You know those seven billion connected devices in the marketplace? They’re generating enormous amounts of data. And the consumers behind all of these phones, tablets, and smart watches want to be engaged with that data. Unfortunately, most companies lack the skill sets internally to figure out the necessary technology.
“The only people building this stuff today are basically ‘rocket scientists’ at tech companies,” says Adam Bosworth, a former Google vice president of engineering, whose résumé also includes pioneering Microsoft’s XML program. “They’re very smart. But they’re not normative and they’re expensive.”
Now salesforce.com’s Chief Strategic Officer, Bosworth and his current team are working on a next generation engagement solution. “What I’m focused on is how to take all of this complicated technology and make it easy for our customers to connect to their customers in a whole new way.”
Why real-time needs data science and scalability
As it goes, the window of opportunity for companies responding to consumers is very small, whether via a web or mobile application. The sooner the information gets to them, the better. Real-time event processing recognizes data is coming in and makes it possible to react in a second or less.
Bosworth compares this to experiencing a flight delay. Passengers who figure it out the fastest are most likely to get rebooked effectively relative to others. But this is just one piece of the puzzle. You also need data science around it. If a fitness tracker repeatedly tells a user he weighs 175 pounds, it’s not remarkable. If it tells him he will weigh 185 pounds in two months if he doesn’t change his behavior, that’s intelligence.
“You’re basically creating a model that looks to see if either the number of things occurring, the trends of things, or the aggregate value of things in some time window mean some value. And the moment they do, it lets somebody know or triggers automatic action,” Bosworth says.
Along with reacting intelligently, the system should respond to messages at a constant speed regardless of how much data is coming in at any moment in time. For companies with a large number of customers, this could mean 100 million, to even a couple of billion messages a day. “That kind of dynamic scale, called elasticity in the cloud, is actually very hard to do,” says Bosworth.
Building programmer logic and monitoring data
Another tricky piece to this next wave of real-time customer engagement involves building a productive programmer model. Its logic will need to make it easy for users to specify how the system can quickly react to incoming events, as opposed to waiting to read the data and then deciding what to do about it.
Instead of a car alerting the driver that it is low on oil, the programmer can give it parameters, based on trends, that disclose when the car’s oil is leaking and on its way to needing to be filled. “Something happened; let me describe what I want to do when something happens, but let me take into account what has been happening all along,” says Bosworth.
A second example: If a customer signals by requesting information that she may want to buy a product, the rules need to dictate what content is pushed to her without making her feel spammed or nagged. According to Bosworth, salesforce.com has already laid the groundwork with its newly released Journey Builder. Its next development step will involve allowing for additional customization.
Once this new real-time engagement solution is in play, the final step is a monitoring element. Companies will need to know at what point engagement was triggered, what sparked it, and how the responsible characteristics can continually be improved.
“Salesforce has pioneered every cloud innovation from social to mobile. This next evolution of real-time data processing will be another game-changer. We have all the pieces in our hands to give our customers this recipe for engagement,” Bosworth says.
This article was written by salesforce.com and not by the Quartz editorial staff.