Announcing interactive web-based decision intelligence
At Quantellia, we’ve been delivering enterprise-scale, desktop- and PC-based decision intelligence models to our clients for a few years now, using our World Modeler™ software. In the last few months, every single one of our clients has asked for our work to be delivered through a web interface, so we’ve been heads-down in delivery and development to meet their needs. These are not available to be viewed by the general public, however, so I’ve spent the last few days building a demonstration to show you what we do, and as part of my answer to a recent Quora question on Agency theory as well as Mark’s upcoming talk on this topic at MLConf Seattle.
The model below (along with decision intelligence as a whole) mixes ingredients from machine learning, big data, intelligence augmentation, predictive analytics, dashboards, complex systems analysis, causal modeling, agency theory, statistical analysis, and more in a way that answers the most practical of all questions: If I make this decision today, how will it impact my future business objectives? Going beyond “data interaction” and “data visualization”, this model generates data from the future, and avoids the data visualization brain upload problem.
My model’s below, and I encourage you to take it for a spin. I built this based on one I built in World Modeler for a recent client.
To set the stage, imagine you are trying to decide how much to invest in a new training program for your company. The primary motivation: projects are systematically missing their schedule dates, and you feel that the problem could be solved, in part, by better training. But you know that training is more helpful with a more junior team, and you don’t know how improvements to skills translate into the bottom line. This model breaks it down. Enjoy!
(Note you’ll need a modern browser; if this doesn’t display, please consider upgrading to the latest version of Chrome, Firefox, Opera, or IE.
- You might like to zoom in a bit to make this easier to use. Ctrl-(number keypad +) does it on my computer. You can also see this on a page on its own at this link.
- Move the Training Investment lever. Observe that this investment makes zero difference if teams are already reasonably well-skilled.
- Lower the skill level by moving the Average skill level slider. Observe that this results in some value, but not enough to outweigh the cost of training.
- Lower the Cost per hour of training and raise the Training investment further. Now we’re getting somewhere.
- Note that sliders represent both decision modeling Levers as well as Externals, depending on your circumstances. You might have control over the cost per hour of training, or you might not.
- See if you can find a combination of slider settings that produce a benefit that outweighs the initial investment
- Disagree with my curves that show how skills translate to training benefits? Drag the blue dots in the curves on the right-hand side up and down to change the underlying model behavior. See this by adjusting the sliders again.
- Note that none of what we’re doing here is “drilling down” into the data, nor data exploration. Instead, we’re using pre-existing machine learning, hand-drawn cause-and-effect curves, and more to answer the question: “If I make this decision today, how will it impact my business objectives tomorrow, given the best data and information that I have available?”
- Please see more decision model examples under the Resources link in the menu at the top of this page.
Would a model like this be useful to you? Please drop me a line, and let’s talk about how I and my team can help you out.