Still today I watch out the aircraft window and try to make sense of the patterns that I see from miles up in the air.
What’s most interesting about patterns is that some emerge only at a distance, like the watershed of a mountain range. Others only appear upon close inspection, even magnification, like the unique patterns of crystals that make up snow flakes.
Patterns are everywhere. In business, recognizing patterns is how we sell more efficiently, spot issues with products, and manage our employees better. What we’ve done to date, however, pales by comparison with what is emerging in the science of patterns.
The acceleration of the creation, storage and processing of data creates an ideal opportunity to apply mathematics to the discovery of patterns. This is just in time, too. Before the world was so interconnected, an enterprise had control over the data that mattered most to the business. The technologies that grew up in this environment didn’t need to take into account data outside the organization’s control. Enter the Internet and application proliferation.
We no longer have that luxury. Not only is there more data, but there are more sources that at a glance may appear to be unique but on closer inspection are really the same thing represented slightly differently. The science of pattern recognition is key to creating trusted records from the many, disparate sources of information available inside and outside the firewall.
When data can be mathematically analyzed for patterns, key information emerges, like:
- Patterns of fraud in the submission of workers compensation claims
- Comparative/competitive pricing of products that are profoundly similar but sold under slightly different ID’s
- Security watch lists that distill various spellings of names to produce a single record that can be easily monitored
- Master patient indexes that allow for a single patient record despite data coming from multiple doctors, laboratories, pharmacies, and other treatment providers
- Event detection when the ‘ground rules’ aren’t already known, like emerging threats from epidemics or credit card fraud
The beauty of patterns is detection of information without knowing exactly what to look for. We’ve always built systems in ‘fragile’ ways that relied on very exact information expectations. Not having that concern is very, very liberating.
Without pattern capabilities, our best efforts are to find and fix these problems after the fact, when the patient has already been prescribed duplicate medicate, the fraudulent claim has been paid, or the dangerous individual is past security. Detecting patterns is the way to insulate ourselves from the ‘duplicity’ of a fast-moving, ever-changing world.
Patterns that we see or find enable us to solve problems without starting from scratch. Systems are now doing what only people could do before, and only when it wasn’t overly complicated. We can get systems to understand ‘close enough’ and take action.
Patterns are a fascinating way to make systems more powerful without making them more fragile.