How can we manage the ever growing number of data inputs for continuous intelligence?
Sort by:
The key is to identify pertinent data. For example, Quest Diagnostics has a portfolio of about 30,000 different tests they offer.
So one would think that if a physician did all 30,000 tests, they would know everything about the patient.
That could not be further from the truth.
Same thing with data inputs. It is so easy to get overwhelmed and obtain data that offers little value. Identify the specific good and valuable data first. Once you do that, you will find you are likely not overwhelmed. Most firms only get overwhelmed when they don’t know why they have the data input in the first place.
What I'm hearing from a lot of people in manufacturing is how do we take the one-to-many relationship of the single piece of data into that kind of multi-tenancy model and apply the different perspectives. That is their challenge. And they're trying to do it in the broadest way possible, whether it's through manufacturing execution systems or product life cycle management, both of which are cloud-based. We're not talking about 20-year-old SAP sitting in the background, it's not that way. It's much more of “how do I do this so that the value that I'm driving for the organization is insight to the customer, the sales guys, the end-user consumer who buys the electronic device. The car, for example, because now everybody's making autonomous EV. There's a tremendous amount of security that's required with it, and how do I qualify or quantify that?
Well, it's intimately related to identity and role and access.
Yeah, I would agree. I think I've used the same description you just provided relative to people interest and applied it to data collected for a city. So when you think about a city, there are interests relative to infrastructure utilization, relative to tourists, relative to transportation, relative to available parking relative to any number of things, whether the season is right for stores to start opening an hour earlier based on traffic, you name it. And the cool thing about that particular model from my perspective is that collecting the data is the easy part and it's also in the long run, the cheapest part. You don't have to sell value from data for very much to justify. I still think somebody is going to make billions of dollars eventually if they can figure it out.
The way I encounter customers is that for every stream of data, there is a hero, there's somebody whose life we have to make easier or better or more productive in their job function. And so an organization is the sum of all the heroes. I think of these almost as layers in a map, if you know GIS, then you could layer different views on the underlying substrate. So you can have somebody who's responsible for a layer and a hero, but not lose the ability to go top-down and see everything. That's what we're trying to build at Swim. And we have one implementation thus far in Dubai. We do all the traffic, so we have every Uber, every taxi, every public transit vehicle in marine and trains and trucks, each of which is owned by a different person and they'll never share data. The fiefdoms are all about how much data you have, but the overall view or the overall intelligence, say around an accident or something, is derived from a top-down view of everything altogether. The goal has to be to make everybody a hero in their world, let them own insights rather than just raw data and then find some way in which organizations can combine different layers and views.
I think what you're working on with Swim is brilliant. Do you see Swim as basically being a part of what you just described where a lot of the layers of data that you throw away are layers that other parts of a city or a transportation authority could probably still benefit from?
There is an interface in the space use case that lets you, depending on who you are (the space agency, the U.S. Space Force, the British MID, Tesla,etc.), dig into different views of all your satellites and space junk around the world. The overall requirement is to find things that are going to collide, so the overall goal is different from the individual tenant goals, which might be for tests of one of my satellites or whatever it is. But the overall goal is clearly cross-plane, it has to cross all these planes. And so we're starting to play with that, it's fun. It's a journey of learning and we are nearly there, appreciated.
It’s about sifting through the ever-expanding mountain of data and reading out meaningful content.
More is NOT better. More is just more.
One of my preferred techniques is combining multiple pieces of data into single indexed results — i.e. making more into less and building digestible meaning.