In the early 1960s, chaos theory pioneer Edward Lorenz famously asked, “Does the flap of a butterfly’s wings in Brazil set off a tornado in Texas?” Lorenz theorized that small initial differences in an atmospheric system could result in large and unexpected future impacts.
Similar “butterfly effects” can surface in the increasingly interconnected and complex universe of enterprise partnerships and supply-chain and cross-product relationships. It’s a world where new or evolving products, services, partnerships, and changes in demand can have unexpected and surprising effects on users and other products, services, traffic, and transactions in a company’s ecosystem.
Monitoring these complex relationships and the potentially important changes that can reverberate through an enterprise’s network calls for an interconnected system of virtual “sensors,” which can be configured and tuned to help make sense of these unexpected changes. As enterprises increasingly interface with customers, partners, and employees via apps and application programming interfaces (APIs), setting up a monitoring network like this becomes a particularly important part of data analysis.
What are Sensors?
Traditional sensors are often defined as “converters” that transform a physically measured quantity into a signal that an observer can understand. Sensors are defined by their sensitivity and by their ability to have a minimal effect on what they measure.
Physical sensors can capture aspects of the external environment like light, motion, temperature, and moisture. They’re widely used in business, too. Retailers can employ them to measure foot traffic outside or inside their stores, in front of vending machines, or around product or brand categories. Airlines use physical sensors to measure how weather patterns affect boarding and take-off delays. Using a diversity of sensors enables the definition of an environment around the usage of a product or service in the physical world.
Besides investing in traditional data processing technologies, cutting-edge enterprises map their digital world by defining and building so-called virtual sensors. Virtual sensors collect information from the intersection of the physical and digital worlds to generate and measure events that define the usage of a digital product or service. A virtual sensor could be a data processing algorithm that can be tuned and configured to generate results that are relevant for the enterprise. The generated alert notifies the enterprise of a change in the environment or ecosystem in which the user is using a product or service.
How to Build a Virtual Sensor Network
Building a network of virtual sensors for your business calls for requirements similar to those of a physical sensor system:
- Sensitivity, or the ability to detect events and signals with configurable thresholds of severity
- Speed, or the ability to speedily collect and process signals to generate business-critical events
- Diversity, or the ability to collect, collate, and combine signals from multiple sensors with the goal of generating business-critical events
To begin charting the web of relationships that impacts the demand and usage of various enterprises’ products and services, businesses should determine which other products and services in the marketplace are complements, supplements, and substitutes to their own. Deep understanding of such evolving and complex relationships can help enterprises with planning partnerships.
- Supplementary products and services enhance the experience of another product or service. For example, flat panel TVs are enhanced by wall mounts, stands, warranty services, cable services, and streaming movie services.
- Complementary products and services work in concert with other products and services to complete the experience for the end user. Demand for car tires, for example, tends to generate demand for gasoline.
- Substitute products and services have an inverse effect on each other’s demand. For example, two retailers offering the same selection of products targeted to the same consumer.
Understanding these relationships is the starting point of creating a network of sensors to monitor the impact of changes in traffic or transactions of an outside product or service on an enterprise’s own products and services. Detecting this change within the appropriate sensitivity can often be the difference between an enterprise’s failure or success.
Take for example, a web portal that aggregates content from several content providers. This portal uses APIs to connect to these third-parties. In many cases, these content providers are automatically queried by the aggregator, regardless of whether an end user is interested in the content. If for any reason there is a spike in usage of the portal on a particular day, this will automatically trigger spikes in the traffic for each of the content providers. Without understanding the complementary connection to the portal and the associated shifting demand properties of the connection, the content providers will find it difficult to interpret the traffic spike, which will eat up resources and leave legitimate traffic unserviced.
Here’s a similar example. Let’s say a service can support 100 API calls spread among 10 partners. If this service receives an unexpected and unwanted spike in traffic from one partner that eats up half of its capacity, then it will only have 50 API calls left to distribute among the other nine partners. This in turn can lead to lost transactions and dissatisfied users.
With an awareness of the network, however, the service would understand that this one partner routinely only sends 10 calls on a normal day, and would be able to put restrictions in place that wouldn’t let the extra 40 calls eat up the capacity of other partners.
In these kinds of situations, virtual sensors can provide the awareness and insights into this web of interdependency, and help make sense of traffic spikes that otherwise might seem incomprehensible.
Sensor-Aware Data Investments
Building a network of physical and virtual sensors entails collecting diverse signals from a complex map of data sources and processing them to generate events that can help enterprises understand the environments around their end users. Investing in these networks enables enterprises to track and monitor external signals generated from sources that have the ability to impact the enterprise’s traffic, transactions, and overall health.
This ability, in turn, helps digitally aware businesses negate potential troubles caused by the digital butterfly effect, and take advantage of the opportunities presented by a strong grasp of what’s happening in user and partner ecosystems.