Metrics In Production Environments
Raul Puentes Friday, August 24, 2018
Share this page with a friend
As we have mentioned in our prior article. Metrics are extremely necessary, we basically use them to measure everything, is a human need, we need to know the time, we created a metric system, divided the day in hours and created a device to tell us the time, we went beyond that and found the need to create a system to actually measure the days that go by, years, decades centuries we need to keep track of where we are in time, so man created a metric system and named it a Calendar, so basically the point is, metrics are well darn important.
Let us dive into an area of metrics application where I believe the accuracy of the data measured should be tip-top. Production Environments.
Production environments could be in any type of company, or factory, financial institution we need to keep up with the production of products, services, including sales, number of employees, big companies believe or not have a huge daily turn around that we may not see, employees leaving, entering just the same way as people are born and die in this world. Sales are a must, we must know how much money our new shoe fabrication business is doing, how much money is making, and how much do we owe Uncle Sam in Taxes. Very basic, right? Well, production environments are the most challenging metrics to obtain accuracy on, the constant production and minute by minute sales make it virtually impossible to keep up. This is again where technology plays an incredible role. Imagine the technology that companies like Amazon invest to keep up with approximately 35 orders per second!? 1.1 Billion Sales a year. Incredible right? The algorithms and systems running constantly need extremely careful supervision, as you know, it doesn’t matter how high tech your systems are, computers crash from time to time.
Let’s take another aspect to show the importance of having accurate data. Suppose you are in charge of the metrics and MIS (management information systems) in one of the biggest banks in the world. More specifically you manage the metrics of the AML department (Anti Money Laundering). So how important is the AML program for a bank nowadays. Well, let me enlighten you.
After the terrorist attacks of 9/11, President bush came up with the USA PATRIOT ACT this is not a just a cool name for an anti-terrorist plan, it actually stands for Uniting And Strengthening America Providing Appropriate Tools Required to Intercept and Obstruct Terrorism Act of 2001. If you are ever curious, Act #3 of such document contains the Anti Money Laundering Program. This forced government financial regulatory institutions like FinCen (Financial Crimes Enforcement Network) and the OCC (Office of the Comptroller of Currency) to push banks to establish an AML program and no, is not an option, if they don’t implement this successfully, Banks and other financial institutions face incredibly high penalties and also risk their luxury of actually staying open for business, is the law.
Now that we laid out the importance of such a program we will dive back into the production environment of such programs. So the technology systems and algorithms monitoring our accounts for any change in behavior (CIB) spit out an average of 100 to 200 cases a month per analyst in an AML department containing about 500 analysts (talking about major banks like Citi, Wells Fargo or Bank Of America, have in mind this are estimates). If you do the math you are talking about thousands of cases generated automatically not including the alarms that are manually set at a branch employee when to guys wearing yellow chemical suits deposit 9,900 dollars in cash every other day (yes is a Breaking Bad reference). As you can see is quite a high volume of cases, all of these having to be revised personally by a human being, and having this human decide if the behavior seen in the account is relevant and makes sense for certain type of costumer, or not and should be escalated to legal departments and eventually law enforcement to go after the Narcos laundering money and cleaning it up to go into our economic system. Each analyst goes over approximately 8 to 7 cases a day, all in the different status (pending closure, during the investigation, on hold for X or Y reason, etc. also this records must be completed during the allocated dates imposed by the government agencies we mentioned above. No pressure, right!?.
Well now that I scared you enough, and you are that human in charge of mapping all these metrics for this oh so important AML department at the biggest bank in America. How do you keep up with hundreds of analysts producing in conjunction thousands of investigations daily, every minute, every hour there is the one generating, and another one being completed, oh wait! Another one being escalated! And the metrics you produce are the guidelines for the production department, to see if they are on track or they are falling behind and need to tell the drum banging guy to bang the drums a little faster so the slaves row faster for the ship to move faster Battle Rhythm!!
Basically the future of the bank after a quarterly audit by government offices rests in your head. Again no pressure.
Here we can see how important, production metrics must be. There is something called “real-time data” which refers to the accuracy of the data produced. If you create a report for the regional director of the department and the last time data was pulled was at 8 am, in a normal banking hour company you are basically showing him what was completed the day before, but if you had to take the kids to school that day cause your wife had the flu and you ran into traffic in the interstate due to a semi flipping their load all over the road and you didn’t make it work by 11 am and sent the report at around 12. Well in 4 hours a lot could be have changed, and your boss is now receiving inaccurate data. It is sad to say that in my experience, this problem today is still a huge issue the company has to deal with. Let me tell you unless you possess servers and technology like the one Amazon has, you will run into this issues, there is really no such thing as real-time data anymore. Thankfully there is quite a number of tech companies, I have my favorite one which I am actually trying to propose at my employer location, that is working tirelessly to get close to producing reports as close to real time as possible. Think about it, in the example, I gave you, where you have to deal with cases that investigate money launderers which in fact may potentially fund terrorist initiatives and other bad guy action, how important do you think your job is? Like one of my favorite memes says… I am just going to leave this here. And remove myself slowly, and let it sink in.