each screen? Is he thinking about the motivation in each screen? Are we asking for one single and clear action on each screen? If the action takes longer than 1 sec, does the product provide feedback to the user of progress? What distraction on the screen may pull the user away? What distraction is the user likely to face?
• QUESTION for PREPARE: how does each person see himself, does the self-concept support action on this screen? Does this look familiar to what users already know, especially when it comes to taking action on the screen? Is it clear where and how to take action? Does the user know what is needed, especially action outside of the application?
• TACTICS TO SUPPORT ACTION: increase the power of the cue (make it clear, blinking text,..), increase trust (make the site pro and beautiful, increase interest and trust, deploy social proof, display strong authority on the subject, bypass automatic rejection, be authentic and personal). Evaluation (motivate users with relevant association, leverage loss aversion, use peer comparison, use competition, avoid cognitive overhead, avoid choice overload, avoid direct payments). Ability (elicit implementation intentions, default, cheat, positive peer comparison, increase the sense of feasibility). Timing (increase urgency frame text to avoid temporal myopia, remind of prior commitment to act, make commitments to friends, make a reward scarce
• Social proof: our behavior frequently conforms to what we believe our peers do. This is reinforced by the idea we have that others are watching and judging us: “the spotlight effect”
• Keep in mind what you define as success: should the product solve the issue for 10% of users or 100%? When in production, Go lean but when chunking always evaluate if the product still meets the behavioral aims of the application

V] Refining the product

• Measure impact to share with finance, with the team, to share with the world (trust on impact), to improve the impact, to learn, to settle an argument
• Define the right measurable outcome, avoid vanity metrics, then experiment to measure the causal impact
• Default confidence level is 95% (alpha = 5%). It means you can expect to incorrectly say there is an impact when really there is not one 5% of the time = FALSE POSITIVE
• Default Statistical power = 80% (beta error level = 1 - 80% = 0.2) = you can expect to incorrectly say there is an no impact when really there is one 20% of the time = FALSE negative
• Multivariate testing
• Staggered rollout
• Matching and quasi-experiments
• MultiArmed bandit techniques, good at driving users earlier toward the version of the page that is most effective (cost = time, it takes longer time to confirm if a page works better, risk if there is a false signal in an early stage of the experiment: https://www.optimizely.com/optimization-glossary/multi-armed-bandit/
• Research, proxy metrics
• Identify obstacles of behavioral change during the test: watch real people using it, check data, find bottleneck/drop in each step of the funnel, segment users, use small and quick test to improve each step
• RE-design for behavioral change existing products: 1. document the target outcome, actor, and action. 2. develop a rough behavioral plan using the product current sequence of steps, what does the product actually encourage users to do? 3. instrument the product to measure user behavior at each step of the way (if not done already). 4. Dig into the data to see what your users are doing? What impact is the product having? What obstacles are users facing? 5. generate ideas for product changes, prio them. 6. in parallel sketch a blue sky version of the product. 7. evaluate if the blue sky is very promising or too close to a current product. 8. add + fix change to the product

VI] Putting it into Practice

• Understand > discover > design > refine
• Understand (create)
• Discover (outcome, actor, action)
• Design (structure the action, design the environment, prepare the user)
• Refine (address impact, develop insights and ideas, change & measurement)
• 5 preconditions to re-engage with a product: cue to do so, positive intuitive reaction, positive conscious evaluation, ability to do so, a reason to do so, a reason to do so at a particular time

Source/others:

• Ideas42
• irrationalLabs
• Persons to follow: nir eyel, bj fogg, joe oppenheimer, jon winman, john beshears
• Tiny habits book