Enterprises using Social Media To Predict Demand: PepsiCo

Social Media Predictive Data

Consumer demand today changes faster than ever before. As a result, new product launches are failing at extremely high rates. After all, accurately assessing the market and predicting demand requires terabytes of data, which ranges from point-of-sale numbers, e-commerce transactions, out-of-stock histories, pricing averages, promotional planning, special events, weather patterns, and a many other factors. To add to that, most enterprises continue to ignore the significance of applying online consumer dialogue to predict future purchasing behavior.

Just a little more than a decade ago, consumer opinion could only be gathered through relatively small sample-sized market research groups. Today, social media and customer review platforms let Enterprises tap into the minds of millions of millions of potential customers and refine their product launch accordingly. Product launch failures occur all the time. But what exactly is the problem?

Product Launch Fails Happen To Everyone

No business is safe from misunderstood consumer demand. Even the biggest and oldest enterprises can wrongly assess the market and launch products that quickly end in failure.

Google Glass, an innovative head-worn device with smartphone capabilities, promised to revolutionize consumer electronics. But soon after its launch in 2012, it became clear that there was not enough consumer demand for the product. Consumers weren’t actively searching for the features that the device contained; in other words, the technology was launched far too early for product demand to mature. Not long after launching, Google discontinued the product, deeming it a failure. 

Staggering product launch failures happen in virtually every industry; let’s take Burger King as an example. In 2013, the fast food chain launched Satisfries, a healthy alternative to french fries, as an attempt to accommodate consumer preferences’ shift toward health-conscious foods. Unfortunately for Burger King, they assumed their customers would prioritize health over other factors, such as flavor and price. It only took months for the product to be discontinued. While the Satisfries product was likely applauded by its research panel, in reality, customers were not planning to go to Burger King with the intent of eating the healthy alternative; the company misaligned interest with intent.  Failure to understand how consumer interest correlates with future sales led to the failure of the Satisfries launch.   

Customers are constantly voicing their interests on the Internet. If more companies drove their launch decisions using alternative data sources, perhaps the storylines would be different.

ecommerce purchase transaction

Bringing Data To The Rescue 

Recent advances in AI technologies, such as Natural Language Processing and Predictive Modeling, substantially reduce the risk of product launch failure. By collecting and analyzing as many data points as possible, companies can paint a better picture of customer desires, learn more about the market’s needs, and increase the probability of product launch success. Demand forecasting is a complex process, which requires constant variable adjustment, consideration of the specific features of business, and frequent refinement of demand forecasting strategies. 

One of the tactics to stay ahead of consumer needs is “demand sensing” – the use of new data sources to predict purchasing patterns. This innovative approach collects and correlates a company’s historical data with other real-time signals, which makes it capable of reducing forecasting errors by 45%. Digitalization paves the way to better understanding today’s consumer whose needs constantly change. Even as customer preferences evolve at great speed, algorithms can predict these changes months in advance and allow enterprises to react before their competitors.

Demand sensing allows retail and manufacturers to achieve higher customer satisfaction and launch efficiency. For Consumer Goods enterprises, demand sensing is vital for anticipating customer needs. Early adopters Procter & Gamble and German retailer Otto already report improved cohesion between the massive data volumes that they collect and how those translate into real-time decision-making. But even low volume businesses, such as aircraft engine manufacturers, have found demand sensing useful for assessing the need for spare parts and repair services. 

retail purchase data

Using Social Media To Predict Demand: The Case of PepsiCo

To predict consumer demand with a greater degree of accuracy, businesses can use social media and customer reviews to supplement their historical data. PepsiCo is one of the companies that embrace this approach. The company exists in a highly competitive environment and needs to find growth points within its market. Notably, the company’s primary goal is to meet rapidly shifting expectations and shopping habits; and they use demand sensing to help them achieve it. Lately, PepsiCo made bold moves in product development and marketing, trying to accommodate consumer demand for more healthy beverage options – a growing trend that was discovered through social media analysis. 

Supplementing econometric demand forecasting with unconventional data sources can answer a variety of questions that surround product innovation – specifically, which products customers want, what tastes they crave, and what colors are growing in popularity, among others. PepsiCo’s approach, already proven successful for product development, is further applied to sales and marketing decisions, including ad spend and messaging decisions.

Enterprises either choose to build in-house systems for predicting demand or outsource the task to solutions like Trendscope or Simporter. These tools aggregate consumer sentiment and produce actionable insights about which products to bring to the market. The technologies can discover the desirable attributes of a product, the marketing messages that motivate purchasing behavior, a competitor’s impact on market share, sales forecasts of the product, and much more. Algorithms can predict what customers will say about a product (and where they will talk) by creating Neural Networks that correlate historical online chatter with historical sales data. Once determined, demand sensing systems are able to translate the dialogue into actionable strategies.

The Future Of Demand Forecasting 

Further research proves that using social media to predict customer demand is instrumental in making better sales predictions and, in general, increasing the efficiency of operational decision-making. Although still in the early stages of adoption, including online data in demand forecasting is capable of increasing the accuracy of prediction models to as high as 98 percent.

Demand sensing and algorithmic forecasting are effective across multiple industries. Retail, automotive, fashion, food and beverage, consumer goods, and pharmaceutical can all benefit from a new approach to demand sensing. Altogether, any Consumer Goods niche with textual customer data can enhance their product launches with additional data.

In the end, an upgraded approach to demand sensing pushes R&D, supply chain, purchasing, and marketing planning to new levels of accuracy, systematically increasing the odds of a product launch success. With growing interest in the market, more companies are likely to follow suit of PepsiCo and look for solutions that offer demand prediction and pre-launch analytics.

Conclusions

Aligning supply to true demand enables companies to launch preemptive responses to the market. Thanks to data mining and advances in AI, it’s now possible to verifiably predict demand and avoid damaging pitfalls when launching a new product. Online data gives unprecedented access to customer preferences, which, if used effectively, can determine the fate of a product launch months before the actual launch. 

What do you think?

This site uses Akismet to reduce spam. Learn how your comment data is processed.