What kind of child do you have?
Your ERP system is a living, breathing thing and much like employees and your children: you only get the best out of them based on the effort you put in them, the attention you give them, the treatment you give them and what education you give them.
The innocent teen-girl artificial intelligence chat robot named Tay that Microsoft launched, turned into a Hitler-loving sex robot based on tweets received from the weirdos on Twitter in just 24 hours. (Source: Telegraph). If you still have an old on premise ERP system, locked away sound and safe, that only has the occasional chats with other business applications, you don’t have anything to worry about, or do you?
If you are running your business on the latest cloud based ERP system, having business intelligence and machine learning capabilities, mashing your trusted data with all kinds of information received from the internet (of things), you might have a lot to worry about, or don’t you?
Who is teaching?
To answer those questions, we have to answer two other questions first. The first question is “who is talking?”. In other words, do we trust the source of the information and where does that source gets his information from?
The second question is “what does your Intelligent ERP system do with the information received?”. If the information received is corrupt and your ERP system is taking actions based on that information, you have something bad going on. An example: your ERP systems prints an aging report of your customers that owe you money. Wouldn’t it be nice to mash your aging report with data telling if any of your regular, well trusted, customers are subject to a lawsuit or are even filed for bankruptcy?
Probably the answer is yes, but is this something that your ERP system should process automatically, or should a human being be the interpreter of the information received? Is it even possible for human beings to process all the information coming from private databases or the internet?
And on top of this all, your ERP system is also getting information from hundreds of employees entering data. Will they input all data correctly or are mistakes made? And will your ERP system behave irregular because of those mistakes?
The answer to all questions above and the reason why Tay got twisted can be explained in one word: “patterns”. If enough entities do or say the same thing, that thing becomes the truth. The one not acting accordingly is an anomaly.
The only point in monitoring all your data coming in into your (ERP) system is to detect the anomaly. Since it is undoable for most human beings to detect the anomaly in terabytes of (external) data or in any of the thousands of tables used by your ERP system, you need an anomaly monitoring system doing so.
Goodbye KPI, Hello Continuous Anomaly Monitoring
The old way of detecting anomalies is to define a KPI and everything exceeding the boundaries is an anomaly. For example, your revenue is between 1.5M and 2.5M each month. Everything outside these boundaries should be investigated. When the company grows the boundary values of each KPI should be adjusted accordingly. In our example to maybe 2M to 3.5M each month. With the company growing each year, this will become an iterative process.
This way of detecting anomalies is fine, but your BI is now mashing up data from your CRM system, your ERP system with machine learning capabilities and whatever it finds on the internet and you need hundreds of KPI’s to measure all that, or simply be happy with the few KPI’s that really define your business, ignoring the rest.
The alternative here is to use a continuous monitoring system that simply processes any data and find the anomaly for you. Such an anomaly detecting system can tell you that employee X for the first time in 12 year didn’t show up, it can tell you that revenue in the BFSI sector in the UK is down but goes up in The Netherlands and it adjusts the KPI boundaries for you based on the values interpreted.
Kaya Consulting tools
Kaya consulting has build a variety of tools helping you to make sense of your data. From anomaly monitoring systems to dashboards where you can define your own KPI’s. From continuous auditing systems to tools enforcing your data in a certain pattern. Our BI Analytics team is here to help you make sense of it all, give us a call to help you to unleash the potential of your data.