IoT and VCR, now onto C

In my (last article) I talked about VCR where V stood for Visibility. In this follow up I want to talk about C, where C stands for Causality.

The purpose of these articles is to provide a business-centric understanding of IoT, not to delve into the technicalities (of which there are many!). Despite the flurry of content available I have found many business people are unclear on why to entertain the IoT conversation (other than that everyone is talking about it).

To be frank, often the business value is unclear to start with, whether due to immature technology, a technology rather than business (case) focused approach or lack of understanding around change management and technology adoption. It is not like selling traditional SAP or Oracle ERP or desktop support services or data centre or even cloud migration services. It is new, exciting, dangerous and risky. I’ve spent 75% of my career working for start up/SME’s, this is what we do.

If Visibility is the “what” (happened) then Causality is the “why” (did it happen) and/or when (might it) happen again. Visibility on its own is useful (in fact vital) and on the IoT journey it can be a useful start, a phase 1 drop perhaps. However, knowing that something has happened is only the start in terms of deriving the maximum value from the event.

Put simply Causality is “cause and effect”, why does something happen, what factor(s) affect my operation? Some may be within direct control, many will not but, at least if we can model and understand the major factors, it gives us a chance to identify when they are likely to occur and hence mitigate their impact.

Now, causality is a complex topic. Some would say the whole world is a massively complex cause/effect network that no one will fully understand because there are too many “moving parts”. I can buy into that.

We’re all familiar with the story of the butterfly wing flapping in one part of the world causing a hurricane in another, but for most businesses the cause(s)and effect(s) are more obvious and more identifiable. Once you have a view on the cause you have the ability to introduce “some” control.

At its heart then IoT causality is applying meaning to the data generated from your sensory network, applied to your specific business. Knowing the temperature of your lorry is 5 degrees Celsius has a very different meaning if you’re carrying ice cream or if you’re carrying fresh fish. But it is not just data from your sensory network, this on its own is limited in effect at best. It also needs to include additional data sources, if you are to really understand the major factors affecting your operations.

For example, if you’re running a logistics business what effect does time of year, time of day, traffic conditions, age of driver, driver health, driver alertness, type of goods, weather conditions, type of vehicle, condition of vehicle and route have on your ability to deliver. It is possible that certain combinations of the above may affect your delivery schedule and that some increase the risk of an accident, potentially fatally.

If you’re a retailer then how can you understand your inventory location, volumes, stock outs, misplaced stock, shrinkage, staff movements, customer movements and customer demographics within your store, linked to sales by item and overall store profitability? It is important to note that some of this needs to be done in real time, so you’re able to effect a response to improve sales and customer service. There is no retailer that can do this now but none of them would deny some cause/effect across these elements. After all you can’t sell what you don’t have on your shelves.

Many manufacturers have limited real-time visibility of their production process and hence limited understanding of their Work In Progress. They are unable to see how production flows through their factory (the visibility bit), effect of the workforce on productivity, the external factors affecting output. For example, the product moved to “the side” for further work which then gets lost.

Agriculture is ripe for the adoption of sensor-based farming to understand how to improve yield based on weather, soil consistency, seed type, animal fertility patterns and many other factors.

This leads us to a major and impending issue. Namely amount of data and the ability for people to make sense of it. Yes, event these clever and somewhat expensive Data Scientists. Put simply the expected amount of data generated from IoT devices AND the influencing factors will overwhelm the ability of people to digest and understand, model it and keep the model up to date. Luckily there is a solution (or should I say area of capability) that can help.

Data Scientists are in short supply at the moment but just as once people used to weave cloth by hand, the time is now here where a single person can’t ingest and understand the volume of data, hence we look to the newly emerging Machine Learning/Artificial Intelligence.

And to be clear, Machine Learning is not magic. It doesn’t look at a data set and “suddenly” infer that your factory is/isn’t operating at maximum efficiency. It still needs human inputs to determine what is right and what is wrong and the shades of grey in between. It needs time to get enough data to be able to spot trends and understand patterns. It also takes a lot of compute power. It is not a magic bullet but it offers a route to deriving meaning in the high volume, data-led world that we are building.

Finally I wanted to share one my critical learnings on IoT. Do not expect your partner or supplier to turn up with a turnkey solution that has been proven 50 times with a clear business case and value proposition. Where we have been successful (read, the customer has seen tangible benefit) is where it has been a co-creation, collaborative exercise. This is not necessarily a natural process for many System Integrator nor for many customers but approaches such as Design Led Thinking and a consultative, business benefit realisation exercises can be effective.

IoT is something of a Brave New World, full of contradictions. It is risky, it is rewarding, it is uncertain, it is valuable, it can easily be a waste of time but it is coming and the time to get on-board is now.

My next article will address R.  As usual, comments and insight gratefully received.