Big Data 101: Quality, context, relevance, and actionability are key

Manufacturers and other businesses often assume that big data is, first of all, lots of it. That’s really a small part of a big data strategy. It’s much more important to generate the right information, ask the right questions, present the outcomes in a meaningful way, and make data actionable. It needs to support your company’s strategy and decisions. Simplifying your approach and questioning the tacit assumptions behind your effort often helps, especially when you want to give your big data strategy the agility that mobility and social can facilitate.

When we talk about big data, the idea of ever-growing data masses is usually present in our minds. As businesses become digital enterprises, they create, harvest, and consume more data than ever. It takes a wealth of data to create digital customer experiences and provide effective systems of engagement—you may need, for example, data related to past purchases and transactions, preferences, locations, products, the user context and environment, and more. Data sources include the ERP system as well as other business applications, but also the devices and equipment in the internet of things, social media, and company-built environments.

For manufacturers, the data ecosystem may include performance metrics, environmental records, and location details of equipment and machinery, as well as engineering records, bills of material, service and maintenance orders, and social conversations. A service engineer, for instance, may access that information on a mobile device, add findings and recommendations, and initiate an engineering change or a modification before a failure occurs.

Big data can translate into innovation

Our customers often come up with big data scenarios that push the envelope of the possible. Lotus F1 Team generates a lot of data of high-quality data from the hundreds of sensors placed on its race cars which is processed at about 40 teraflops, generating yet more data that directly drives innovation. It uses that information in revised designs for better-performing parts in faster cars, and then produces them in the two-week sprints between one race and another. In addition, the organization gathers data to help race car drivers perform at their best level and make sure they enjoy optimal physical and mental health. Human bodies react strongly to the extreme acceleration and the challenging cockpit environment of race cars. Drivers can lose up to 18 pounds in a race, and their heart rate and other values are being monitored very closely. Finally, another set of data streams document the social interactions and feedback of viewers during races, helping the marketing team expand the Lotus F1 brand through events, sponsorships, and media coverage.

That kind of big data requires substantial resources. At Lotus F1 Team, close to 500 people consume the post-race data records and translate them into race car parts and performance. Vast data storage is also necessary, although its cost is relatively modest.

Proper perspective and the right questions

Reliable, relevant information is a starting point for big data. To make that data actionable and meaningful, you need to ask the right questions from the right perspective. From my own experience, I believe we often look at business data from much too close, without the right vantage point. It’s easy for the huge quantities of information to become distracting. Even with its masses of information, Lotus F1 Team generates exactly the data it needs, not more. The question to be anwered is “How do we make the cars faster?” Answering it requires analyzing and synthesizing all the data points involved, considering all the designs and components that are part of a race car.

Manufacturers need to strategize how to create value from big data, without drowning in masses of information—which can pose a security risk, and can needlessly consume storage and management resources—or asking the wrong questions. In a well-known market research anecdote, apple pie appeared to be consumers’ most favorite pie, but it turned out that only applied to large-size pies and was the result of household consensus. Retail managers invested in small-size apple pies and saw large numbers of them go to waste, because individual preferences were not the same. In this case, the data was solid, but it was not the sort of information that was really useful, because different conditions applied to the marketing of family-size vs. individual pies. The question asked was not relevant to the marketing of individual pies since it disregarded a key detail of the research.

Tayloring the approach so big data becomes actionable

Overcomplicating the results of data analysis is an unproductive habit that can trip us up even if we have the right data, ask the right questions, and have meaningful findings to present. Sometimes, we create complex graphics and charts when a simple list and a couple of recommendations might be more meaningful. You need to share your findings in such a way that they are actionable for the people who need to make decisions that lead to better products, more competitive strategies, or more compelling customer experiences.

The takeaway

Thus, to sum up big data 101:

- Your data has to be sound and reliable.
- Quality is better than sheer quantity.
- Don’t create more data than you can process...
- ...and don’t process and present more data than people can consume and use.
- Data needs to be actionable and meaningful for your business, so it can drive action.

The right data analysis and presentation tools also make a difference, especially when you create systems of engagement that involve mobility and social within a complex manufacturing ERP environment. At STAEDEAN, we’re developing the apps and solutions that give you a powerful, scalable foundation for achieving this.

I would love to hear about your experience in creating a big data strategy, so please get in touch with me or contact STAEDEAN.