Start free trial
Trending New Genres Authors
EnglishEnglish
EspañolSpanish
简体中文Chinese
繁體中文Chinese (Traditional)
FrançaisFrench
DeutschGerman
日本語Japanese
PortuguêsPortuguese
ItalianoItalian
한국어Korean
РусскийRussian
NederlandsDutch
العربيةArabic
PolskiPolish
हिन्दीHindi
Tiếng ViệtVietnamese
SvenskaSwedish
ΕλληνικάGreek
TürkçeTurkish
ไทยThai
ČeštinaCzech
RomânăRomanian
MagyarHungarian
УкраїнськаUkrainian
IndonesiaIndonesian
DanskDanish
SuomiFinnish
БългарскиBulgarian
עבריתHebrew
NorskNorwegian
HrvatskiCroatian
CatalàCatalan
SlovenčinaSlovak
LietuviųLithuanian
SlovenščinaSlovenian
СрпскиSerbian
EestiEstonian
LatviešuLatvian
فارسیPersian
മലയാളംMalayalam
தமிழ்Tamil
اردوUrdu
Searching...
SoBrief
Start at the End

Start at the End

How to Build Products That Create Change
Build products around the behavior people need, then find out what actually changes it.
by Matt Wallaert 2019 256 pages
3.81
433 ratings
Listen
Immersive
V2.3
Amazon Kindle Audible
Try Full Access for 3 Days
Unlock listening & more!
Continue

Key Takeaways

Design from the behavior you want, not the product you want to build

Start with the destination. Matt Wallaert argues that creation should begin with a clearly described action, not a product idea, advertising concept, or inspiring vision. A product matters only if it changes behavior in a way that serves a real need. His Intervention Design Process, or IDP, works backward from that outcome: identify an opportunity, validate it, define the behavior, map what encourages or blocks it, design interventions, and test whether they work.

Bing in the Classroom shows why. Microsoft suspected students were searching less than their curiosity suggested. Classroom observation revealed the obstacle was not a lack of curiosity, but teachers’ concerns about inappropriate results, advertising, privacy, and classroom disruption. Addressing those barriers helped raise school searches by 40 percent and home searches by 15 percent.

Analysis

The shift from product-first thinking to behavior-first thinking is a useful antidote to feature worship. It also changes how teams discuss success: instead of debating whether a product feels compelling, they can ask what people actually do and whether that action matters. The Bing example shows why this distinction is practical, not merely philosophical. A campaign promoting curiosity would have treated an abundant motivation as scarce. Still, measurable behavior is not automatically valuable behavior. Teams must also ask who benefits from the change, and whether the chosen metric captures that benefit rather than a convenient proxy.

Treat every promising insight as false until independent evidence converges

Confidence is not evidence. A potential insight is an observation about a gap between the current world and a better possible one. It may come from data, direct observation, workplace lore, or research outside the organization. None should be treated as true simply because it sounds plausible or comes from a senior person. Wallaert recommends validating an insight through convergent validity: independent kinds of evidence that point toward the same conclusion.

The classroom investigation illustrates the method. Search logs showed students made fewer than one query per student per day. Watching classrooms confirmed that students used computers without searching much. Later observation showed that teachers, not children, were the key barrier. Separate quantitative and qualitative work can expose assumptions that a single method or a powerful person might miss.

Analysis

Triangulation is valuable because each method has different blind spots. Logs reveal what happened, while observation can suggest why; interviews provide participants’ accounts, which may or may not match their behavior. Keeping these sources distinct before comparing them also reduces the risk that a team simply repeats its own assumptions in different formats. Yet converging evidence can still share a hidden bias, especially when samples come from the same institution or population. Good validation therefore depends not only on multiple methods, but on asking whose experience is missing from the evidence.

Define the desired behavior in measurable terms and real-world conditions

Make the goal observable. A behavioral statement specifies the population, its motivation, the conditions outside the designer’s control, the behavior sought, and the data used to measure it. In plain terms: when a certain group wants something, and relevant conditions are met, what should they do, and how will anyone know? This prevents vague goals such as customer love from masquerading as outcomes.

Uber’s early goal was unusually clear. People wanted to travel between locations, and users in San Francisco needed a connected smartphone and electronic payment. The target action was taking an Uber, measured by completed rides. By contrast, counting computer sales could make a company look successful even if customers never used the machines. The statement should be ambitious, but it must evolve when a company’s market, capabilities, or customers change.

Analysis

A behavioral statement functions like a contract between strategy and measurement. It forces teams to distinguish the actual outcome from the intermediate signals that are easier to count, such as registrations or downloads. Its focus on motivation is also a useful safeguard: behavior change should connect to something people already want. The risk is false precision. A binary sentence can clarify scope, but real motivations, access, and behaviors often vary by degree and context. The statement is a disciplined starting point, not a complete description of human life.

Map what encourages action and what quietly makes it harder

Behavior reflects competing pressures. Promoting pressures make an action more likely; inhibiting pressures make it less likely. The behavior that results depends on their balance, so a weak outcome may reflect excessive friction rather than insufficient desire. Pressures are context-dependent: a price can discourage one person and signal quality to another, while a product’s playful image may suit a casual setting but feel wrong in a formal one.

The M&M example makes hidden influences visible. Taste and hunger encourage eating, as can bright colors and familiar branding. Distance from the candy, an opaque container, and concerns about calories can discourage it. In an office study described in the book, workers ate more when candy was nearby and visible. The lesson is to map both sides before building an intervention, rather than reflexively adding rewards or promotional messages.

Analysis

This model directs attention toward the often overlooked conditions surrounding a choice. Changing the placement of a snack or simplifying a payment can matter more than inventing another message about why the behavior is desirable. It also helps explain why identical products can produce different actions in different settings. The model is intentionally simple, which makes it usable, but real pressures can interact and change direction. A map should therefore be treated as a testable account of a particular population and moment, not a universal inventory of human motives.

Remove unnecessary friction before spending more to attract attention

The overlooked barrier may be the best lever. Teams naturally brainstorm ways to make an action more appealing, yet reducing obstacles can be cheaper, more durable, and useful to more people. Uber’s automatic payment, wider driver availability, and lower waiting times addressed burdens that customers did not necessarily describe as the central problem. The service made an existing motivation, getting somewhere, easier to act on.

Small barriers accumulate. A task can be blocked by cost, distance, uncertainty, time, awkward steps, or fear of making a mistake. In the book’s subway example, countdown clocks made waits feel about 30 percent faster without speeding up trains, because passengers could plan around clearer information. Friction reduction is not always the answer: some people enjoy effort or want to spend attention on choosing. Find out which effort your population wants removed.

Analysis

Friction reduction links behavioral design to the broader study of choice architecture: the same option can become easier to take when its steps, timing, or uncertainty change. The subway clocks are especially instructive because they improved the experience without changing the underlying transit system. This suggests that perceived control can matter alongside objective speed. But making actions effortless is not inherently beneficial. Removing a pause from a consequential choice can undermine reflection, so designers should distinguish pointless burden from useful deliberation.

Generate many solutions, then pilot distinct bets rather than favorites

Novelty is not the goal. Once pressures are validated, teams translate them into interventions, the practical changes intended to shift those pressures. One pressure may support several interventions, and one intervention may address several pressures. Generate a broad set before judging feasibility, then select a varied group with limited overlap. This gives pilots a better chance of revealing what works, rather than repeatedly testing near-identical ideas.

The Meetup example defied an easy assumption. A required pledge during event creation seemed likely to deter organizers, since extra steps often reduce sign-ups. The pilot instead reduced spam and increased successfully created events by 16 percent. The reminder strengthened organizers’ sense of purpose. The point is not that extra steps always help, but that plausible assumptions are not substitutes for testing. A pilot should be small and quick enough that failure is inexpensive.

Analysis

Distinct pilots turn uncertainty into useful comparison. If every test uses the same kind of message, a team can learn little even when the numbers differ; a wider range of approaches can expose which pressures matter. The Meetup result is a reminder that a step can both add friction and reinforce motivation. However, combining many components can make a successful intervention hard to explain. That may be acceptable when the goal is a workable outcome, but it becomes limiting when teams need to know which component is safe, effective, or worth scaling.

Make behavior change serve people’s motives, and expose its trade-offs

Good intentions do not settle ethics. The book distinguishes an intention-action gap, when someone wants an outcome but struggles to act, from an intention-goal gap, when someone wants a goal but rejects the proposed route. Designers should not assume that a desired outcome gives them permission to push any behavior. The behavior should serve a motivation the population actually holds, and its benefits should not outweigh costs to other important motives.

Transparency adds accountability. In a flu-shot project, some members wanted vaccination but faced practical obstacles; others did not intend to get one. A message about protecting family or congregation could connect vaccination to an expressed motive. By contrast, frightening people with a false claim about harming their relatives would distort the choice. The team should be willing to describe its goal and intervention publicly, and to accept responsibility for resulting harms.

Analysis

The ethical framework is strongest when it treats people as agents whose own goals matter, rather than as targets to be moved. Its emphasis on transparency also counters the tendency to judge an intervention only by its intended benefit. Public disclosure is not a perfect safeguard: people may disagree about costs, and some interventions can be disclosed without becoming fully understandable. Still, inviting scrutiny from people who do not share the designer’s incentives makes self-justification harder. Ethical review should examine both what action is encouraged and the means used to encourage it.

Pilot cheaply, test seriously, and monitor results after scaling

Use evidence in stages. A pilot is a small, fast, deliberately low-cost attempt to see whether an intervention shows promise. A test exposes a more developed version to a larger group and assesses whether its impact justifies the resources required. Scaling follows only when the expected behavior change is worth the effort. Small pilots also reduce the emotional and organizational investment that can make teams defend a failing idea.

The process continues after launch. An intervention can lose effectiveness as circumstances change or collide with other interventions aimed at the same behavior. The book calls this interaction the piranha effect: a growing portfolio can dilute or cannibalize the impact of individual efforts. Track the target behavior continuously, use alerts rather than relying on someone to open a dashboard, and be ready to revise or stop work that no longer helps.

Analysis

Staging decisions is a practical way to balance learning against the cost of exposure. It also creates permission to abandon ideas without treating every abandoned project as wasted effort. Measurement needs care, however. A small pilot can suggest a direction, but weak evidence should not be presented as proof, especially when the consequences of error are serious. Nor does a positive result guarantee durable impact. Continuous monitoring makes the intervention accountable to changing conditions, while a record of null and negative results lets future teams learn from more than successes.

Use identity and attention as contextual pressures, not fixed customer types

People respond through shifting identities. Wallaert treats identity as a source of both encouragement and restraint: a person may act to affirm a group they belong to or avoid resembling a group they reject. Priming makes an existing link between identity and action more noticeable; moderation strengthens or weakens that link; mediation creates a new connection through a shared value. A campaign can therefore do more than remind someone of a label. It can reshape what that label means in a specific context.

Attention changes the strength of cues. People have limited cognitive resources, so defaults, salient reminders, and clear information can have greater influence when someone is busy or overloaded. Yet not everyone wants effort removed. Some prefer automatic choices; others value researching and customizing. Treat identity and attention as questions to investigate, not stereotypes to assume. The same intervention can support belonging for one group and feel alienating to another.

Analysis

The identity lens is useful because it explains why a single message can affirm one person and repel another. It also encourages designers to examine the values attached to a behavior, rather than relying on demographic labels as shortcuts. But identity research is especially vulnerable to overgeneralization: group averages do not reliably describe each person, and the relevance of an identity changes by setting. The same caution applies to attention. A default can reduce overload, but it can also quietly steer someone who would prefer to deliberate.

When eliminating a behavior, replace the motivation it used to satisfy

Removing an action can leave a vacuum. The same process used to increase behavior can be reversed to reduce it, by strengthening barriers or weakening attractions. But if the behavior served an underlying motive, suppression alone may send people toward a substitute. The replacement could be equally harmful or simply less visible, so designers should return to the motivation and identify a better way to meet it.

Smoking illustrates both approaches. Warnings, restrictions, and taxes increased barriers, while anti-smoking campaigns also challenged the image of smoking as attractive. Yet the desire for social connection, identity, and ritual remained available to be met by alternatives such as e-cigarettes. A different example in the book concerns adding salt to bland food: rather than merely urging people to stop, the intervention offered spices that could make meals appealing. Replace the function, not just the behavior.

Analysis

This principle resembles substitution in systems design: removing one component rarely removes the need that component served. It is especially important in public health, where a behavior can provide social connection, relief, status, or routine as well as its obvious physical effect. The replacement need not reproduce every feature of the old behavior, but it must be compelling enough to compete. Designers should also watch for unintended substitutes, since reducing a visible behavior can make a problem appear solved while its underlying pressures remain unchanged.

Analysis

The book’s central contribution is not a new discovery about human motivation so much as a disciplined bridge between behavioral science and the decisions organizations routinely make. Its strongest move is procedural: define the behavior first, insist that assumptions face multiple forms of evidence, map forces on both sides, and treat implementation as a sequence of reversible bets. The Bing classroom story gives this method a memorable proof of concept. Curiosity sounded like the obvious lever, but classroom constraints, privacy concerns, and teachers’ responsibility explained the gap more convincingly. The result was a better intervention and a measurable outcome.

The framework also usefully reframes design as changing conditions rather than controlling people. That distinction supports both analytical clarity and ethical restraint, especially when an intervention helps people act on motives they already hold. Its warnings about identity, attention, and replacing behaviors extend the process beyond simple interface friction.

Some claims require careful handling. The book’s discussion of p-values simplifies statistical interpretation: a p-value is not the probability that a specific intervention is ineffective. It describes how surprising data would be under a specified no-effect model. Decisions should combine effect size, uncertainty, costs, possible harms, and the quality of the study, rather than treating a single threshold as a verdict. Likewise, the claim that inhibiting pressures are often especially effective is a productive corrective to promotional thinking, not a universal law. Removing a barrier can be powerful, but barriers can protect privacy, deliberation, or safety.

The IDP is most persuasive as a learning system, not a recipe guaranteeing success. Its lasting value is organizational: it makes goals explicit, makes failure informative, and requires successful interventions to keep earning their place. The method can improve products, services, and public programs, but only when teams measure outcomes that matter to the people affected, not merely outcomes convenient to the organization.

Summary in 30 Seconds
Start by defining a measurable action that serves a real human need, then verify the problem through independent evidence. Map what encourages the action and what obstructs it; reducing friction can work better than adding persuasion. Classroom search rose 40 percent, home search 15 percent, and a Meetup pilot increased successful event creation by 16 percent. Test varied, inexpensive interventions, assess harms and trade-offs, and keep monitoring after launch.
Contains spoilers

Last updated:

Report Issue

Review Summary

3.81 out of 5
Average of 433 ratings from Goodreads and Amazon.

Start at the End receives mixed reviews, with an average rating of 3.83/5. Readers appreciate its practical approach to behavior-driven product development, finding it insightful and applicable across various fields. The book's casual tone and real-world examples are praised by many, though some find it overly chatty. Critics note the book could benefit from more concrete examples and visual aids. Overall, readers value the book's framework for creating impactful products and its emphasis on ethical considerations in behavior change.

Your rating:
4.4
120 ratings
Want to read the full book?

Glossary

Intervention Design Process

Behavior-change design cycle

The book’s structured method for designing and evaluating changes intended to produce a measurable behavior. It moves from identifying and validating an opportunity to defining the target behavior, mapping pressures, designing and selecting interventions, reviewing ethics, piloting, testing, deciding whether to scale, and monitoring performance.

Potential insight

Possible opportunity for change

An observation that the current world differs from a more desirable possible world in a way that behavior change might address. Potential insights can arise from quantitative patterns, qualitative observation, organizational assumptions, or information from outside the organization. They are hypotheses to validate, not established facts.

Behavioral statement

Measurable behavior goal

A concise description of the behavior an intervention seeks to produce, including the target population, its motivation, relevant preconditions outside the designer’s control, the behavior itself, and how that behavior will be measured. It translates a broad aspiration into a specific outcome teams can evaluate.

Promoting pressures

Forces encouraging behavior

Factors that make a particular behavior more likely. They may include motivation, convenience, identity, social context, attractive features, or cues in the environment. Their influence depends on the population and situation, and they operate alongside inhibiting pressures rather than determining behavior alone.

Inhibiting pressures

Forces discouraging behavior

Factors that make a behavior less likely, such as cost, effort, distance, uncertainty, limited availability, or concerns about consequences. An intervention can increase a behavior by reducing these barriers, or decrease it by strengthening them. Their effect varies with context and can interact with promoting pressures.

Convergent validity

Agreement across evidence sources

Support for a conclusion from different kinds of evidence, such as behavioral data, direct observation, interviews, and external research. In the book’s process, convergent validity reduces reliance on any one method and helps teams challenge assumptions before designing interventions.

Optimum distinctiveness

Varied, low-overlap options

A principle for selecting interventions to pilot. Rather than testing several nearly identical ideas, teams choose a differentiated set that covers more possible approaches with limited overlap. This increases the range of information gained and improves the odds that at least one intervention will change behavior.

Piranha effect

Interventions erode each other

The reduction in an intervention’s impact when it operates alongside other interventions aimed at the same or competing behaviors. A portfolio may produce less combined change than its separate interventions suggest, because efforts can compete for attention or cannibalize one another’s effects.

Priming

Activate an existing identity link

An intervention that makes an identity more salient so that an already established connection between that identity and a behavior becomes more influential. Priming is most useful when the identity-behavior association is clear, relevant to the situation, and meaningful to the people involved.

Moderation

Change an identity link’s strength

An intervention that strengthens or weakens an existing association between an identity and a behavior. Unlike priming, which makes an identity more salient, moderation changes how strongly that identity is connected to the behavior. The book uses the Always campaign about running like a girl to illustrate weakening a limiting association.

Mediation

Create a connecting value

An intervention that creates a new pathway between an identity and a behavior by linking them through a value or intermediate idea. The book’s example connects computer use with care, making technology more compatible with a nurse’s existing identity and values.

About the Author

Matt Wallaert is a behavioral scientist and author known for his work in applying psychological principles to product design and behavior change. He has experience working with major tech companies, including Microsoft, where he contributed to the development of Bing. Wallaert's writing style is described as conversational and engaging, often incorporating personal anecdotes and humor. He emphasizes the importance of ethical considerations in product development and behavior change strategies. Wallaert's approach combines academic knowledge with practical industry experience, making complex concepts accessible to a wide audience.

Download PDF

To save this Start at the End summary for later, download the free PDF. You can print it out, or read offline at your convenience.
Download PDF
File size: 0.37 MB     Pages: 17

Download EPUB

To read this Start at the End summary on your e-reader device or app, download the free EPUB. The .epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.
Download EPUB
File size: 1.43 MB     Pages: 21
Want to read the full book?
Follow
Listen
Home
Swipe
Library
Get App
Try Full Access for 3 Days
Listen, bookmark, and more
Compare Features Free Pro
📖 Read Summaries
Read unlimited summaries. Free users get 3 per month
🎧 Listen to Summaries
Listen to unlimited summaries in 40 languages
—
❤️ Unlimited Bookmarks
Free users are limited to 4
—
📜 Unlimited History
Free users are limited to 4
—
📥 Unlimited Downloads
Free users are limited to 1
—
Risk-Free Timeline
Today: Get Instant Access
Listen to full summaries of 26,000+ books. That's 12,000+ hours of audio!
Day 2: Trial Reminder
We'll send you a notification that your trial is ending soon.
Day 3: Your subscription begins
You'll be charged on Nov 25,
cancel anytime before.
Consume 2.8× More Books
2.8× more books Listening Reading
Our users love us
1M+ readers
Trustpilot Rating
TrustPilot
4.6 Excellent
This site is a total game-changer. I've been flying through book summaries like never before. Highly, highly recommend.
— Dave G
Worth my money and time, and really well made. I've never seen this quality of summaries on other websites. Very helpful!
— Em
Highly recommended!! Fantastic service. Perfect for those that want a little more than a teaser but not all the intricate details of a full audio book.
— Greg M
Save 62%
Yearly
$119.88 $44.99/year/yr
$3.75/mo
Monthly
$9.99/mo
Start a 3-Day Free Trial
3 days free, then $44.99/year. Cancel anytime.
Unlock a world of fiction & nonfiction books
26,000+ books for the price of 2 books
Read any book in 10 minutes
Discover new books like Tinder
Request any book if it's not summarized
Read more books than anyone you know
#1 app for book lovers
Lifelike & immersive summaries
30-day money-back guarantee
Download summaries in EPUBs or PDFs
Cancel anytime in a few clicks

Why one summary?

Scanner
Find a barcode to scan

⭐
⭐
⭐
We have a special gift for you
Open
38% off
DISCOUNT FOR YOU
⭐
⭐
⭐
$79.99
$49.99/year
Only $4.16 per month
Continue
2 taps to start, super easy to cancel
Settings
General
Widget
Loading...