The Challenge
In the contemporary manufacturing landscape, bridging the gaps between the stages of a product lifecycle is both a strategic priority and a formidable challenge. Our client, an international manufacturing giant, confronted significant hurdles in linking design, manufacturing, and service data. The disparate systems across these stages caused inefficiencies, data silos, and redundancies, hindering their ability to respond rapidly to market demands and innovate effectively.
The client needed an integrated solution that offered visibility and synchronisation across the complex web of product lifecycle stages. This required a robust digital infrastructure capable of assimilating data from various sources into a coherent, actionable dataset that could inform decisions at each point of the product's life, from the conceptual blueprint to the after-sale support.
The Solution
We at Adyantrix crafted a comprehensive Digital Thread solution tailored to transform the client's data landscape into an interconnected ecosystem. Our implementation strategy involved three critical phases:
Phase 1: Assessment and Architecture Design
Our team conducted an extensive analysis of the existing systems and workflows. This included stakeholder interviews and a roadmap design that highlighted key touchpoints in the lifecycle requiring digital thread integration. A cutting-edge architecture was developed to ensure scalability and flexibility, anchoring on microservices and cloud-based technologies to facilitate seamless data flow.
Phase 2: Implementation
A phased deployment approach was implemented where initial integrations focused on high-priority areas with rapid ROI potential. We utilised advanced API management solutions to integrate legacy systems with modern platforms, ensuring real-time data synchronicity. For example, as part of the design phase enhancements, CAD tools were directly linked to the manufacturing floor, enabling instant updates and feedback loops.
Phase 3: Integration and Optimisation
Leveraging AI and machine learning, we enhanced data visibility and predictive analytics capabilities throughout the lifecycle. Service data was incorporated to improve predictive maintenance and customer support functions, ultimately impacting the product quality and customer satisfaction. Regular workshops were held to refine processes and drive organisational adoption of the new system.
Key Results
The implementation of the digital thread solution yielded remarkable outcomes. Foremost, cycle times for design to manufacturing adjustment experienced a reduction of 30%, significantly diminishing time to market for new products. Moreover, the interdepartmental data silos were virtually eliminated, which boosted collaborative efforts and innovation across teams.
Service efficiency saw a 20% enhancement due to improved predictive maintenance scheduling derived from integrated analysis of service data. As a result, customer satisfaction scores improved by 15%, reflecting higher-quality products and more responsive service processes.
The holistic integration also equipped the client with powerful insights, leading to data-driven decision-making capabilities across all levels of the organisation. This not only enhanced operational effectiveness but opened new avenues for strategic product innovations.
Ultimately, the Digital Thread strategy enabled our client to transform its product lifecycle management into a seamless, continuous drone of innovation and efficiency, moving closer to the vision of a truly smart factory.
Technical Approach
The digital thread architecture we built for this client was designed around a canonical data model hosted in an Azure-based integration platform, with Azure API Management serving as the central gateway through which all lifecycle systems exchanged data. The canonical model defined a consistent schema for product identity, revision state, and attribute data that every connected system—regardless of its native data structure—was required to conform to when publishing or consuming lifecycle information.
On the design side, the client's primary CAD environment was PTC Creo, with product structure and configurations managed in Windchill PLM. We developed a bidirectional integration between Windchill and the canonical data layer using PTC's REST API extensions, ensuring that every approved design release was immediately reflected in the integration platform and available to downstream manufacturing and service systems without manual data entry or file transfer.
The manufacturing execution layer used SAP S/4HANA for production order management, interfaced through SAP's OData services. A key technical challenge was mapping PTC's part-and-revision identification scheme to SAP's material master structure, which used a different versioning convention: we developed a translation service hosted as an Azure Function that normalised revision identifiers in real time, eliminating the manual reconciliation that had previously consumed several person-days per product release.
For the service data layer, IoT telemetry from field-deployed products was ingested through Azure IoT Hub, processed by an Azure Stream Analytics job that applied product model and serial number lookups against the canonical data store, and surfaced through a Power BI dashboard consumed by both the service engineering team and the design team. This closed the feedback loop from operational product data back to the design environment—a connection that had not previously existed.
Implementation Highlights
We structured the implementation to deliver value incrementally rather than waiting for a full-system cutover. The first integration deployed was the Windchill-to-SAP design release pipeline, chosen because it had the highest frequency of manual intervention under the legacy process and the most clearly quantifiable time saving. Within six weeks of going live, this integration alone eliminated an average of 14 manual data entry events per working day, each of which had previously required a production planner to manually transcribe part numbers, revision codes, and material attributes from a Windchill export report into the SAP material master.
The most technically complex phase was the IoT service data integration. The client's installed base of field products spanned five product generations with different telemetry protocols: three generations used proprietary serial communication formats, one used MQTT, and the most recent used OPC-UA. We developed protocol adapters for each format, normalising all telemetry streams into a unified JSON schema before ingestion into Azure IoT Hub. Developing and testing the adapters for the proprietary protocols required the embedded firmware team to provide protocol documentation that had not been formally maintained—in two cases, the documentation did not exist and had to be reverse-engineered from firmware source code.
Organisational change management was as demanding as the technical integration work. The design engineering team had historically operated with complete ownership of product data up to the point of design release, after which responsibility transferred to manufacturing and service functions. The digital thread created shared visibility of product data across all three functions simultaneously, which initially generated tension: manufacturing planners could now see draft design revisions before they were released, and service engineers could flag field failure patterns directly to design teams without going through a formal quality process. We facilitated a series of cross-functional workshops to establish governance protocols for how different categories of digital thread data would be accessed, acted upon, and escalated, converting what had initially felt like a threat to functional boundaries into a shared asset.
Measurable Outcomes
The 30% reduction in design-to-manufacturing cycle time was measured across 47 new product variant introductions in the twelve months following go-live, compared to the same metric across the preceding twelve months. The previous average cycle time from approved design release to first production unit was 11.3 working days; post-implementation, this fell to 7.9 working days. The primary driver was the elimination of manual data re-entry and the associated correction cycles that had previously delayed production order creation.
The 20% improvement in service efficiency was most apparent in the reduction of mean time to diagnose (MTTD) for field failures. Before the digital thread, service engineers investigating a field failure had to request design documentation from the PLM team, typically waiting one to two working days for the relevant drawings and specifications to be located and shared. With direct access to the canonical product data layer, service engineers retrieved the same information in under three minutes. Across a service team handling an average of 340 field events per month, the cumulative time saving was substantial.
The integration also surfaced three design-quality insights within the first six months that would not have been visible without the service data feedback loop. One was a correlation between a specific assembly sequence parameter in the manufacturing execution data and an elevated rate of early-life field failures in one product family—a correlation that the quality engineering team had suspected but could not previously quantify. The digital thread made the correlation visible and statistically verifiable, leading to a process change that reduced early-life failure rates for that product family by 38%.
Lessons Learned
The most important lesson from this implementation was that the canonical data model must be defined before any system integration work begins, and it must be defined collaboratively with representatives from all lifecycle functions. On this project, the initial canonical model was drafted primarily by the IT architecture team and later required two rounds of revision when the manufacturing and service functions identified attributes that were critical to their operations but had been omitted or structured incorrectly. Each revision required retrospective changes to already-deployed integration adapters. A more thorough upfront data modelling workshop would have avoided this rework.
The second lesson concerned protocol diversity in IoT integrations. The effort required to develop and validate adapters for legacy communication protocols was underestimated in the initial project plan. For future engagements involving IoT-enabled product fleets, we now recommend a formal protocol audit as a discrete early-phase deliverable, with adapter development effort estimated separately for each protocol variant rather than being treated as a single line item.
Finally, the governance frameworks established for cross-functional data access proved to be as important as the technical integrations. Without clear protocols for how draft design data could be used by manufacturing planners, and how service failure data could be escalated to the design team, the increased visibility created by the digital thread would have generated more organisational friction than value.
Why This Approach Worked
The digital thread succeeded because it was grounded in an architecture principle that prioritised the canonical data model over point-to-point integrations. Many manufacturing organisations that attempt to connect lifecycle systems build direct integrations between pairs of systems, resulting in a fragile web of bespoke connections that is difficult to maintain and nearly impossible to extend. By routing all data exchange through a canonical layer with a single, governed schema, every new system that joins the ecosystem requires only one integration—to the canonical layer—rather than N integrations to every existing connected system.
The incremental deployment strategy was equally important. By delivering measurable value with the first integration before the architecture was complete, we maintained organisational momentum and senior stakeholder support through what is inevitably a long and complex implementation. Manufacturing technology programmes that attempt to deliver everything in a single big-bang cutover frequently encounter the combination of delayed value realisation and accumulated organisational fatigue that causes programmes to stall or be descoped. The phased approach ensured that the digital thread was operational and generating returns before the most technically complex integrations were tackled.
Speak with our Data Engineering team at Adyantrix to find out how we can support your next project.
Work with Adyantrix
If you are looking to tackle a similar challenge, Adyantrix has the expertise to help across the full project lifecycle. Our data engineering practice covers pipeline design, streaming, and data infrastructure. Our data analytics practice covers BI reporting and self-serve analytics platforms. Our IT consulting practice covers technology strategy, architecture review, and digital transformation advisory. Our cloud & DevOps practice covers cloud infrastructure, CI/CD, and platform engineering. Get in touch to discuss your requirements — no commitment required.
