First impressions

Connected devices experience many challenges in their first few hours and days of life, as they first wake up and connect to the cloud. And this is exactly the time when the customer receives their critical "first impression" of your service.

Challenges in these early times include:

• Securely connecting the device to the internet and cloud service may require user intervention, to provide e.g. network and service access credentials
• If the device has been in storage for a while, it is likely to need a software upgrade, which can be a delicate process
• Problems with physical installation, e.g. poor radio connection
• Hardware failure rates are highest because of the so-called "bathtub curve"

It is often quicker, easier and cheaper to fix these early problems quickly, perhaps even while the installer and the customer are on-site and paying attention to this newborn device. Even if there are some teething problems, customers are much more prepared to give a new device a second chance while it is still new and exciting. Indeed, problems are an opportunity for the service provider to demonstrate just how effectively they deliver support - which paradoxically can lead to even higher customer ratings than if there is no problem at all.  

Therefore wise Operations and Customer Support teams will set-up metrics and processes to acutely monitor and manage devices during these sensitive early days, to ensure a good first impression and to minimise support costs.

Identifying "Kindergarten" devices

We can use the power of DevicePilot to analyse performance and automate problem-resolution for early devices, just as at any other stage. But how to tell DevicePilot to apply measurement or automation only to devices which have just been born? The answer is to apply a device filter in the "Scope" section of the DevicePilot KPI, which selects for analysis only those devices which are in - say - their first 48 hours of life.

We create this filter using a Time property. Normally, this creates an "$online" property to track how long it's been since the last time that anything was received from the device. But by clicking the Advanced box as you create the filter, you can select a specific property to keep track of. The trick is that the $id property (the unique identity of each device) has a timestamp which is only updated the very first time that the device is seen.  So by selecting $id in the Advanced tab, you can create a filter that is true for the first N hours or days of a device's life, and then goes false. Make this the Scope of a KPI and - voila! - your KPI will capture only devices that have just been born and are therefore in your "Kindergarten".

Now you can create dashboard metrics and automation processes especially for those devices.