How to make the digital transformation of manufacturing happen?

The Digital Transformation is here: devices, technologies and suppliers are ready to bring manufacturing enterprises to a new level with increased productivity and more added value.

But if you go for a walk in production now, you’ll probably not see smart sensors, edge analytics-packed machinery, forklift drivers with smart-glasses and foremen yelling orders to their assistive bots.

The macro-level reasons why it is not happening or it is just slowly happening are covered in many posts in the Industrial IoT/Industrie 4.0 Viewpoints blog: there are cultural issues, a standards mess, and more.

Examples of cultural issues delaying transformation:

“Many engineers just aren’t comfortable with black-box solutions”

“Execs … are reluctant to embrace change”

Manufacturers not being disruptive ready

“The linear thinking and inertia of humans and human systems”

Some people’s inability to fully embrace change of any kind, even if that change is for the better

Examples of a ‘standards mess’ delaying transformation:

Heterogeneus, proprietary, non-inter-operable OT (Operational Technologies) systems

Slow life-cycle of hardware and software: “the slow rate at which these systems are refreshed, typically 15-20 years today”

Examples of other delaying factors:

Confusion between the competing platforms / technologies / protocols / standards

Myths such as supposed lack of clear ROI, lack of data security, gathering of useless data, absence of a track record

Fast product cycles in IIoT (similar to those in consumer electronics) confuse buyers in the industry

From a strategic point of view, it’s really a war between the incumbent technology ecosystem and the new ecosystem, rather than between the technologies themselves.  (See Ron Adner and Rahul Kapoor “Right Tech, Wrong Time” Harvard Business Review, November 2016)

Besides those reasons, there are also micro-level reasons that come up along the business decision process when a specific digital transformation project is proposed within a manufacturing company.

In the end it’s just like any other project: you spend upfront a given amount of money and hope to get a return of investment (ROI).

The trouble is that an innovative Industrial IoT project will be complex and high-risk, and there are several ways it could fail:

  • it proves impossible to build (technology or organization issues);
  • it does not deliver the ROI promise (lower returns or higher costs than expected);
  • it does not get traction (the target users will not use it / buy it);
  • it turns out to be unmaintainable in the long term (workforce turnover, product cycles).

At least a third of all Information Technology (IT) projects fail (See IEEE Spectrum), and probably even more of those of the high-risk type. These rates of failure are unacceptably high for manufacturing, where the average project failure rate is probably less than 10%.

So if the proposed project has a good ROI on paper, how to keep these risks under control and make it a success?

One approach we know from IT is outsourcing most of the trouble to a reliable supplier, and get the solution you need with a software-as-a-service (SaaS) arrangement. Off-the-shelf SaaS is currently booming in IT: see what Adobe, Autodesk, Microsoft, Salesforce and Trello are doing.

With SaaS, there is no upfront cost, and pay-per-use scales linearly with the number of users; typical figures for consumer-oriented services are 5-10 €/per user/per month and for business-oriented services 10-200 €/per user/per month.

So SaaS is a perfect fit for a small enterprise without an IT infrastructure, or for a larger organization that prefers to keep the internal infrastructure slim. It is also suitable for an innovative project, where user acceptance progresses slowly and the numbers may be initially low.

But all of the SaaS vendors above offer off-the-shelf, standard tools – nobody is offering your-own-IIoT-as-a-service yet! For that you need tailor-made SaaS.

Giovanni Battista Moroni “Il Tagliapanni” (The Tailor), circa 1570

Giovanni Battista Moroni “Il Tagliapanni” (The Tailor), circa 1570

With tailor-made SaaS, a solution provider will build the IIoT solution based on the requirements of the manufacturing company, with the agreement that:

  1. the provider will not apply their full margin on the upfront costs;
  2. the provider will keep the intellectual property of the solution;
  3. the manufacturing company will perpetually pay for the use of the solution in a SaaS fashion.

With the pay-per-use model the gain for the solution provider will come later if the project is successful, while the OPEX for the manufacturing company will grow gradually as the solution is deployed and gets traction. This setup can slash the CAPEX for the upfront costs due to the initial effort of adapting, integrating and customizing the chosen platforms / technologies by a factor of 2, while creating a strong commitment for the provider to make the project a success.

Of course it’s a generic framework that can be adapted with any option and variant your legal and financial advisors can imagine. For example, the pay-per-use business models we know from IT SaaS can be creatively adapted to the OT environment by stipulating per-installation/per-hour fees.

The bottom line is that if you find an agreement with a trusted supplier, with tailor-made SaaS you can share the road towards digital transformation with them, and make it happen for real.

About Your Guest Blogger

Paolo Greppi is one of the founders of the Italian startup simevo s.r.l., a supplier of flexible technologies for the simulation of industrial continuous processes and of financial cash flow. Paolo comes with a mixed IT and chemical engineering background and has a vision where an ecosystem of suppliers can speed up the application of innovations in the industrial and financial sectors.

Comments

  1. This is a good article, and it provides food for thought. A common misunderstanding is that, to be successful in the digitized world, it requires expensive IT foremost. Wrong. What you need first is a concept for your applications – whether and to which degree to make your products, your manufacturing processes, and your internal structures smarter. Then, you need a business model – how to make money with your revised applications. Next, you need to do your homework – ensure that your master data make sense, that you adhere to industry standards, and that your solutions are technically compliant. And only then is the time to think about integration, to implement an IIoT, and to spend a lot of money on IT. As Paolo correctly points out, SaaS is an intelligent way to reduce heavy investments into IT and to gain access to a scalable solution that is flexible enough to grow with the company or to follow changes to its business environment.

    • Interesting link. The digital transformation IS INDEED happening here and there: it’s just not happening ON AVERAGE !

      There is an unlimited variety of billing schemes that can be devised; what works in one case may not work in another case.
      But if you want to go in the direction of servitization, the common example being the drill, what you want if “pay per hole”. “Pay per hole” is the same as “pay per use” (USE, not USER): the more you use your drill, the more value you get from the drill, the more you pay. Also “pay per hole” minimizes the upfront cost.

      If you go for “pay per user” i.e. “pay per tag”, that’s the same as “pay per drill”. It basically means you have to pay it all upfront, in this case when you install the sensors on the steam traps. Also fine, but different.

      I assume in the case of Denka Singapore the business case was so convincing that they decided to happily sustain the upfront costs: good ! The core argument of the article is that if convincing the end user to sustain the upfront costs proves impossible, it’s worth trying to propose a “pay-per-use” model.

Leave a Reply

Your email address will not be published. Required fields are marked *