Social Forecasting is the aggregation of employee and/or expert knowledge and its conversion into quantitative business KPIs and forecasts. It is used for forecasting new product potential, sales figures, as well as strategic scenarios. This flavor of crowdsourcing with employees is being used by various corporations, which have embraced a simple insight: “Employees often know more about products, markets and competitors than you think”.

111208_CMSwire - Social Forecasting.jpg

In the late 1990s, the first adopters of Social Forecasting applied it in business to forecast sales, project deadlines and even the probability of a competitor entering the market. Nowadays, in a 2010 survey of 3,249 executives by McKinsey Social Forecasting, Prediction Markets has been named one of the Top 10 Web 2.0 tools in modern enterprises.

How does Social Forecasting Work?

Social Forecasting is based on a simple concept, which combines the approach of crowdsourcing with an incentive mechanism for users to contribute their best knowledge. Crowdsourcing participants are usually consumers or lead users. In Social Forecasting, a company utilizes its employees' knowledge of the market, the consumers, the competitors, etc. to answer complex strategic questions and generate quantitative forecast about sales, new products, new product ideas, etc.

Byron_2.jpeg
 

Employees visit a Social Forecasting web application based within or outside the company. The single user can access specific topics and then pick a certain product to make a sales forecast about (by region and SKU, for example). The twist is that the user also has to place a stake of virtual play-money on every forecast made. The more confident the user feels about the prediction, the more play-money can be staked. At the end, if the user’s forecast turns out accurate, s/he will turn a profit. Otherwise, the stake is lost. This creates a significant incentive for users to forecast as accurately as they can.

The crowd forecast is then the aggregated result of all the individual forecasts. Each user input updates this crowd forecast, making it a live indicator of how external events impact the forecast. At any point in time the current crowd forecast can be downloaded and used to aid the business process further down the line, e.g. in the supply chain or as a support for management in strategic decisions.

From the employee point of view, Social Forecasting is a way to contribute to important decisions by providing one’s own insights. Based on one’s performance, the employees are also rewarded with prizes or other incentives. For management, Social Forecasting is a unique way to consolidate hundreds or thousands of employee opinions into a single number -- the crowd forecast -- making this a more actionable piece of information than endless amounts of corporate wiki articles and blog posts.

As an example, the following table compares Social Forecasting with predictions derived from election polls. As can be seen, the social forecast was more accurate than the corresponding survey 78% of the time.

Ivanov_table.jpg

Social Forecasting in Business

Numerous companies have begun to use Social Forecasting to predict business-relevant KPIs. The following chart provides a few examples across industries, including high-tech (GE), Consumer Goods (Henkel), Agribusiness (Syngenta), telecoms (Deutsche Telekom) and retail (BestBuy).

Ivanov_new3.jpeg

General Electric uses Social Forecasting to assess new product and technology ideas in different business units. The high-quality of ideas resulted in 2-6 patents per business unit.

Germany’s Deutsche Telekom -- the 4th largest Telco firm worldwide -- uses Social Forecasting to gather the knowledge of over 240,000 employees to receive quantitative estimates and forecasts on demand forecasting, marketing, HR and strategic issues.

Consumer goods giant Henkel has achieved a 22% increase in the accuracy of their S&OP process. This resulted in 8-digit positive EBIT effects annually in North America alone. As a result, Henkel is introducing Social Forecasting also to its core markets Germany, France, Spain and Italy.

In retail, BestBuy has been an early pioneer in Social Forecasting. Nowadays BestBuy applies Social Forecasting throughout the company and uses it for demand planning for all top products, as well as forecasting new products sales, product launch dates, store opening dates and large special events, e.g. Christmas sales. In these application areas the Social Forecast consistently beats the internal analysts.

These are only a few examples of Social Forecasting in business.

Deploying Social Forecasting

Byron5.jpeg

The important thing in Social Forecasting, just as in wikis, blogs and surveys, is having relevant topics. They must be relevant for the managers, who are the end users of the generated results, but also for the participants. Topics such as new products, strategy, sales and marketing are good examples.

After one has chosen the topics, the next step is selecting the right participants – the wise crowd. Topics, participants and activity are closely related. Of course, the more the better, but if you are not sure how many people you need for your Social Forecasting, take this as a rule of thumb: about 150 participants for 30 forecasts per week, e.g. sales of 100 products each, with 3 different flavors.

The final step in the setup is choosing the right platform and good incentives. The whole setup mustn’t be spoiled by being too complex and without any means to encourage the participants to use it. The influence of such incentives is huge, as a study by LaComb has shown. The results provided by incentives-driven Social Forecasting were better than those without.

Last but not least: Social Forecasting is not a perpetuum mobile -- it can’t be turned on and left without any supervision. It is a social phenomenon that has to be constantly maintained and updated, much like a corporate social business platform, which Social Forecasting tools are often integrated with. Newsletters, regular blog entries and frequent introduction of new topics ensure that the participants won’t lose interest in using a Social Forecasting platform.

Conclusion

To summarize: Social Forecasting takes in many opinions, similar to crowdsourcing, but combines it with an incentive mechanism for people to give their best possible estimates. This increases the quality of the inputs and thus results in highly accurate forecasts. The resulting crowd forecasts have proven to be more accurate than traditional surveys and statistical models, especially when forecasting new products or demand in highly-volatile markets.

Editor's Note: You may also be interested in reading: