Is it time to dump survey-based insights?

The surprising U.S. presidential election results raises questions on survey or poll-based insights and predictions.

 In the restaurant industry, we too typically rely on surveys and research to determine our business and marketing strategies. Based on a grand sum of 1000 consumers we change our menus, prices, creative, concept design. Generally speaking, a larger sample helps control for issues of honesty, but a larger sample with the same bias will just give more statistical confidence in the answers, many times, which are wrong. 

 Fortunately in restaurant marketing, when we are wrong we have an opportunity to course correct by using actual purchase data to reconcile what people say they do with what they actually do. We may have 20/20 hindsight into the polling models post election, which shows us the sample bias that led to incorrect projections - but it's too little and too late to react to. For marketers, using a variety of research and analysis methods including reconciling expected results with actual purchase behavior we can help create a learning process. In high frequency marketing campaign deployment, 'errors' are an opportunity to learn more about who customers are and what they respond to. A properly constructed data and analytics process is built as a perpetual cycle that includes adding new data sources and constantly learning and recalibrating from results. 

 Data is powerful. For companies relying solely on survey or self-reported customer response data, it’s time to expand your data sets and your understanding of your customers. By connecting their transactional behavior with their lifestyle, psychographic and engagement data, a true picture is possible. Through these insights you can create the right business and marketing strategies and win their share of wallet and share of mind by delivering contextual messages in the right channel at the right time to the right customer. This is where the rubber hits the road and a disciplined approach to segmentation, deployment, measurement and refinement must be established and maintained.