Trend Forecasting

What Is Trend Forecasting?

Trend forecasting is the act of identifying market trends before they come into existence. Companies need accurate and comprehensive audience segment and social media data to predict product and market trends.

Why Is It Important to Have a Good Trend Forecasting Model?

Forecasting and trend analysis helps you monitor your market and take advantage of opportunities to meet business goals.

What Internal Data Should I Have for a Good Trend Forecasting Model?

To develop a good trend forecasting model, you will first need to clearly define your goals. For example, appealing to a new consumer segment and developing a five-year business plan need different strategies.

Secondly, you will need to use your own customer segmentation and transaction history. Even if you are trying to appeal to a new market or creating a new product, brand perception and similar products you used to sell will guide your forecast.

What External Data Is Essential for a Good Model?

A good trend forecast uses external data from events like conventions, trade shows, and fashion shows first. Increasingly, however, the forecasts use social media. Researchers analyze blogs and social media users with large followings and apply NLP and image recognition programs to their content to find the newest and most popular topics.

Good trend forecasting models also account for events (for example, holidays, royal marriages), news, and economic forecasts.

What External Data May Prove Useful for a Good Model?

Other useful data includes supply chain risk management, competitor analysis, industry data, and anything that may impact your business. The last point is broad and vague but essentially you must include external data that you think you need.

For example, you may sell wearable fitness devices and decide to track of political news coming out of the country in which a supplier mines cobalt for batteries you use in your devices because you fear changes to the costs of mining or trading this mineral. Alternately, you may consider this addition to your model excessive.

What Are the Main Challenges of This Use Case?

There are many challenges to building a good trend forecasting model. First, there can never be any certainty about the results and you may have to wait months or years to see if your forecast was helpful. Secondly, it can be difficult to determine which data to include in your model and how to account for failures to meet KPIs—is a drop in expected sales, for example, a temporary situation or it is indicative of a major shift in market expectations?

Finally, major unexpected world events (like a global pandemic) can destroy your entire forecast and signal permanent market changes that you could not have accounted for previously yet now must consider in all aspects of your business.

Interesting Case Studies and Blogs to Look Into

Heuritech: How to Predict the Success of New Accessories Right after the Fashion Show?
Jena Nesbitt: Creative Direction Case Study

Tangible Examples of Impact

There are three basic types—qualitative techniques, time series analysis and projection, and causal models… These differences imply (quite correctly) that the same type of forecasting technique is not appropriate to forecast sales, say, at all stages of the life cycle of a product—for example, a technique that relies on historical data would not be useful in forecasting the future of a totally new product that has no history.

Harvard Business Review: How to Choose the Right Forecasting Technique