Big Data Survey 2015: 4 Core Insights for More Success With Data
Big Data Survey 2015 is the first national research about the use of Big Data within Dutch organizations and aims at providing an insight in the current role of Big Data, level of adoption, intentions and potential pitfalls. The survey is an initiative of trade show Big Data Expo and the data engineers and data scientists of GoDataDriven.
This article shares four core insights from Big Data Survey. Organizations with a data driven vision and a management that strives for a company culture that supports experimentation, are able to actually use data. Employer branding plays an increasingly critical role to attract the right people. Big Data Survey shows that organizations are aware of the fact that they should be operating more data driven.
“Big Data is not just about data, technology and skills", said Rob Dielemans, “we have learned that especially the organization and processes are critical success factors for success with data”.
Data success starts in the board room
More and more markets are overturned by small start-ups that, unencumbered by a legacy of hierarchy and fixed processes, are able to adapt quickly to market opportunities. Next generation energy companies and telecom providers shoot up the ground, and not for long, the same will happen to banks and other industries. Companies like AirBnB, Uber, and Alibaba disrupt the traditional industries of respectively hotels, taxies, and product trading. Change is indeed the new normal. At the moment only 5% of organizations claims to be far ahead of competition, while 39% is somewhat behind the competition. Time to embrace the opportunities of new technology and data.
And where does this change start? Dutch companies point at their management to play a vital role in the adoption of a data driven strategy. Management should provide direction and should support the organization in its process of change.
“A successful data project clearly begins with support from the management and a proper strategy. That gives direction and the support the organization needs to move data driven processes to the business”, adds Walter van der Scheer, chief marketing at GoDataDriven.
A large part of organizations (37%) claims that their management has appointed a strategic role to data, 27% partly agrees. For nearly a quarter of the companies (23%) a lot is to be gained on this part.
Organizations need to develop experimentation skills
Although organizations are aware of the necessity to apply data to everyday processes, adapting the organization to a new way of working is often a big challenge.
What do you think is mentioned as the main challenge in the field of data? Collecting data and investing in the right infrastructure perhaps? No, not at all, these are in fact the smallest challenges. Organizations find it particularly difficult to (find the time to) experiment. Organizations seem to be mostly process-oriented and have not embedded experimentation in the normal work routing yet.
Start-ups and IT-departments have been in the know for long: in order to respond to new insights, like market developments or results, a flexible work process is essential, also known as agile. Corporate agility means working with flexible project teams and short iterations focusing on the completion of manageable jobs. Traditional 20th century companies mostly are organized hierarchically and process oriented, leaving little room for agility. Experimentation, let alone unsuccessful experimentation, does not live within these companies. When experimenting, having a few unsuccessful experiments that can’t be used in the business is inevitable, but should definitely not be the standard.
Before an organization begins with data experiments, obviously it is required to collect sufficient amounts of data of good quality. For most companies, however, data is relatively easily made available, and current available (open source) technology allows companies to extract, transform, and load data efficiently.
Most often, the challenge is the organization itself: time to experiment, making data from various sources available for the business, and attracting the right talents.
Data offers an opportunity for every organization
According to 90% of the participants every organization should use data to improve processes. Most organizations apply data on the customer side, like within online marketing, webshop or online application, market analysis and customer service. The level of data applications disappoints: only 1 in every 10 organizations personalizes its website in real-time based on the behavior of a visitor.
Only a small part of organizations applies data to back-office processes like supply chain management and logistics, finance and administration, and HR. Clearly, organizations expect quicker and better results from marketing related processes, possibly also because of the data that is already collected from these processes. For companies that take their first steps with data this is perfect as positive results contribute to support from the management. Still, back-office processes should not be overlooked, as often substantial profit can be gained here.
Most organizations apply data for marketing purposes, like:
- direct marketing and online marketing (19%)
- online, like webshop and online application (16%)
- market analysis (14%)
- customer service (8%)
“Organizations often opt to develop data applications for processes that have an immediate business impact. Especially for a launching project it makes sense to explore the data opportunities within marketing. Often, there is business value to be found in less obvious processes as well. Think about predictive maintenance, production optimization based on predictive models or fraud detection for a financial institution”, remarks Rob Dielemans, managing director at GoDataDriven.
The rise of the data scientist
In the past, high cost of storage and limited processor power of computers made it difficult, if not impossible, to process and leverage large amounts of data. Business intelligence was used to look at data from restricted time periods. Useful for insights, but not so actionable, as the data used was not at all real-time. Nowadays it is possible to use nearly unlimited amounts of data for real-time analyzes. Thanks to the real-time component it is possible to use incoming data for immediate predictions. This has lead to the rise of the data scientist; academics who are not only math and statistics savvy, but who also mastered the art of software development. And last but not least: Who are able to translate insights to business decisions.
Over 60% of companies claim to have business analysts and data analysts in house, against only 38% who employ data scientists. Interestingly, the larger an organization is, the more it has data scientists.
In order to attract data scientists, it is important to be attractive as an employer. Corporate behavior, a management oriented culture, and fixed processes, are not part of this image. As transforming a large corporation is not done over night, this often is the reason for larger, traditional, businesses, to create a new, more agile, organizational structure outside of the existing business.
Being attractive as an employer is not all about paying the highest wage. “Data scientists look for challenging and stimulating environments, that provide room to work together with like-minded people, and to experiment with data and technology, so that the data scientist can develop predictive models that truly have an impact on the business", explains Friso van Vollenhoven, Chief Technology at GoDataDriven.
What is the best next step for organizations?
Many organizations have difficulty leaving hold of a process driven way of working and to re-organize. When organizations aspire to do more with experimentation it is necessary for the management to develop a vision and support the business. This is more than just releasing budget, but also the development of corporate agility, where it is possible to try many new things in short iterations. Fail fast.
Data applications are still in its infancy. Websites mostly are and do not adapt to visitor behavior, data is used from CRM or marketing database and is mostly numerically.
The future looks bright: participants claim to work on the implementation of advanced technologies like predictive models and artificial intelligence. Within three years 50% of the organizations expect the first applications of advanced technology. Data scientists play a crucial role in this advancement.
Download the full Big Data Survey report, in Dutch.
Follow us for more of this
Fairness in Machine Learning with PyTorch
May 22, 2018
Elitist shuffle for recommendation systems
May 13, 2018
GoDataDriven open source contribution: April 2018 edition
May 06, 2018
Event Report: Blockchain Ask Me Anything
May 03, 2018
Towards fairness in ML with adversarial networks
April 27, 2018
GDPR: "May 25th is when it really begins
April 10, 2018