Cloud for Business Processes

Authors: Jānis Tupulis, chairman of Latvian Open Technology Association, Rasa Gulbe, Member of the Board at DATI Group

In our data age, the true value of data depends on our ability to process it and use it in a meaningful way. For smaller businesses an Excel table is usually enough, but for larger enterprises business analytics software and apps are invaluable. With their introduction the need for additional computational power increases as well. This is where a cloud can come to aid by providing resources and balancing them with other business processes.

Before a software purchase, we are more and more often offered to test it free of charge. After this trial we can purchase the solution as a service or deploy it in our infrastructure (On Premises). How do you choose the best option for your company?

Identify the need
First of all, the need for business analytics-specific infrastructure solutions is dictated by the volume of data storage. The information needs to be relevantly structured in order for the company’s management to be able to efficiently use the analysis tools. Research by IBM shows that up to 80% of time currently allotted to data analysis is spent on retrieving and structuring the data for analysis [1]. Furthermore, companies often have multiple dimensions of interest in data, for example, across time periods, regions, business operations, etc. Therefore the operating system source data can be just a single table while the volume of data prepared for analysis can increase several times over, because a separate data set is needed for each data section.

Secondly, infrastructure needs are dependent on the frequency of data updates, which, when it is necessary to gain operative insight, can require larger computing resources. It is important to balance computing power, for example, by mainly loading the operating systems during the day and performing analysis at night. If this is not possible, additional resources have to be allocated to analytics. This can be provided, for example, by using cloud services.

Calculate the returns
The use of complex data analytics tools doesn’t always require the company’s own data center. The switch from CAPEX to OPEX over the last decade has often meant that IT solutions are subscription-based or, in other words, software is purchased as a service (SaaS). Small and medium enterprises often choose this path. SaaS doesn’t require you to think about infrastructure, because the end application is being rented from the vendor who maintains it.

However, not all solutions are available as a service. In that case, the company has to combine elements in order to introduce and maintain the solution. If it is impossible to purchase the solution as a service, the company must either look for necessary infrastructure providers, rent cloud computing power to run business applications or provide part of the solution themselves – on its own IT set-up.

When choosing a purchase model, in the case of business analytics, the transfer of data from the operating systems to the business analytics infrastructure must be specifically considered. Meaning that if all systems are located in a single data center, it is relatively easy to provide short transmission and waiting times (latency). If, however, the company purchases all the infrastructure resources as a service or uses a hybrid model, then both specific infrastructure and functional solutions must be provided for data transmission.
Before purchasing a cloud or creating a data center, the company should calculate the return on investment in processing time and power.

Consider system and security risks
Business analytics data is one of the most valuable company assets, therefore it is important to evaluate all aspects of security – not only how secure the purchased cloud is, but also how secure the processes of the company itself are.
Purchase of a solution as a service can initially provide a secure base, but, as data or the number of people with data access grows, a self-managed infrastructure will give you more control.

When choosing an external cloud service provider, extra attention must be paid to its reputation and implemented security standards. Find out from the software developer in what infrastructure and in which country your solution would be hosted. Check what ICT security standards the provider works with and whether the services of interest are certified by an external security auditor. If the plan is to deploy the solution abroad, pay close attention to the data security legislation of the country in question.

Regardless of which option you choose, you shouldn’t forget about employee trainings. The World Economic Forum Global Risks Report 2022 shows that 95% of cybersecurity risks are related to human error. [2] In order to reduce the leakage of data and insights used in analytics, it should be ensured that all team members are able to observe digital hygiene – taking care of passwords, backups, two-level authentication, etc.

In order to select the best solution, it is enough to define what insight you want to gain and what data will be necessary for that insight. However, to understand whether to purchase software as SaaS or deploy the solution in the infrastructure of your choice, the analytics process has to be evaluated.
   • Will storage and processing of large amount of data be necessary to gain insight?
   • Will frequent data access and data updates be needed to gain accurate insight?
   • Does hosting data on an infrastructure you don't manage increase data leakage risks?
If the answer to all of the questions is “Yes”, then you may need to evaluate the option to set up your own infrastructure (On Premises) before subscribing to cloud resources.

‍