Kari works at Gofore as Senior Service Architect. He has been working 20+ years e.g. with data architecture and management, Business Intelligence, data warehouse automation and CRM with both public and private sector customers. Kari wants to demystify current data-related hype. AI, IoT, Big data, these things come and go, but the key is to build useful solutions utilizing those technologies, no matter how 'hot' they are in the market right now. The key to getting the best out of these solutions and technologies is good quality data.
Data and information are or will be, inevitable prerequisites for success, no matter what your business is. Data is said to be the new fuel, new services, new tools, new professions (very sexy ones), etc. Data is also said to be risky, vulnerable, unsafe, out of control, etc. All these statements are true and at the same time very confusing. This post gives some guidelines on how to tackle the ever-growing data and information overload.
Problem or opportunity?
The ever-growing amount of information can be seen as a problem or as an opportunity. Both these options are, not good or bad, but dependent on the relevant approach to data and information. By selecting your approach, you might even select your future career. Cybersecurity professionals love to solve technical data privacy problems whilst business development people see endless opportunities for new services and platforms. So even viewing information as a problem might be a business opportunity.
Bring the data together
Data and technologies become more and more complicated and diverse all the time. And sad to say, too often data and solutions are very isolated and siloed. Silos are a very big problem. Successful organizations are transparent. This means you need to get rid of data silos, especially mentally, and bring data together in one platform. This applies not only to your own data but to all data which might belong to other departments, other companies or to publicly available data.
Technically this is not a problem, there are a huge number of tools and platforms which enable you to put data in one place and to combine and integrate it. Usually the problem lies in people’s thinking and in organizational cultures. There can also be some legal restrictions, but still it is often possible to have some elements of the data transparently available.
Don’t try this alone
It is impossible for a single person, and in most cases even for a single organization, to do everything alone. You need a team, people for certain roles. On the football pitch, you cannot be the goalkeeper, midfielder and forward at the same time. In realizing the opportunities lying within your data you need people for different roles, e.g. data engineer, data scientist, developer, business analyst, security expert, etc.
The roles in your team depend on the approach you take and what you are trying to achieve. Tackling information problems requires engineering, security, quality skills, etc. Grabbing information opportunities requires e.g. experimenting, prototyping, visualization, AI, design and sales roles. To be successful in the long run you need to assess both problem and opportunity approaches. Like Tom Davenport says, “Great data teams play both offence and defence”.
The way forward
The glue between data, the development team, the production team and the technological platform is the way of working. You might say these are the processes, but they represent a strict and siloed way of working. The nature of data related development whether it is dashboard building, advanced analytics, AI model development, etc. is such that when you start, you don’t know where, and with what results you will end up with. In addition, new requirements will very probably emerge during the work. Traditional project methods for development are not capable of handling these ever-changing situations fast enough.
The line between development and production is very vague in the field of data and analytics. The transition from development to deployment and to everyday use must be seamless and continuous. You must understand constantly changing needs and constantly increasing data. Agile and experimental ways of working are the key. You must be able to show results, iterate them and adjust your direction continuously with your clients in ever-shortening cycles.
Execution brings results, prepare for change
The problems very often lie in the execution. Not only ready-made solutions but also small and more ‘academic’ experiments should be taken into use and deployed into production. If you don’t try them in production with real data, you don’t know what they bring to you. Another aspect of execution is actions. No solution or information is useful if it is not used or no actions are based on it. Usually, it is most effective, also cost-wise, to build up and kick-start new solutions with help from a partner(s), even in this case you need to take care of the action-part. When new solutions are deployed and new ways of working have become business as usual, then you can rethink what resources and competencies you need to have yourself and how you continue with partnerships.
The field of data & AI, like many others, is in constant development and transition. And like all development, it is not only about tools and technologies. People, competencies, ways of working and organizational culture are key parts of success. New development drives constant change, and in order to succeed in a changing environment, you, or at least your thinking, must also change!