Second Wave of Digitalization - the whats and hows

Challenges with the Second Wave of Digitalisation

Not yet quite here, and already cumbersome... Efficiency of invested money into digitalisation in the „first wave“ is disappointing. If you were around in this business in the early 2000s, you'll have a déja-vu: figures mimic those return rates of IT related technologies back then. Did we learn anything? Can we expect any better ROI from the second wave? We attempt to give a practitioner’s answer.

Flashback – case 1: ca 20 years ago project members were trying to consolidate a fragmented legacy database to import into the new ERP solution. They efforts were in vain: the solution had to go live without homogeneous historic data – management reporting was therefore heavily compromised.
Fast forward – case 2: it‘s early years in Industry 4.0, an ambitious digitalisation team was facing reality, when historic maintenance data couldn't be brought to the quality levels on which predictive models could run. They needed to go back to square 1: defining & collecting the data they needed.
And just recently - case 3, a powerful team of maintenance and data experts are desperately trying to make sense of the terabytes of data collected from their machines over the years. After months of hard work they end up without anything meaningful. In the next round they includ even more data – to come to the same result.

This little retrospective (and surely you can list your memories too) is to highlight, that underneath the packaging through new terminology, the “beast” is the same: DATA. So, what can be learnt, what practical suggestions can be made for today‘s data analytics projects:

DEMYSTIFY AI! Make sure you have a fair understanding about it! Machines do not “learn” and they do not get “intelligent”. Random forest, decision trees, and so on are statistical methods developed many decades ago. What is new, is the availability of computing capacity that makes those statistics approaches accessible and utilised also in the corporate world. Consulting independent, experienced centres of excellence (relevant department of High Schools for example) and literature can help understanding the mechanics of data analytics / artificial intelligence / machine learning, ... (you name it) before you set your expectations and invest actual time and money into data & AI related projects.
FOCUS YOUR EFFORTS: Know what you are after. What is being looked for in/with data must be in great granularity clear and aligned with strategic objectives from the beginning. Less is very often more, especially when it comes to data analysis efforts (costs). "Evolution instead of revolution" - as they say.
QUALITY IS THE KEY to success. Garbage-in – garbage-out: like it or hate it, the good old maxima is still valid – algorithms can only deliver meaningful insights on quality data with the right timing and reliable recurrence. Artificial Intelligence is no magic glue, crystal ball or a new version of alchemy. Sorry to say: low quality data will result low quality analytics. Quantity is neither indicative for quality nor for results. Please note: data quality cannot be achieved in silos: it requires a cross-company approach and ownership.
PREPARE YOUR ORGANISATION. We do not intend to join the debate whether AI will replace or augment current jobs. Certain is, that organisations and individuals need to find their way around by having clear guidances (processes), developing digital dexterity and learning to co-exist with AI empowered machines. Managing the transition in its entirety is the responsibility, and good interest of the management. It certainly needs to be started well before the first interaction between human and machine takes place.
PATIENCE & PERSEVERANCE Expectations must be set and kept realistic. Not just that a lot of data must not mean a lot of insights; companies excelling in data science today built their capabilities step-by-step over the (many!) years. They handle and develop data and analytics as an ecosystem with long term view and objectives.

We hope, you found this article useful for your initiative. Although it might all look obvious, it's deceiving. Majority of the companies are struggling with putting Artificial Intelligence in context and getting real returns out of it. It is possible, but it is a long-shot game.