Trend analysis: Is it just a hype or a real trend?
Dr. Tuomo Kuosa goes into detail about how Futures Platform's futurists differentiate a trend from a hype.
FUTURE PROOF – BLOG BY FUTURES PLATFORM
What is the right time to go to an emerging market?
If you are too early, the markets are not ready, and you end up losing time and money. If you are late, someone else will grasp the lion´s share of the growing market, and you end up being just a follower. Hence, how to know when the trend is real and the market is ready?
A trend is a prevailing development tendency of something. A trend is something that we consider to represent a development in the real world. It always has a time element and a direction. It grows, weakens or fluctuates on its way from history to the present and the future.
For example, a market trend is a perceived tendency of financial markets to move in a particular direction over time.
In theory, all trends can be quantified, but in practice, it may be too difficult to quantify some prevailing development tendencies with any reasonable data.
For instance, the development of the VUCA world is very difficult to be verified with any useful data set, as it is a meta-level phenomenon.
In other words, it is a combination of many quantifiable trends that together, and only together can be reasoned to be forming a larger phenomenon called VUCA. Its existence cannot be inferred from just one or two of its sub-trends such as exponentially increasing amount of data and speed of data transfer. Instead, there are dozens of things that, only when put together, allow us to make such a conclusion.
Alongside metalevel phenomena, many other types of trends and phenomena can justifiably be seen to be growing now or in the future, but are hard to be verified with historical data. This is an especially acute thing within different types of emerging issues, such as weak signals and wild cards.
According to the theory of foresight, if we can identify a thing that doesn´t exist yet, but which has powerful pushing drivers behind it and no major development restrictions, we may conclude that there are good chances that the issue will emerge and evolve into a growing trend at some time range.
The restrictions that could hinder the emergence can be, e.g., laws, political ambitions, exiting infrastructure or trends that move in the opposite direction. The pushing drivers that allow us to conclude that the emergence of trend strengthening is plausible can be quantifiable trends, such as the ageing of the population, but they may also be historyless events, such as a sudden political decision to invest in something or a sudden change in customer needs or the public mood or values.
Start a free trial of Futures Platform to take a closer look at the new feature and stay up to date with the latest developments of trends.
Are you already a user? Login to your account to explore how the new feature can benefit your work.
S-CURVE ANALYSIS
In Futures Platform, we use S-curve analysis or diffusion analysis, and so-called meta S-curve analysis as a basis for trends and phenomena analysis, categorisation and timing. We aim to use quantitative data sets and the principles of diffusion analysis as a backbone of all the analyses whenever it is possible.
In many cases, we can find useful data points from statistics or time series so that we can estimate the S-curve directly from the available data points without any need to add other drivers or explanatory functions to the econometric modelling. Yet, in most cases, the econometric modelling requires many modifications as the available data set doesn´t fit the dynamic of the simplest diffusion logistic function such as:
P(t)=δ_1/(1 e^(-δ_2-δ_3 t) )
The most commonly used variations of standard diffusion modelling, for getting suitable curves out of the available time series, are Gompertz-function or Bass´s diffusion model. The biggest modification to standard diffusion models that we have been using in Futures Platform is Trend Impact Analysis (TIA). That method forms a baseline trend from standard trend extrapolation but combines to it a set of disruptive events that have the potential to convert the trend from its expected historical track. In the end, all the diffusion analysis models use quantitative data points and attempt to identify what kind of factors affect the diffusion speed, and how the adaptors spread in the modelled timeline.
In some cases, such as in Trend Impact Analysis, qualitative assessments can be given numeric values that have been used alongside the historical data points.
Figure 1: S-curve analysis for e-commerce in EU area
The S-curve in figure one presents the extrapolated probable development of the main branches of e-commerce in the EU area until 2040. The used data points are from the years 2010-2018. The used standard model includes all known drivers and other affecting trends at the time of the modelling, as that is embedded in the used data, but it doesn´t contain any potential surprise elements, as is the case with the Trend Impact Analysis model.
HYPE CYCLE
Sometimes a newly emerging theme rises lots of interest before the actual trend emerges. This can be made visible if we add, e.g., Google Trends topic search data on top of the existing trend. The hype cycle in figure two comes from global searches on the topic and is presented in relative changes between different years, instead of actual numbers. The trends themselves show the total cumulative percentage of the companies that have been using the technology.
Figure 2: Hype cycle and the actual trends in usage of Big data analysis in EU area
Sometimes the hype may grow enormously, when the actual trend is still in its infant phase, as was the case with e-commerce and WAP services that enabled mobile internet use at the end of the 1990s. The incredible increase in value created by the hype was broken by the bursting of the dotcom bubble at the turn of the millennium. After the busted bubble, the public interest in e-commerce was very low and stayed that way. Yet the actual E-commerce trend continued to grow slowly in the first half of the 2000s and ended up being a new accelerating growth trend in the 2010s.
For example, tablet computers, various reliable mobile payment systems, artificial intelligence assistants, Instagram, and augmented reality came into the market, enhancing the online shopping experience and boosting online marketing.
It is quite common that a new technology or innovation to gain considerable attention and accelerated public interest when the journalists first start to write about the possibilities of the emerging thing. First, only a small group of experts and forerunners know about the new thing. Its technical details are debated just in field-specific journals until a significant news outlet takes it into its agenda, or a large corporation starts to rise a hype around its new innovation with its vast marketing budget.
A classic example of that was the Google glass. Google had made a prototype of the new type of AR glasses and started to rise hype around it in 2013. The hype became huge, but in 2014, Google withdrew the product from production before it ever came to the market. In that case, there was only the hype and no trend. The big question is, how can you know when the huge attention around a new thing is really about a real trend and not just hype?
META S-CURVE ANALYSIS
Alongside the different variations of diffusion analysis Futures Platform is using a Meta S-curve analysis. Its basic principle is that we form a theoretical S-curve for a broader theme area by aggregating all trend knowledge that we can have regarding it. The larger theme can be for example “Agriculture and food production”. Some things concerning agriculture, such as spinning jenny, have reached 100% of their potential a long time ago. Some things have nearly reached that but are still largely in use like the plough. All these are considered to be located at the final end of the curve and are therefore weakening trends marked with blue colour in our data bank.
Figure 3: Meta S-curve analysis
Some things in “Agriculture and food production” are still in their infant phase. These can be new technologies or practices. A new item, e.g., an AI robotic technology or a new super fertiliser is in its weak signal phase as long as only a small group of experts and forerunners know about it. At that phase, its technical details are debated among researchers and just in field-specific journals.
When the journalists start to write about the new thing in more popular papers, the thing becomes a strong signal. At that point, some exciting items go into a hype cycle and gain early oversized expectations. When the expectations are naturally not met right away, the interest in the thing goes to the valley of death for many years.
At the same time, the real trend lives its own life. At some point, the second cycle of the early adaptors and opinion shapers start using the new item and convince the larger group. That is usually the right time to go to the market.
The market beaks out when the larger group becomes “users” or adaptors. That is when the fast, extensive growth begins. At that phase, the market has more demand than supply, meaning that all suppliers can flourish. Later on, the competition in the market intensifies, meaning that suppliers must either boost their processes, scale-up, differentiate or merge to be successful in the new situation. At some point, the former “new” thing becomes obsolete as a new and better product replaces it, or it becomes fully utilised, and its trend cannot grow anymore. In both cases, the trend has reached its market saturation phase and starts to be considered a weakening trend, i.e., not growing anymore.
The benefit of the Meta S-curve analysis is in its ability to provide a grounded overview of the things that are new, old, emerging or in a fast-growth phase regarding the theme. It contains knowledge about the maturity of its different trends that helps to locate the maturity and probable growth pace of a new thing by comparing it to the other trends. In theory, similar trends and phenomena are strongly interlinked, and they move in the curve more or less together. They face similar restrictions, bottlenecks and drivers, and therefore accelerate simultaneously. If a new thing can be linked to an existing cluster of trends, we get an idea of how it probably behaves. That helps to differentiate potential oversized hype from the real trends.
If you are interested to learn more about trend and phenomena analysis, we recommend you to read the follow-up article “Cybersecurity Spending – Three Scenarios with Trend Impact Analysis“.
View more phenomena and quantitative analysis in our digital foresight solution by starting a 15-day trial.
To stay ahead, organisations must widen their lens, exploring not just their own industry but what’s happening on the edges.