Advanced Analytics is still a trend topic. It can form the basis for new business models, more attractive products and services or be used for process optimization. However, it is often unclear what is already feasible and what is unfortunately only desirable. But how are advanced analysis methods actually used today? Which companies predominantly use Advanced Analytics? How far along are the individual industries? Which use cases are actually implemented? The results of the latest BARC Advanced Analytics user survey provide answers to these questions. The annual study analyses the feedback from more than 250 participants from the DACH region. This time, differences between companies that are in the prototyping stage and those that are still predominantly in the operationalization stage will be highlighted.
With the results from our user surveys, discussions at conferences and experiences from projects, we have compiled the most important recommendations for action to be successful with Advanced Analytics in the third part of our blog series.
Maturity – the market is progressing
While the spread of advanced analytics is hesitant, existing initiatives are maturing. Companies still have the opportunity to belong to the group of pioneers that can secure competitive advantages at an early stage through Advanced Analytics. In order to ensure a smooth transition from prototypes to operationalization, the validation of prototypes should be taken into account as early as the solution design stage. This is where most problems occur or where most projects fail. By clearly defined use cases, the identification of benchmarks and KPIs, which measure the added value of the solutions or the targeted alignment of the projects with the corporate strategy, this validation can be considered early in the process.
Commercial software is worthwhile for operationalization
Commercial software also offers advantages in the difficult phase of operationalization, in which prototypes are converted into stable IT solutions with regular maintenance cycles. For example, advanced analytics platforms or directly usable standard web services provide more support than open source software.
Automation enables scalable processes
Manual operationalization in the form of exporting and sending the results is often sufficient in early phases of operationalization, but can tie up considerable resources, especially if several solutions are in use. In the long run, only the automation of model scoring allows scalable solutions and processes. This primarily requires IT resources, which most companies still lack.
Discover new data sources
Log data is used more and can enrich transactional data. Many companies are already analyzing data generated by their website or webshop to gain additional insights about their customers. Interesting use cases can also be derived from sensor logs and information from IT networks.
Take care of your data quality
Advanced Analytics is useless without the right data. It is about improving existing data in terms of quality, access and interpretability, collecting new data and linking external data. In this context, data management is an often underestimated but central discipline.
Invest in your staff
Many companies are launching massive training initiatives – Bertelsmann uses Open Online Courses from Udacity for this; Airbnb has developed its own training program. Training focuses not only on data scientists, but also on business analysts and data engineers. Advanced analytics can only be implemented through analytics teams in which all roles are available and well trained.
Get to know further strategic, technological and organizational approaches as well as best practices for the productive use of Advanced Analytics at the Data Festival on March 20 and 21, 2019 in Munich (https://datafestival.de) and take the opportunity to network with other participants, exhibitors, experts and like-minded people and exchange ideas in moderated workshops.