How brands should use trend forecasting in marketing strategy
How does Netflix know what type of shows you would like to watch next, and what content to produce or acquire next? The answer is trend forecasting. Netflix uses trend forecasting to optimize its marketing and personalization strategies.
Similarly, if you look at Spotify, another streaming platform for music, it uses trend forecasting to create the ‘Discover Weekly playlist’ with new trends in artists and songs in the music industry. Accordingly, it promotes them through its editorial playlists, podcasts, social media, etc.
Another example is Lush, the cosmetic brand. A few years ago, Lush observed a growing trend among its consumers: a heightened awareness of the environmental and social implications of their purchases. They sought increased transparency and accountability from the brands they endorsed. In response to this shift in consumer consciousness, Lush initiated the ‘Naked Range’ campaign.
Through this initiative, the company aimed to raise awareness about packaging-free products that minimize waste and carbon footprint. Here trend forecasting played a role to help Lush’s marketing team grasp what truly resonated with its audience in the first place.
In the sports industry as well, betting experts rely on historical data of players and statistical models to accurately forecast the winners of nearly every stage of the tournament.
Is forecasting and prediction the same thing?
In many cases, both terms are used interchangeably. That’s because both use data, research, and other information to conclude a pattern. However, the main difference lies in how forecasting is used in the strategies.
Trend prediction helps us determine what new pattern shifts will emerge, while trend forecasting uses data to determine how already-identified pattern shifts will play out.
Short-term vs. long-term trend forecasting:
Trend forecasting often falls into one of these two time-related categories — short-term and long-term.
Short-term trend forecasting helps us understand how trends, especially micro-trends, will progress in the immediate future. This kind of forecasting looks ahead to the next month, quarter, season, or couple of years.
Long-term trend forecasting projects how trends, especially macro-trends, will progress over years, decades, or longer.
In general, long-term forecasting requires data sets that include longer periods than short-term forecasting.
Why is trend forecasting important?
If you are a fashion brand, you might want to know what colors and patterns will be in style for the coming season. Accordingly, you may want to optimize your inventories to meet that demand, making your brand more profitable.
Furthermore, if you are developing a technology-powered product, knowing what consumers will want in twelve months and tailoring the product accordingly will make the product launch more successful.
For your beauty brand, you need to use trend forecasting to build a marketing communication strategy for different audiences segmented according to their shared interests.
Do you use any audio or video streaming app? The answer is probably ‘yes’. Your apps enable you to create personalized playlists, which are curated selections of content such as songs and music tailored to match your taste, mood, or activity. Right?
In this context, the app utilizes trend forecasting by analyzing the user’s listening history, preferences, and feedback. These data points are then processed by algorithms and machine learning techniques to identify similar songs, artists, and genres that the user may enjoy.
Trend forecasting methods:
The specific steps you take to forecast trends will depend entirely on your needs, objectives, business environment, and more. In general, though, they will fall within 5 categories of trend forecasting methods:
• Time series model: This model assesses a trend based on its performance over time. So, you’ll look at trend data to predict a trend’s future growth.
• Economic model: Economists use this model to predict supply and demand curves.
• Judgmental model: This method is done formally or informally and can be the only option when hard data isn’t available. However, it is also much more susceptible to human error and bias.
• Delphi / qualitative model: This model uses a panel of subject matter experts and conversation leaders to form a decision. The idea is that these experts’ collective knowledge is more powerful and less biased than a single individual’s opinion.
• Statistical model: This model relies on large data sets (big data) and complicated, computer-assisted analysis to predict the future course of a trend. This is more accurate and objective than judgmental forecasting, but without the right data, it can also miss important context.
A practical guide to using trend forecasting:
The first step to trend forecasting is finding under-the-radar trends. So instead of identifying what’s hot right now, you want to find budding topics that show strong growth potential yet haven’t peaked.
This process is deceptively tricky because if a topic is too easy to find, it might have already peaked (or you’re too late to develop and launch a product that meets that trend).
If you’re using social media to find trending topics, you’ll be looking at trends that are hot right now. You can use the tool called “Exploding Topics” to surface early trends that are steadily gaining traction.
Now you need to prepare a comprehensive list of popular topics.
Then validate those popular topics with data and analyze their growth patterns.
Here you need to understand if the pattern has a spiky growth and sudden decline (known as fads) or if it steadily tends to have compounding growth over an extended timeline (known as trends).
If you want to use short-term forecasting, you can choose the topics from the fads. Else, you need to look for permanent, macro-level trends that will be relevant for the coming years or even decades.
Here you are all set with your list of topics. Choose one and discover related subtopics and the broader meta trends. Every meta trend may introduce you to another 10–15 related trending topics. For each topic, don’t forget to check the search volume data.
Now that you have a handful of promising trending topics. Narrow it down based on the search volume and growth pattern. But that’s not enough.
Here you need to look at consumer reviews, sales data, and competitive intelligence data, subject matter experts’ articles, and conversation leaders’ interviews to accurately forecast a trend’s market demand.
Once you select the trend to work with, you should keep tracking that regularly and check related sub-trends as well.
Thus, you can also keep track of short-term trends that may be used seasonally.
Tools used in trend forecasting:
Exploding Topics — to identify emerging trends in fashion and technology
Google Trends — to identify current trending topics and their search volume
Glimpse — an expanded version of Google Trends, to discover the trend by category or type in a keyword
Trend Hunter — to identify current trending B2C topics
Trends by Hubspot — to discover community (social media and tribe) identified business trends
CB Insights — for tracking tech funding trends which are used by entrepreneurs, tech investors, business executives, etc.
WGSN — to identify advanced fashion trends
Meltwater — for trend spotting and forecasting for PR
