Cloud Key in Data Strategy for Mexican CompaniesBy Jan Hogewoning | Wed, 10/07/2020 - 09:38
Q: What has changed in the application of data analytics models in Mexico over the last two years?
A: In the banking and retail sectors, there has definitely been an acceleration in the use of analytical models. We see that companies do not just stick with static models that tell them who could be the best client. They are now using models that generate a real-time response on their platform to offer a personalized product at the moment in which they have contact with the client. For example, in an online marketplace, companies adapt the advertisement banner to the behavior that the customer is exhibiting at the very moment of their visit. While not everyone has reached this stage, most companies are definitely moving in this direction. Obviously, the volume of data that companies are using has also grown.
Q: What technologies are at the basis of this acceleration?
A: It is a combination of different technologies but the most important is cloud. SAS is particularly active in the banking, insurance, retail and telecom sectors and everyone is moving to the cloud. The cloud environment offers three main advantages. First, it offers significant cost savings. Companies do not have to invest in hardware infrastructure to manage and analyze large amounts of information. Second, the cloud offers elasticity. If you need more capacity, you just ask for it. Third, the cloud allows you to test models much faster, with more flexibility and at lower costs. Companies that have the most success generally let their data scientists play in sandboxes, with multiple models and small populations. To do that without the cloud is much more time-consuming and expensive.
The Mexican banking sector is the most advanced in artificial intelligence-based models for analyzing data. Now it needs to integrate everything into the cloud. The industry is considering what can be uploaded to the cloud and what cannot. This is subject to regulation. Data needs to remain anonymous through encryption. SAS has special security management schemes that facilitate this process. Retail is also starting to hit this phase.
Q: Considering these new tendencies in the Mexican market, how has SAS adjusted its services and products?
A: SAS has been building data analysis models and software for 40 years. A recent Gartner report demonstrated that for more than 10 years we have been the global leader in data transformation and manipulation. However, we are broadening our services in several areas. The area in which we are growing fastest is credit card fraud prevention. The primary threat is identity theft, where a credit card account is requested with that stolen identity. During this pandemic, fraud has grown rapidly, particularly in online purchases. Companies are no longer dealing with a few large crime groups but many individual criminals. This is an area where we offer a specific product that allows companies to detect fraud faster, with more specific targeting and greater prevention capabilities.
SAS also has a strong risk analysis area. We help banks understand which products will fall out of favor, how much reserves they should have, which segments they should focus on, to what level they can take a risk with the unbanked sector and more. We have been doing that for years. More than 80 percent of banks in Mexico use us to carry out credit risk assessment.
Another area where we are active is customer intelligence, where we analyze the customer’s journey, generating real-time understanding using both the web and applications. In the case of banks, you use this to understand the capacity of debt to provide a loan that helps the customer achieve what they want without becoming over-indebted.
Q: How does your software adapt to different cloud providers?
A: We work with all cloud providers and will remain non-exclusive. However, in June, we signed an agreement with Microsoft and Azure is now our preferred cloud. The company is the biggest cloud provider with the largest amount of data. With SAS being the largest global analytics provider, covering more than 35 percent of the market, this alliance will allow us to improve our service. Automatically, Azure clients will be able to get SAS solutions through their cloud. What used to take six to eight months, as companies had to install the hardware, now can be implemented in a matter of weeks.
Q: How does the cloud help the Mexican manufacturing industry?
A: With the arrival of IoT, manufacturers can now benefit from analysis of production processes in real time to reduce inefficiencies and incidents. For a manufacturer that produces thousands of units a minute, doing this manually is practically impossible. Using all types of data collected by sensors and cameras at the production plant, you can employ artificial intelligence to identify potential problems. The cost advantages are huge. Two or three years ago, this was only emerging in Mexico. Many companies now have the hardware in place and using the cloud they can run models continuously.
Q: How much competition do you face in the services you provide?
A: There are many open source models that come out continuously. Often a client comes to us with a new open source model that was released in the US, for example, that aims to innovate a particular aspect of their business. However, when it comes to actually making these models operational, companies can run into many problems. The main one is the lack of support. Many open source models are not backed up, which is where we come into play. We can ensure that a model functions 24/7 and not just in one location but all of them. Another issue is scalability. To create a functioning model in multiple servers, you need to grow your capacity of analysis. This requires a high level of programming with experts in multiple areas. When we execute an open source model within a SAS environment, we ensure it is scalable. We have over 180 consultants in Mexico dedicated to this. The ability to absorb other models out there and to make them run well is one of our key strengths.
Q: How are your helping to grow the pool of data scientists in Mexico?
A: Data science is one of the most rewarding jobs in the sector. At SAS, we have two programs that train data scientists. One is the Beca SAS (SAS Scholarship), where about 60 students who have completed their university studies undergo a three-week training course in SAS models and platforms. We then introduce them to our clients and almost all are contracted within three weeks. Another program is SAS Academy for Data Science, a course of six to eight months that aims to provide formal certification as a data scientist. Here, students receive in-depth training in the manipulation of large volumes of information, machine learning, artificial intelligence and statistical modeling analysis. With this program, students can achieved the SAS Certified Big Data Professional credential and the SAS Certified Advanced Analytics Professional credential, earning the distinguished title of a SAS Certified Data Scientist. A good data scientist needs to understand each method of data manipulation, even though today most of the work the data scientist will carry out is to monitor if automated models are functioning well. We provide this long program once a year and usually between 20 and 25 individuals graduate. It is often used as a path to a career switch by professionals.
Q: What specific benefits are e-commerce clients seeking with data analysis and ai models?
A: During the pandemic, we have adjusted our offer for e-commerce. Our solutions are focused on two areas, one of which is demand planning. Our real-time models allow a client to generate better prognostics of fast-changing consumer behavior. This allows them to optimize placement of products at distribution centers and reduce delivery times. Mexican consumers are not only purchasing more online, they are also demanding much faster delivery times. The other area is to improve the online offering. This means analyzing what users do in the portal, outside the portal, where they come from and what their profile is. In e-commerce, the number of people who actually end up making a purchase remains low. By helping to customize what the customer sees in that moment, the average ticket per customer really goes up.
SAS is a global provider of analytics software and solutions. It provides a wide range of products for the banking, healthcare, life sciences, manufacturing and retail sectors and also to the public sector