i

Description

Char de parade de cirque (« bandwagon ») richement sculpté et peint, aux roues multicolores

Source

Carol M. Highsmith · original page

Licence

PD Public domain

Original photo, unretouched. The background is a version of it stylised by the site.

The Bandwagon Effect (Trend Effect)

When everyone is doing it, we rarely ask whether it is actually in our own interest.

In brief

• What: the popularity of an option becomes, by itself, a reason to choose it.
• Why: following the majority is often a good shortcut (they may know something), and diverging from it carries a social cost.
• Detect: the main argument is “it’s what everyone does”; no one can explain what the option solves for us.
• Avoid: state your opinion before hearing others, start from the need, require a stopping criterion and quantified feedback.

On this page

1. Definition
2. The mechanism
3. Where it shows up
   1. In project management
   2. In management
4. Famous cases
   1. Solomon Asch’s conformity experiment 1951-1956
   2. Segment: moving from microservices back to a monolith 2018
   3. Prime Video: a monitoring tool moved away from microservices March 2023
5. Detection signals
6. Avoidance reflexes
7. With AI: detect, fix, mitigate
   1. Detect
   2. Fix
   3. Mitigate
   4. Prompts to integrate into your workflows
   5. How to integrate them
   6. Precautions
8. Limits
9. Sources and verification
10. See also
    1. Combines with, or reinforces
    2. Related biases
    3. Counterweights and related reflexes

Definition#

Fictitious illustration

The bandwagon effect (bandwagon effect, literally “the marching band wagon effect”) refers to the fact that a belief, behavior, or choice becomes more attractive as the number of people who have adopted it increases. Economist Harvey Leibenstein described it in 1950 for consumption: some people buy a good because others are buying it.

In organizations, it takes two related forms. The first is conformity in a meeting: you give up your opinion, or do not voice it, because others think differently. The second is the trend effect: a technology, framework, or architecture becomes the “obvious” answer because the whole industry is talking about it.

The mechanism#

Two influences are intertwined, distinguished by Deutsch and Gerard as early as 1955:

• Informational influence: if many people make this choice, they may know something I do not. That is often reasonable, but it becomes a pitfall when everyone relies on others without anyone truly checking.
• Normative influence: we want to be accepted, not be seen as the only one who does not understand or who is holding the group back. We align with the common opinion even when we believe it is wrong.

Economists Bikhchandani, Hirshleifer, and Welch (1992) showed how these sequential choices produce informational cascades: after a few initial matching choices, everyone ignores their own private information and imitates, so the collective decision rests on very few real signals and can stay wrong for a long time.

Where it shows up#

In project management#

• Technical choice: “big players in the industry use microservices, containers, generative AI, low-code”; we adopt before measuring the benefit for our team, our data volumes, our maturity level.
• Method: a large-scale agile framework is rolled out because neighboring companies did it, without defining the problem it is supposed to solve.
• Design review: the first reviewer approves; the next ones validate “since it’s already been reviewed.”
• Recruiting and training: we look for skills in a technology whose value has not been established, simply because job postings for it are multiplying.

In management#

• Team meeting: after three matching opinions, the fourth falls in line, even if skeptical; silence is taken as agreement.
• HR practices: an evaluation tool, ritual, or metric is adopted because “every company does it.”
• Company policy: a reorganization, “transformation plan,” or employer brand is launched by copying a highly visible competitor.

Famous cases#

Detection signals#

• “Everyone does this now.” / “It has become a standard.” / “We can’t stay out of it.”
• Justifications cite company names or conferences, never numbers from our own context.
• An option is adopted in a few weeks, while a more modest, better-documented idea waited for months.
• In meetings, the first opinions voiced are followed by a few nods and then unanimous silence.
• It is impossible to state what would make us give up the option, or the specific problem it was supposed to solve.
• Skeptics are described as “behind,” “resistant to change”: labeling rather than a counterargument.

Avoidance reflexes#

1. Collect opinions before discussing. Each person writes down their position (or estimate) before hearing from others; reveal them all at once. The earlier the manager's or majority's opinion is known, the more weight it carries.
2. Start from the need, not the tool. Write down the problem, constraints, and success criteria before naming a solution. A trendy tool must address one line in this list.
3. Assign a challenger. An explicit devil's advocate role (rotating): disagreement becomes a task, not an act of opposition.
4. Ask for comparable lessons learned. Who did it, at what scale, and with what measured result? A testimonial from a very different company proves little.
5. Plan a time-boxed trial and a stop criterion. “Two-month pilot on this scope; we continue if a given indicator reaches a given threshold.” The decision is made reversible.
6. Look at failures and abandonments too. Published lessons learned are biased toward success (see Survivorship bias).
7. Protect minority voices. Actively seek out divergent opinions; thank the first person who says no instead of treating them as a blocker.

With AI: detect, fix, mitigate#

Prompts to integrate into your workflows#

1. Critical review of a trendy option. Before adopting a technology, method, or tool that is getting a lot of attention.

We are considering adopting: [option]. Our context: [team size, volume, constraints, skills, timeline].
I don't want to know whether the option is popular. Do a critical review:
1. under what conditions it produces poor results, and how our context is similar;
2. the hidden cost of implementing and operating it;
3. three simpler alternatives that would meet the same need;
4. what we should observe to decide against it.
Do not invent any statistics or company cases; flag what still needs verification.

2. Start again from the need. When the discussion starts with the name of a solution.

Here is what is being proposed: [solution]. Ignore its name.
Based on the following description, formulate the problem to solve, measurable success criteria, and constraints. Then propose at least three ways to meet the need, including one that requires no new tool, and compare them against these criteria.
Description: [context]

How to integrate them#

• Before the decision meeting: prompt 2 produces a need statement that everyone reads, then rates with their opinion.
• In the decision file: prompt 1 provides the critical review, to be added to the appendices.
• After one year: compare the outcome against the success criteria written at the start.

Precautions#

• A model can be influenced by wording: if you write that the option is “essential,” it will look for its strengths.
• The company cases and figures cited may be invented: verify every reference before reusing it.
• AI itself has a trend effect: it more readily recommends what is abundant in its data.
• Do not send confidential company data to an unauthorized tool.

Limits#

• Following the majority is not always a mistake. Under uncertainty, common practice is often a good starting point; imitation is a bias when it replaces analysis.
• Asch's results were obtained with American students, in the 1950s, on a low-stakes task. Replications found varying levels of conformity across periods and cultures (Bond and Smith, 1996).
• A technical rollback (Segment, Prime Video) is a company case: it shows a risk, not a law.
• Do not swing to the opposite extreme, the systematic rejection of what is popular, which is another form of dependence on others' opinions.

Sources and verification#

• Asch, S. E. (1951), “Effects of group pressure upon the modification and distortion of judgments,” in H. Guetzkow (ed.), Groups, Leadership and Men, Carnegie Press — the line experiment.
• Asch, S. E. (1955), “Opinions and Social Pressure,” Scientific American, 193(5) — effect of majority size and the dissenting partner.
• Asch, S. E. (1956), “Studies of independence and conformity: I. A minority of one against a unanimous majority,” Psychological Monographs, 70(9)
• Deutsch, M. & Gerard, H. B. (1955), “A study of normative and informational social influences upon individual judgment,” The Journal of Abnormal and Social Psychology
• Leibenstein, H. (1950), “Bandwagon, Snob, and Veblen Effects in the Theory of Consumers' Demand,” The Quarterly Journal of Economics
• Bikhchandani, S., Hirshleifer, D. & Welch, I. (1992), “A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades,” Journal of Political Economy
• Bond, R. & Smith, P. B. (1996), “Culture and conformity: A meta-analysis of studies using Asch's line judgment task,” Psychological Bulletin
• Noonan, A. (2018), “Goodbye Microservices: From 100s of problem children to 1 superstar,” Segment engineering blog (July 2018).
• Kolny, M. (2023), “Scaling up the Prime Video audio/video monitoring service and reducing costs by 90%,” Prime Video Tech blog (March 2023).

See also#