Agile Integration: Part Two--Characteristics of a Purpose-built Integration Platform for Supply Chain Planning and Optimization; Roadmap to Agile Integration
on Mar 14, 2017
We explore what to look for in an integration platform to achieve rapid integration for supply chain planning and optimization systems. We also describe a potential roadmap to agile, incremental integration and value realization.
Full Article Below -
( This article is excerpted from the complimentary report Agile Integration: Accelerating Time-to-Value in Supply Chain Implementations,
available for download here. )
In part one of this series, we looked at the challenges in integrating supply chain planning and execution to existing ERP and other systems. This is almost always the long pole in any implementation project, and hence, each month of reduction in integration time usually yields a month reduction in the overall project. Here in part two, we describe what to look for in an integration platform that has been purpose-built for supply chain planning and optimization. We also present one possible roadmap to agile integration and incremental value realization.
Preparing for the IoT Data Tidal Wave
With sensors, intelligence, and connectivity built into more and more of our physical objects and surroundings, the Internet-of-Things has the potential to provide new high value streams of data to supply chain planning and optimization engines. It also has the potential to generate overwhelmingly high volumes of data. Here are examples of potential IoT applications in supply chain:
Predictive maintenance—Predictions of future failure can be key inputs for forecasting and optimizing scheduling of skilled technicians, spare parts inventory, and the required logistics.
Consumption monitoring—Machines can provide predictive data on when consumables will need to be replenished (e.g. soap for washing machines, syrups for soda machines, oil or brake pads for cars).
Machine usage—Forecasting when each individual customer is likely to need a new machine or system, based on their current usage, could be an input to longer-term forecasting and planning.
Supply-side visibility—Data from sensors on the supplier’s plant floor and dock doors, on trucks and ships, in yards and warehouses, ports and airports can be combined with weather, traffic, and other data to provide much more accurate lead times and precise ETA, feeding S&OP planning, production planning, and risk management systems.
Integrated retail store sensor data—Can provide insights into consumer behavior, influencing merchandising and planogram development, as well as promotions and forecasts.
Cold chain temperature exposure data—Can be used to optimize distribution, enabling a First-Expired First-Out strategy.
These are just the tip of the proverbial iceberg. To be ‘future-proof,’ supply chain planning solutions need to be ready to deal with the variety, volumes, and real-time streaming of all these new IoT data.
What to Look for in an Integration Platform
The desired attributes of an integration platform depend on the intended purpose. The requirements for a general-purpose integration platform, intended to integrate any system to any other system, are different than the requirements for integration that is specific to a supply chain planning and optimization system.
Purpose-built Integration for Supply Chain Planning and Optimization
Thus, not all integration platforms are created equal. For example, virtually all integration platforms have adapters for each of the major enterprise systems (SAP, Oracle, etc.) and many also have some sort of templated mapping to bring that data into their internal canonical data model. However, these tend to be a mile wide and inch deep. In contrast, a best-of-breed, purpose-built integration platform for supply chain has depth in its specific domain. Here are some of the key characteristics to look for from an integration platform for supply chain planning and optimization:
Proven purpose-built templates—This means the templates have real depth of mapping, incorporating all the types of master data that are important to supply chain planning and optimization. Ideally the level of depth, maturity, and completeness of mappings evolves and increases over time, as the solution provider encounters more situations, accommodates new data sources, and adjusts to changes in existing data sources. Thus, it is important that the solution provider actively invests in the templates on an ongoing basis. This is part of what makes integrations sustainable over time.
Flexible, future-proofing, no-coding—No pre-built mapping can be 100% complete for all circumstances. Thus, it is critical that the integration platform provides the flexibility to adapt and evolve. Practically speaking, attributes to look for include:
No-coding/visual approach to mapping
Rules-based, easy to customize
Self-documenting and easy to understand and maintain by newcomers who did not create the integration
Does not break upon upgrades on either side of the integration
Transformation logic that is independent from mapping logic
Robust data validation.
Figure 1 – Solution Attributes That Address Various Challenges
Willingness and capability to take responsibility for integration—If the solution provider is willing to take total responsibility for doing the ‘grunt work’ and making the solution work, it helps take the pressure off of constrained internal resources. It also demonstrates that the solution provider has skin in the game.
Ongoing management of integration post go-live—A vendor with strong purpose-built integration has a clear advantage in building out capabilities at scale for managing, supporting, and maintaining integrations after the initial go-live events. Taking on this responsibility enriches the vendor’s R&D efforts by forcing them to further improve the solution, as they continually learn, fix issues, and broaden the use cases and systems they cover. Be wary of providers that don’t offer to take responsibility to manage the integration once the system has gone live. Their lack of commitment may be a warning sign of the difficulties of maintaining and evolving the system as things change.
Price-certainty/time-certainty—Look for a solution provider that has enough confidence to offer a fixed price and guarantees for the timeframe of the integration.
Automation of supply chain master data management—The solution should include tools that can crawl through your ERP’s item master and other master data, and automatically drill down on multi-level BOMs to find all the elements, correlate elements between the systems, rationalize naming conventions between different systems, identify missing data, and so forth. This is another aspect of ‘purpose-built,’ as this kind of automation requires an understanding of the semantics and syntax of the supply chain data needed to feed planning and optimization engines.
Data validation—The system should be able to identify missing data, data that is out-of-range, and provide dynamic, rules-based reasonableness testing logic. This could include various user-editable formulas or look-up tables to ensure the data is valid. It should include mechanisms for logging and alerting of data quality issues as soon as they occur, so they can be identified, diagnosed, and fixed quickly, as they arise, rather than after-the-fact when the damage is already done.
Delta-change data loading—If the integration platform reloads the entire database from scratch each time, it will cause unacceptable performance and inability to get rapid data updates. The system should have the capability to determine what data has changed in the source systems and then upload only the changed data.
Unstructured data handling—Increasingly intelligence is being extracted from unstructured data, such as social media, email, news feeds, IoT data, and so forth. Machine learning and AI are getting better and better at seeing the patterns and finding meaning and useable information from unstructured data. The method of integration depends on where this machine learning resides. If there is an external analytic engine ingesting and making sense of all this data, it may generate specific signals and insights that are then fed into the supply chain planning and optimization system. On the other hand, if the machine learning analytics are built into the planning and optimization system, then it requires the capability to ingest high volumes of streaming data and unstructured data. Makers of planning and optimization systems are currently figuring out what architecture works best for them. It is a good idea to find out how they plan to handle unstructured and streaming data.
These guidelines should help you select a supply chain planning and optimization system that has the right integration capabilities. This will be a key enabler to make rapid, agile, future-proofed integration possible.
An Incremental Roadmap to Value
An agile implementation approach drives value realization in incremental bite-sized steps. In this spirit, here is a possible sequence of steps to realizing incremental value:
Problem-focused initial integration—Using a supply chain solution with a purpose-built integration platform, focus on a single use case, finding the minimum data set needed to solve it. Don’t skimp on data validation, quality, or completeness. Often the key constraint is finding the existing data that has the quality and completeness required. Therefore, data quality and completeness should be part of the initial use case feasibility assessment and selection process. If you discover the data is just not ready, it may be smarter to move to a different use case for which the right quality data is available. Flexibility in which use case to tackle first can allow you to pick the true low-hanging fruit, enabling higher confidence of a quick first win.
Use case/data expansion—Building upon the initial foundation you’ve built, and the organizational momentum from initial success, further use case can be implemented in succession. In addition, work can proceed on integrating additional data into existing use cases, to increase their value. This often includes further cleaning up and enhancing the data.
Build business agility and performance—As more and more use cases are implemented, these new capabilities can be used to improve the supply network, add new products, acquire new companies, and run the business more effectively.
Integrated business planning—Adopt Integrated Business Planning (IBP) to get to the next level of performance. Once you have the right, relevant, timely data all in one place, potential supply-demand issues can be spotted early enough to actually do something meaningful about them. Then you can do interactive ‘what-if’ modeling to deal with changing business conditions, and come up with the best end-to-end plan, taking into account everything from end-customer demand, to the logistics, and supply constraints.
IoT and unstructured data—Look for opportunities to experiment with various use cases that leverage smart, connected machine data, as well as internal and external unstructured data. An agile, fail-fast approach to experimentation is good at this stage, to try out and discover what works and doesn’t, and learn as you go.
With the right purpose-built, supply chain-focused integration tools, the implementation of supply chain planning and optimization can be done very rapidly (within weeks or months). Rapid time-to-value builds excitement and pride with the team, and momentum and support from upper management to fund further projects. Picking the right platform and using the right agile implementation approach can make the difference between getting bogged down in a painful, repeatedly delayed project vs. being the hero that brings the company to the next level of business performance.
To view other articles from this issue of the brief, click here.