Threshold Commitment
Commitment doesn't build up steadily; it occurs in sudden shifts. Companies that view relationship-building as a process of accumulating goodwill fail to notice the truly important moments.
Almost all retention and conversion efforts are based on a mistaken understanding of how commitment develops. The actual process is one of punctuated change: provisional interest continues, occasionally for many years, and then a single qualifying action makes the relationship durable. When the threshold is not reached, nearly all of the relationship's value remains latent, but when it is reached, the relationship becomes self-reinforcing.
A health system has a patient portal with 60,000 registered users, 12,000 of whom have logged in more than once, and 3,000 of whom have arranged an appointment using the portal. Four hundred have done something the system did not anticipate: they have set a recurring reminder, saved a provider as a favorite, uploaded a document, or taken some other action that provided no immediate practical benefit. These 400 have therefore surpassed the other 59,600 users. However, the portal team is not aware of this; they only measure the number of logins.
The difference is important since the behavioral pattern of these four hundred is fundamentally unlike that of the rest. They have actually put in a small but real effort towards the relationship. This effort results in psychological pressure toward consistency: since they have treated the system as something worth personalizing, they are now more likely to use it again, to finish the tasks they begin in it, and to say that the system has value when asked about it. When people are below the threshold, convenience dictates their behavior; when they are above it, identity takes over.
That is the kind of thing that commitment theory refers to as escalation of commitment—the well-documented tendency for future investment to increase as a result of prior investment, not because the reasoning has changed but because of the identity that has been established. Although Robert Cialdini showed the mechanisms behind commitment and consistency with great clarity, the implications for organizational design go beyond what he originally presented. It is not sufficient merely to understand that commitment is self-reinforcing once formed. What leaders need to consider is how to design the situations in which it occurs, and how to realize when such situations are being overlooked.
What Commitment Is Thresholded, Not Linear
In most marketing and retention efforts, the prevailing approach views commitment as a gradual process: each interaction contributes a little to the relationship's equity, and only after a sufficient number of such contributions is loyalty achieved. Although this model is intuitive and can be measured with engagement metrics, it is largely incorrect as a description of how commitment actually develops.
Over several decades of behavioral research on cognitive dissonance and self-perception theory, a punctuated model has emerged. People do not gradually become committed to a particular course of action; instead, they start with a temporary attitude towards it, one that involves low switching costs and little identity investment, until they carry out an action which is costly enough, noticeable enough, or public enough to necessitate an update of their self-concept. This update marks the threshold; before it, the relationship can be broken off, but after it, the individual has a stake in maintaining consistency with the position they have now adopted, either implicitly or explicitly.
The result is that two audiences that appear similar when assessed by behavioral indicators can, in fact, be in completely different situations. A member of the audience who has attended four webinars, opened six emails and visited the pricing page twice can be considered behaviorally engaged, even though they have not reached a commitment threshold. On the other hand, an audience member who completed a one-question form asking for their company size has made a smaller behavioral investment but has nevertheless carried out something qualitatively different: they have taken a self-defining action. By doing so, they have defined themselves as someone who engages in such behavior, and this action has the kind of commitment weight that attendance at webinars does not.
Two audiences that appear identical when looking at engagement figures can be in completely different situations: one has built up a series of interactions, while the other has passed a certaint threshold. Only one of them is interested in carrying on.
The Threshold Identification Problem
Most organizations won't tell you what their commitment thresholds are. Although they are aware of their conversion rates, churn rates, and engagement curves, they don't know whether any particular actions indicate a durable commitment better than accumulated engagement does.
The identification process is based on empirical evidence rather than theory. The question then becomes: which early behaviors are better at predicting long-term retention than behavioral volume alone? In the case of a software product, this could be the action of completing a particular setup step that involves the user linking their existing data; for a nonprofit, it could be making a second donation within ninety days of the first one, no matter how large or small the gift; and for a health center, it could be whether a patient brings a family member to an appointment, whether they contact the center between visits about a non-urgent matter, or whether they tell someone else about the organization.
Research into user retention carried out by Duolingo has produced a well-documented example. They found that users who performed a particular early action—different from merely logging in or finishing lessons—had retention curves that diverged markedly from those of users with a similar level of early engagement who hadn't performed that action. Although the threshold involved was not the most obvious or easily understood behavior in the product, it was determined through empirical analysis by identifying the smallest behavioral indicator that predicted long-term retention more effectively than overall usage figures did. As a result, the product design was changed to make it easier to meet that specific threshold.
The method used across different industries is identical: it involves selecting groups of customers who have maintained long-term relationships and then looking back to identify the common characteristics these relationships had at 30, 60, and 90 days—which the relationships that eventually churned lacked. The explanation is generally not the most common early behavior; rather, it is a particular action, frequently one that involves a small identity cost, which sets the committed customers apart from those who are only engaged.
The threshold is usually not among the first things people do; it involves a small sacrifice of identity. To locate it, one has to look back at the relationships that have been kept, not look forward to the engagement figures.
Public Commitment Amplifies the Effect
Not every act of committing is the same; research on public and private commitment shows that commitments made publicly, or those visible to people whose opinions the individual values, are considerably more durable than commitments made privately.
The process is one of reinforcing social identity. When someone makes a private commitment, they experience cognitive pressure to remain consistent with themselves, whereas making a public commitment produces the same pressure, along with the additional social disadvantage of having to reverse their position. A person who tells their friends they are training for a marathon is much more likely to follow through than someone who commits privately. Similarly, a person who states in a meeting that they support a new initiative is much more likely to continue to back it when difficulties arise in implementing it.
For organizations, this raises a particular design question: where in the relationship one can encourage a public expression of commitment that is both genuine and suitable for the stage of the relationship in question; for a professional services firm, the request for a referral is a typical example. It is not merely a business development move to ask a client to introduce you to a colleague; it is an invitation involving a crossing of a threshold, since the client who makes the introduction has now publicly aligned himself with the organization and there is in fact a real sense of cognitive consistency pressure to maintain that alignment. Companies that request a referral early, before the relationship has generated enough value to make the client completely at ease, obtain fewer referrals, and those that do are weaker. On the other hand, those organizations that request at the exact moment when the relationship has reached a stage of real client satisfaction secure referrals with the strength of a true endorsement rather than just polite compliance.
Community health organizations also have similar arrangements through patient advisory boards, community health worker networks, and ambassador programs. The patient, who is a member of an advisory board, has made a public commitment to the organization that the patient who attends appointments has not. Although both kinds of relationship are valuable, the first is more enduring, and so too is the influence this patient has on the other members of their network. Setting up opportunities for patients to make public commitments at the right stages of their relationship is one of the most effective retention investments an organization can make. Yet, most health organizations do not intentionally pursue this approach.
Steps to Design for Threshold Crossing
The approach to designing with threshold crossing differs in structure from that of designing to accumulate engagement. The aim is not to increase the number of touchpoints but to have fewer moments that are well placed and encourage a qualifying action.
The way to proceed is to determine the threshold empirically before designing anything based on it. Identify the group of users who have held on the longest and are the most deeply committed, then work backward to find the early behaviors that set them apart from those who have churned or become disengaged. Carry out this cohort analysis before taking it for granted that the threshold is the most obvious behavior, the easiest to measure, or the one intuition leads you to believe is most important. The threshold is generally neither the first high-engagement behavior nor the one that requires the greatest effort; it is the first action that indicates an investment in identity rather than instrumental use.
The next step is to reduce friction specifically around the threshold action, not to reduce friction generally. While reducing friction everywhere may improve engagement metrics without necessarily leading to more threshold crossings, targeted friction reduction involves asking: what is making it harder than it should be for someone ready to cross the threshold to do so? In the case of a patient portal, this could be the setup step needed to save a provider, which requires too many clicks and is located deep inside a settings menu that no one goes to. By moving that step earlier and making it part of the prompted onboarding process, you are not only improving UX but also making a threshold design decision.
The third step involves creating a micro-commitment ladder which leads up to the threshold, rather than considering the threshold as the very first request. There is clear evidence from behavioral research on the foot-in-the-door phenomenon that a small initial commitment increases the likelihood of a larger subsequent commitment. This ladder is not intended as a means of manipulation; instead, it is about ensuring that requests follow naturally from the stage of the relationship, so that each one becomes a natural follow-on to the previous commitment rather than a sudden escalation. It would be unreasonable to ask a new patient at a health center to fill out a full social needs assessment on their first visit since the relationship has not yet built up the trust that makes such a request appropriate; on the other hand, a center that asks for one additional piece of information at each visit and ends up with a more complete picture after three appointments is building the ladder in a manner that takes account of where the relationship really is.
The fourth step is to provide a visible sign indicating that a threshold has been crossed when possible. The commitment effect is most effective when a person can see what they have done; confirmation screens, the welcome messages sent when joining a program, a named status change, a physical item, or social recognition all serve the same purpose, which is to make the fact of the threshold crossing clear to the individual who has just crossed it, thus strengthening the identity update and the consistency pressure that comes after. Amazon Prime's annual membership renewal is as much a form of threshold architecture as it is a business model; the decision to renew, made each year explicitly, constitutes a recurring act of commitment. Companies that permit threshold crossings to occur silently are failing to fully realize the commitment effect.
What Changes When You Design This Way
When organizations shift from accumulating engagement to setting thresholds, they make different kinds of investments and measure different outcomes.
Precision
The focus is now on precision rather than on volume. Rather than trying to increase the number of touchpoints, the issue becomes one of identifying the single action that will alter the course of the relationship and removing any obstacles in its way. This usually involves reducing what is done, not increasing it: running fewer campaigns but placing them more deliberately, making fewer requests but sequencing them more carefully, and monitoring fewer metrics while paying closer attention to the specific behavioral signal that actually predicts customer retention.
Threshold Crossing Rates
The focus is now on threshold crossing rates rather than engagement rates. What proportion of newly engaged relationships reach the threshold within a specified time frame? What is the conversion rate from provisional to committed within the first 30, 60, or 90 days? What is the difference in retention between relationships that did cross the threshold and those that did not? It is these figures that provide a different indication of what the organization should be doing than click rates, open rates, and page views.
Relation
The way in which an organization relates to its audience also changes. Those organizations that aim to accumulate engagement see every point of contact as a chance to add value and assess whether the audience has noticed. On the other hand, organizations that aim to cross a threshold treat most contact points as leading up to a small number that actually matter, and they give disproportionate attention to the quality and positioning of those few. The audience on the other side does not always consciously perceive the difference; instead, they come to feel that the relationship was worth committing to at some stage and then appeared to carry on all by itself without them having to put in any effort. That is no coincidence; it is the result of threshold design.
Continuous nurturing builds the relationship to the threshold. What happens at the threshold is a different design problem entirely, and most organizations have never solved it.
Some ideas are worth discussing in the context of your organization.
