A recent blog article highlighted obstacles for micro-sized companies taking advantage of servitisation and internationalising opportunities due to resource restrictions. Similar problems can be anticipated with respect to opportunities within digital transformation. While a larger company can have sufficient personnel resources to have specialists within information technology and data science, micro-sized companies with fewer than 10 employees are more likely to have staff that are required to be generalists rather than specialists, and not have the capacity to assign staff to specialise on digital developments.
Large-scale process industry has an interest in SME suppliers having a sufficiently high digital capability to be able to interface with them. The activation energy barrier that SMEs have to overcome to participate in digital transformations is the basis for the establishment of Digital Norway. This division of Norway’s ‘Top Industry Centre’ is established by large companies to give assistance to SMEs in digital transformation capabilities.
There are different approaches that can be taken by SMEs. One way is to outsource the technical development aspects and focus on the implementation strategy and practicalities. There are many commercial platforms, of a ‘plug- & play’ nature, for example IBM Watson , Google AI or Microsoft Azure . There is still a need to understand some technical details of what the algorithms do, but one does not need to write programmes from scratch.
These commercial offerings are typically made available free to install, but require payment for the features used – such as amount of data storage, which algorithms are used, etc. Without prior knowledge of the data processing requirements it can be difficult to anticipate the costs, and for micro-sized companies it can become a significant load, and SMEs are more sensitive to unexpected costs.
Another approach can be to use freeware, which however requires greater in-house competence. It can be used as an initial approach to identify which algorithms are relevant to use and make a more accurate estimate of the likely scope and costs to implement using a commercial provider. To a large extent, the different data processing algorithms, such as within machine learning and artificial intelligence, are available for free in libraries made available from programming communities such as python or R.
There are advantages and disadvantages with different approaches. Outsourcing can result in losing control over knowledge and understanding of the company’s own process and becoming dependent on external expertise. Data programming has become more complex, but at same time more can be achieved through relatively simple scripts that call functions from libraries. Doing more of the technical development in-house requires some competence-raising to understand what possibilities are available and how to access and realise them. There is currently a shortage in IT-competent employees, and this is expected to worsen. There is tight competition for IT competence, and even large companies face difficulties in hiring the needed competence, which can lead to a further squeeze on SMEs. Building up in-house competence can be a good strategy for ensuring long-term business opportunities.
Ross Wakelin
Northern Research Institute Narvik A.S.
ross@tek.norut.no
(47) 99 252 485
