One of the strengths of the Shepherd platform is its flexible and comprehensive alerting: we use SMS, email and landline & mobile calls to alert configured responders. This means that we can reach people in way that is efficient but suitable for the level of risk we have detected.
During on-boarding each item being monitored (e.g. temperature, pressure, etc) can be assigned a Criticality (High/Med/Low/Notify Only). When an anomaly is spotted or predicted we use this criticality along with other factors to determine which escalation level to alert at: e.g. for a sensor with a criticality of High showing a severe anomaly at 2am we would begin at Level 4, where as, a sensor of criticality Low with a mild anomaly at 10am on a weekday would begin at Level 2.
You have full control of who is contacted at each Level: so lower criticality alerts can be sent to on-site staff, for instance, who can investigate and correct issues before more senior staff are troubled with them. However, all alert levels of 2 and above auto-escalate if responders do not tell Shepherd they are investigating.
This diagram is downloadable as a .pdf and is available as an attachment to this post.