Digitisation? Disruption? Decisions!

0

Can you still hear it? The "D" word. For years it has been haunting conference keynotes and through clever and less clever books. The Axel Springer Group has made it. The German automobile manufacturers have not yet but are catching up. My CD player at home didn't survive – even though it started.

It is, of course, about digitisation. And the resulting "disruption" – the second-favourite word with "D." Basically, however – and this is becoming increasingly clear – it is a matter of "decisions."

Hidden Insights: Digitisation? Disruption? Decisions!

Hidden Insights: Digitisation? Disruption? Decisions!

Why are decisions important?

You are sure to have some analog processes in your company that could have been handled digitally long ago. Documents and forms are examples. Nowadays, these can be easily digitally captured or filled out by app. This will undoubtedly make you a bit faster and more efficient. But such things alone are not enough to gain a competitive advantage.

The real progress lies elsewhere: Digitisation makes it possible to automate decisions. This is the important next step! We are in the midst of a radical change in the analysis of data in companies. Previously, BI and analytics were primarily used to help people make decisions. Comprehensive data sharing and new technologies such as machine learning enable systems to make decisions – often better decisions – automatically and at scale.

For us, the question is to what extent are we prepared to leave the autonomy of decisions to technologies? Discussions continue about autonomous driving and modern diagnostic procedures in health care, for example, as well as questions about the ethics of algorithms. All of this shows that we are on the verge of taking this next step.

We must drive forward operationalisation, the automation of decision-making processes. When you effectively combine human and machine, you gain the real edge.

What do you need for that?

On the one hand, we have to react quickly to changes and understand new situations in unknown territory. So we need great freedom and flexibility in both the business and IT areas. I call that "choice."

1. You need maximum flexibility and speed

The departments need the right tools and analytical applications: machine learning, speech processing, image recognition, etc. You are going to need R, Python and SAS. Sometimes this depends on the type of question, sometimes on who is currently available in the team with which skill set.

If you develop models, you need rather large in-memory capacities. In deployment or live operation, on the other hand, modern container technologies help you to absorb load peaks in computing power.

In addition, there is the factor of speed – today often referred to as "fast fail" – that is, speed from the idea that arises somewhere in your organisation to the qualification. Is the use case load-bearing? How can the opportunities and risks of implementation be assessed, and how can bad investments be avoided?

IT must be set up in a correspondingly agile way. Perhaps you operate in one environment today and another tomorrow, outsource services, keep others in house. Your data scientists always want to try out the latest algorithms. You need competent IT contacts so that you can scale exactly where performance and availability are needed.

2. You need maximum control

On the other hand, you need high stability and reliability. I call that "control." If there is a copy of your customer data on every desktop for agile development, then you are no longer compliant. When the digital lab in Berlin works on visions in a vacuum, you generate more friction than progress.

Ask: What must your analytical platform achieve in the future? Click To Tweet

The decisive aspect for success, however, is trust. Your employees' confidence in the quality of the model results, your customers' confidence in the decisions made and, most importantly, the supervisory authorities' confidence in your responsible handling of data. Without this trust, you cannot achieve automation.

What must your analytical platform achieve in the future?

In order to keep pace with the requirements of the future – that is, with your requirements – we are rebuilding the SAS Platform with all our might in the direction of containerisation and cloud capability. SAS Viya 3.5 will be based entirely on container technology. So if we apply these catchwords "choice" and "control" to our performance promise to you, what does that mean in concrete terms?

You get broader choice and more control:

  • New analytical possibilities: ML, NLP, VDMML, open source – all this is available to you with SAS Viya as an analytical platform.
  • Then we'll talk about the cloud: Which approach do you prefer? Private cloud? Public cloud? Your choice.
  • The whole is supplemented by container technologies, such as Docker and Kubernetes. What that means for you is simple deployment, simple startup, simple shut down and hybrid scale. More flexibility – that is, choice – is not possible.

You also get model management to know if the model is still good or bad. Has the data changed so much that you have to change the model? This includes compliance so you know which model with which data is the current one.

And finally, there is keyword automation, deployment of analytical models in live streams, on an edge device, in batch, in containers, etc. The best model is of no use if you cannot bring it into production. This brings us from model management to intelligent decisioning. And from intelligent decisioning to what digitalisation is all about at its core: the true automation of decisions in an integrated ecosystem. What do you think of that? How far along are you in migrating your analytical platform to the cloud?

SAS conducted a survey on the topic and if you would like to receive an industry report with insights into how you industry compares to the market, please register here.

Share

About Author

Andreas Gödde

Director Business Analytics

Andreas Gödde is specialist for strategies around Big Data Analytics, Digitalization and Internet of Things, helping organizations to get insights from data for business decisions. He leads the presales organization for Business Analytics of SAS in Germany, Austria and Switzerland. Andreas has a 25 years background in advising companies around Business Intelligence, Data Warehouse and Big Data concepts and projects. Andreas graduated in business informatics in Mannheim. He joined SAS in 1994 helping developing and growing the professional services organization in different management roles. In 2006 he moved to the presales organization building up teams for technical and strategic advisory for customers and for emerging technologies and trends like Big Data and the Internet of Things. Before joining SAS he worked for BASF in Ludwigshafen.

Leave A Reply

Back to Top