Current State of affairs
Digital transformation seems to be a buzzword at the moment in business. Everybody is talking about Digital transformation and planning for the big change that Industry 4.0 is bringing. Yet, very few seem to be actively transforming their business to be able to take advantage of the opportunities it brings.
Furthermore, a few of the companies that have started their journeys toward digital transformation have not realised the benefits that they expected.
This can be credited to many different factors. The adoption rate can easily be explained by plotting the Gartner Hype Cycle and the Technology Adoption Lifecycle graphs together as seen below (figure 1). For this article’s purpose, the focus will be only on a few specific technologies. This shows why most of these technologies that directly affect our systems of record, differentiation and innovation have not found mass adoption yet. This is to be expected.
Figure 1: Gartner Hype Cycle vs. Technology Adoption Lifecycle
Looking at the key technologies such as AI and Blockchain and where they are on the Hype Curve compared to where they are on the Technology Adoption Lifecycle it is easy to understand why we have not yet seen mass adoption on most of the hyped technologies. Most of the technologies have not even “bridged the chasm” yet. The only one of the highlighted technologies that has reached the early majority of customers, is Cloud.
The highlighted technologies (which are mostly still in the innovators and early adopters phases) will have a significant impact on the way we use our business software in the next 5 to 10 years.
The first bad experience, will occur when companies and implementation partners do not stick to the five basic principles that should always be considered when investing in any innovation:
Figure 2: The 5 Innovation Principles
The 5 principles being:
- Desirability: does my business or customers want this?
- Feasibility: is the technology mature enough or my business partner skilled enough to deliver this?
- Viability: what is the business value of implementing the new system or change? Strategic or monetary.
- Scalability: how quickly and effortlessly will the solution scale through my entire enterprise to gain maximum benefits and ensure a standardised way of working?
- Sustainability: how long will this solution last before being obsolete or replaced?
How are the technologies currently affecting our business systems?
Cloud offerings are seeing mass adoption at a fast pace.
Figure 3: On-Premise vs. Cloud, The On-Premise ERP Iceberg, SAP Nation LinkedIn, June 2015, https://www.linkedin.com/pulse/on-premise-erp-iceberg-ben-stewart/
Customers are seeing a much lower TCO (Total Cost of Ownership) as well as solutions that keep them ahead of the curve through updates/upgrades at the expense of customisation. The customisation gap left by cloud solutions are however being bridged by custom Apps built on innovation platforms. In my opinion, this is a very welcomed change.
IoT (Internet of Things)
IoT on its own does not provide much value. What you do with that data coming from IoT is what matters. Combining IoT with AI, RPA (Robotic process automation), Blockchain and advanced analytics is where businesses will generate the most value.
Quite a lot can be said around this topic. For the purpose of this article I will stick to ERP (Enterprise Resource Planning) and planning systems.
In some cases, we are already realising great benefits by using machine learning algorithms in forecasting. Achieving increased accuracy (Average of 27% FC accuracy increase) by finding correlation between the sales history and external data like weather conditions, sales prices, social media sentiment and promotions to name a few.
This technique can be used across almost any industry from fashion, food & beverage, utilities to manufacturing and spare parts planning.
This raises a question around DDMRP (Demand Driven MRP) vs Multi-Echelon-Inventory-Optimisation (MEIO). If the system can automate my forecasting and dramatically improve my forecast accuracy, is there a place for DDMRP?
There are two schools of thought on this topic:
First, getting further away from our reliance of forecasts in the short to mid-term (DDMRP)
DDMRP is easy to use but not scientifically sound or optimal, therefore has limitations. Also, very labour intensive to manage due to buffer level adjustments and tweaks to the planning parameters. These buffer adjustments however, can, and certainly will, also be automated through AI by automatically adjusting the buffer levels.
Personally, I am a big fan of DDMRP even though I believe there is still room for improvement in some areas such as the “Red zone” buffer calculations as well as the resource capacity calculations.
Second, using AI Algorithms like Gradient Boosting to get more accurate forecasts combined with MEIO (Multi-Echelon-Inventory-Optimisation).
This would strive towards, but will never achieve, the perfect state of 100% forecast accuracy (= zero safety stock required for demand variability). The question here is, where is the break-even point for forecast accuracy for your company where the results of MEIO would be better than DDMRP?
The bottom line is that you would want to be able to choose which works best for your business. You would also want to be able to simulate what that breakeven point is for your business. Some of the software packages already allow you to simulate the scenarios and compare both and choose the right option to achieve the right outcome.
Clustering will also play a big part in better understanding your business.
The application of AI in ERP and planning systems will continue to grow. The application opportunities are seemingly endless.
Advanced AI - Self-Regulating Adaptive models
We are also approaching self-regulating-adaptive-models.
Firstly, this implies that the system can learn (Adapt) using traditional machine learning techniques (Supervised, Unsupervised and Reinforcement).
In addition to the ability to Learn/Adapt the system should also be able to self-regulate. To self-regulate, the system would need to be able to adjust its learning algorithms incrementally without human intervention.
Currently, there are different forms of adaptive models available which provide us with decision making support.
One application of this is, automatic system suggestions for exceptions. This system can automatically identify exceptions, in some cases predictively, and automatically present different suggestions to resolve the problem.
Figure 4: Example of an Adaptive System Providing Decision Making Support
E.g. A manufacturing sites IoT sensors detect a vibration on the production line, a predictive maintenance algorithm informs the production manager that the line will be performing at 60% efficiency for the next three weeks unless the line is stopped for maintenance. The system’s adaptive learning engine knows that there are three different solutions to this exception/event. Stop the line and perform maintenance, missing the customer order. Continue running the line at 60% efficiency with a 43% change of a complete breakdown. Lastly, to sub-contract the orders and make a smaller margin on the sale. The system predicts the outcomes of each scenario and ranks them accordingly. The production manager selects his/her preferred option and approves it.
The part that is not currently in use is the self-regulation.
The development of a self-regulating adaptive system eventually drives towards a touchless planning system. This means that it is a realistic assumption to make that in the next 5-10 years we will see a near-fully-automated ERP system, where 80% of the core functionality is automated. I believe this will include decision making capability as well.
RPA (Robotic Process automation)
Chat-Bots are already used to automate system maintenance activities, support activities, act as personal assistants, automated communications and analytic support or similar tasks.
I believe that it is not long before RPA is combined with AI to automate the core of business process freeing humans to perform more creative tasks.
Digital twins enhance the decision-making process as well as live monitoring and predictive maintenance.
I believe we will see this being adopted more to model the entire business – like the physical warehouses, docking bays and manufacturing facility, and not only key assets as it is generally being used for at the moment, to simulate the business plans developed in the planning system to ensure that the plan can actually be executed physically without any issues.
Blockchain has been described as the “trust revolution” – this rings true for both finance as well as within supply chain. Enough has been said and written about blockchain and its applications in finance. From a supply chain point of view, blockchain will also help with traceability and visibility of the supply network. Collaboration between your suppliers and customers as well as direct transactions between suppliers and customers. Simplifying the process as well as providing cost savings.
It will also be used to provide the end-user of products transparency of the product’s journey through the product value chain.
E.g. Tracking and recording the eggs journey, handlers, temperature, expiration date, sustainable practice and quality From the farm to the 3rd party logistics, to the processing plant, to the packing process, to the retailer DC, to the store and finally to the consumer. The end-user and every party along the value chain would be able to have full transparency of the value chain.
Most of the major software vendors have similar development roadmaps. They are all moving toward cloud solutions. All are focusing on quick time-to-value for the customers. All are moving slowly toward automation and ultimately touchless systems. Cloud solutions mean less customisation for the customer.
I believe the key differentiators will be industry specific focus/solutions as well as the platform they provide their partner network to innovate on. The better the innovation platform’s functionality and ease of use, the better the partner solutions that will be developed.
I also believe that the commercial models will play a big part in the software vendor’s success going forward. This applies to the major vendor’s partner-commercial-model as well. When given a choice between two software vendors with similar offerings or chance of success, the partners will opt to promote the software vendor with the best partner-commercial-model.
Consulting Partners will continue to see lower income per implementation project with the cloud solutions. They will also see lower support income per implementation due to the cloud solutions’ nature. This is not all doom and gloom though as, there will be more projects due to the reduced price-tag.
With the ERP software products using more and more advanced technologies like AI and RPA, the implementation partners will have to invest in these specialised skills.
With cloud solutions providing the core solution and most of the custom requirements and innovation projects being delivered on technology/innovation platforms, the software vendors rely on their consulting partner network to develop next-gen solutions. This gives the consulting partners an opportunity to build solutions that can be offered as a service to generate more scalable revenue.
Consulting Partners will also have highly specialised employees. These employees are generally not cost effective to employ by the Customers. The further technology evolves, the more specialised resources would be required to set up and manage the technology solutions. This means that it will also become more financially viable and compelling for Customers to outsource some of their business tasks to consulting partners. E.g. Supply-chain-planning-as-a-service.
For Customers, the challenge will be to navigate their digitisation journey. It is not a task many will be able to do successfully on their own. The key to success will be to find a trusted and knowledgeable partner to advise them with their decision-making process.
Also, as previously mentioned, I believe more customers will start outsourcing business tasks that are not core to their business such as, forecasting and supply chain optimisation.
As Heraclitus of Ephesus said, “the only constant is change”. I believe it is up to the strong leaders to carry the torch. You have a choice – lead or be led.
If you would like to know more about Supply Chain Planning, Supply Chain Optimisation and Forecasting or you’d like to view a live demo, please feel free to contact me, Ruan.VanVuuren@olivehorse.com, I would be delighted to discuss your needs and specific requirements. I do hope you enjoyed this blog.