Business Analytics – Tackling critical business processes


More and more often, data specialist SMT helps customers to perform better and achieve their strategic objectives with Business Analytics. CTO Lex Crielaars explains how this works.


“First, we always look at the critical business processes of the customer,” Crielaars begins his story. “They differ per organization, but are always there. Because it can always get better, no matter how well it’s going. For example, companies with a web shop want to know more about visitors on their website who are not becoming customers. At which point do they drop out? That is crucial information for the company’s revenue model. ”


Red blocks

Once that has been mapped out, SMT starts tackling one process. “We define the components within that process and the associated KPIs. That sounds easier than it is, because you have to agree for each component what it takes before you can say that something works well. What values ​​do you attach to this? Only when you have done this for all the blocks, you can lace the entire chain together and get end-to-end insight. You then act on the points in the chain where things still go wrong, the red blocks. ”


The modeling of these processes can be done manually, but also automatically by using process mining, a form of artificial intelligence. “Usually it is still done manually, but organizations are increasingly opting for the second option. The possibilities are becoming ever greater.”


The chain is fully recreated in the SMT EDAP® (SMT Enriched Data Analytics Platform®). “If the chain is clear, we will be twisting the knobs together with the customer. Eventually you will be in a situation where everything, from technology to people, is coordinated in such a way that a clear improvement has been achieved that can also be measured. You have moved from a situation where you had no insight whatsoever, to a situation where you see exactly how you are doing in all kinds of areas. Turning these knobs is then a continuous process. Markets change, so you have to keep tweaking to achieve maximum performance. That is only possible if you have real-time insight into your processes and you continue to manage them for optimum results. ”

Restore sales

Crielaars clarifies with an example. “For an insurance company we implemented a solution for the process of taking out insurance. At one point the pipeline showed only a quarter of the number of new insurance policies that we were expecting at that time. We could see that it was not the process or the infrastructure; everything was working as it was supposed work. We investigated where customers were going and noticed that a competitor had started a large advertising campaign with a discount. Our customer then started his own campaign and was able to quickly restore the level of sales. We only succeeded because we had real-time insight. If it had been discovered later, somewhere in a monthly report, the damage would have been enormous. ”


With another example, Crielaars indicates what else is possible if you convert combined data into knowledge. “Another insurance company had a machine learning algorithm to analyze a fire insurance dataset. It turned out that in the coldest month of the year, the fire insurance paid out ten million euros more than usual. The data showed that it relatively often involved old houses with a thatched roof. The background became clear. Old houses are usually poorly insulated and often have a chimney and a fireplace which is used during winter. If the chimney is not properly swept, it can cause a fire. The company then invested in a discount on chimney sweeps, specifically for that type of old house. The costs of this were not in proportion to the six million that, according to the calculations, had to be paid less.” Being able to calculate those amounts is crucial, Crielaars says. “If an employee of that insurer had gone to the board of directors to just tell them to spend money on a chimney sweep discount, that person would have been laughed at. But if you can prove with data that it costs half a million euros and saves six million, then you are a hero.”

In infinitely many other situations you can improve business processes by analyzing data, Crielaars argues. For example, by combating fraud. “With the self-scanners in supermarkets, for example. Whether you will be checked is not random. Among other things, they measure how long you have been inside the store and what products you have scanned. If that ratio is suspicious, you’ll be picked out sooner. We work in the same way for a telephone company where we can identify fraudulent conversations on the basis of all kinds of data.”


Three pillars

Because of the real-time insight that is required, the SMT EDAP® is crucial. It forms the basis for all SMT activities. Crielaars: “The big advantage is that with that one platform you can serve the three pillars, in addition to Business Analytics, also IT Operations and Information Security. There is not a single company that looks at the organization from just one of those three perspectives. You look at IT Operations because you want to improve the business. And that business wants it to happen safely, because otherwise you still won’t any get further. Everything is inextricably linked.”


July 2019