A curve that is easy to remember
The hype cycle is a picture many people have seen: a line that climbs sharply to an early peak, drops into a deep dip, and then climbs more gently to a steady level. It is closely associated with the research and advisory firm Gartner, which has published versions of it for emerging technologies since the 1990s, and the term has since spread well beyond that setting to describe almost any topic that gets over-excited attention.
Its appeal is obvious. It offers a simple narrative for an experience that many people recognise: great promises, followed by disappointment, followed by quieter usefulness.
The five stages
The model is usually described in five phases:
- Innovation trigger. A breakthrough, demonstration or announcement draws first interest, often before there is a working product.
- Peak of inflated expectations. Publicity grows, early success stories circulate, and expectations rise beyond what the technology can currently deliver.
- Trough of disillusionment. Experiments fall short, some ventures fail, and interest and attention decline.
- Slope of enlightenment. More realistic understanding emerges as organisations learn what works and where it fits.
- Plateau of productivity. The technology becomes mainstream and its benefits are broadly accepted.
The vertical axis is typically labelled “expectations” or “visibility”, not usage or quality. That detail matters. The curve is meant to describe how attention and expectations change, not how many people use something or how well it works.
What the model gets right
It captures something real about human and institutional behaviour.
Early excitement often outruns capability. New ideas tend to attract promises that are easier to make than to keep, particularly when the people making them gain from enthusiasm. Media amplification and promotional incentives can push expectations higher.
Disappointment often follows. When reality fails to match inflated hopes, enthusiasm can swing the other way, and sensible applications get dismissed along with the exaggerated ones.
Useful things sometimes survive the dip. Some technologies do become unremarkable parts of daily life after the noise dies down. The reminder that early hype and long-term value are separate questions is valuable.
It encourages patience and scepticism. Used as a prompt rather than a prediction, it can help people ask whether current expectations are realistic.
Why it is a story, not a law
It is drawn after the fact. The curve is largely a stylised description, and it is easy to fit past examples to it by choosing which events count as the peak and which as the trough. A model that can be adjusted to match anything has limited predictive power.
Many things do not follow it. Some innovations rise steadily without a dramatic dip. Some peak and vanish without ever reaching a plateau. Some never attract much hype at all but still change daily life. The cases that fit the curve are the ones usually offered as illustrations, which is a form of survivorship bias.
Stage boundaries are fuzzy. It is hard to know, at any given moment, which stage a technology is in. A topic that looks like it is in the trough could be heading toward the plateau, or toward oblivion. Without a way to tell, the model offers little guidance on timing. The same difficulty applies to forecasts of how work will change, as seen in automation forecasts, which often diverge from what happens.
The timelines are not fixed. Statements about how long each phase lasts are broad generalisations. Real durations vary greatly and there is no solid basis for assuming a standard schedule.
It treats a technology as one thing. In reality, a field can contain many separate applications, some rising and some falling at once. A single curve can hide this variety. Applying one curve to a loosely defined area, such as a whole category of consumer technology, can obscure more than it reveals.
Attention is hard to measure. Visibility is not a single quantity. Search activity, news mentions, investment and social posts can move differently, and each has its own biases. How to read the signal from search interest charts is one example of why the vertical axis is not as clean as the drawing implies.
It can become self-fulfilling talk. Once the idea is well known, people may use it to dismiss anything new as “just hype” or to excuse failures as a normal stage, neither of which is a fair assessment on its own.
Using it sensibly
The most reasonable way to use the hype cycle is as a description of a common pattern in expectations, not as a map of any specific case. Useful questions include: what is being promised, and who is promising it; what evidence exists of actual use; how much does current excitement depend on future claims; and what would count as evidence that a given area is moving into a more realistic phase.
Recent areas of intense attention, such as crypto adoption, virtual reality or chatbot adoption, are often discussed in hype-cycle terms. Whether a particular example fits depends on how it is defined and measured, and confident claims about where it sits on the curve should be treated as opinion rather than finding.
The model’s lasting value is modest but real. It reminds us that excitement and usefulness are different things, that both exaggerated hope and exaggerated disappointment are common, and that the sensible response to either is to look for evidence of how people actually use something over time.