Building Your Foundation for Digital Transformation

Digital Transformation

A Comprehensive Guide for Manufacturers

Table of Contents

Chapter 1:
Planning for Success

Aligning your capabilities and goals to ensure success

As mentioned in the introduction, a unified data platform is the foundation for enabling a wide range of digital objectives. How prepared and equipped are you to start building and applying this foundation?

There are two aspects of readiness to consider:

1. Organization

Do you have the commitment, organizational buy-in, budget, and skill sets for your digital initiative to succeed?

2. Technology

What level of supporting infrastructure do you already have in place? Are your critical production machines outfitted with sensors? If so, are they networked? What information is being captured? And — most importantly — is your data accessible, ready, and fit to enable successful digital projects?

Choosing objectives you’re not ready for is a recipe for project failure.

Why Readiness Matters

Manufacturers often start out by attempting to solve a use case they’re not ready for. In midstream, they find out that the data they have isn’t relevant to the problem, and the inputs they do need are inaccessible. Or they don’t possess the staff expertise, management buy-in, or budget — and sometimes all three — to build the necessary technical capabilities.

Alternatively, should a one-off pilot actually get off the ground, most often it doesn’t scale across multiple machines — either because the required additional data sets are not available, or because adding them in breaks the logic and data models that were targeted against the original use case. The net result is a series of dead-end pilot projects, a scenario so commonplace that it has a name: pilot purgatory.

Any of these situations can doom a project from the start. The solution is to pursue both technical and organizational readiness upfront. That’s what the next two steps are about.

Choosing objectives you’re not ready for is a recipe for project failure.

Chapter 2:
Organizational Readiness

What are the key success factors in terms of funding, management backing, staffing, and governance?

Technical issues usually receive primary emphasis, but organizational capabilities are equally critical for success. A digital initiative requires changes to established processes and traditional lines of communication in manufacturing enterprises, and new ones that bridge across teams and chains of command.

We’ve seen projects run out of steam or get mired in cycles of endless piloting because key organizational success factors have not been dealt with.

These [dynamics] fall into three key areas:

Commitment and Budget

Digital projects frequently fail because they don’t win broad support beyond the originating team, function, or individual. It’s worth putting in the time upfront to develop and articulate a vision for where the enterprise is going, and to explain how a unified data platform and its initial use cases support the long-term roadmap. Set your sights on obtaining:

  • Management buy-in to give data-driven decision making urgent priority and an allocated budget
  • Project backing from corporate, plant, and IT leadership, and operational teams as well
  • Alignment on objectives among these functions and factions

Identify and cultivate project champions and evangelists within different departments and echelons — including management of production, IT, and corporate functions; factory-level teams who will use the insights digital provides; and data architects and scientists.

Engage all departments and teams during planning. Roll-out is too late in the game to try to recruit broad-based support. If you wait till then, you’ll encounter resistance.

Skills and Resourcing

Laying the groundwork for a unified data platform requires a blend of capabilities from groups that aren’t used to collaborating, and in fact may never have done so before. Key human resources include:

  • Manufacturing experts who understand the available data and physical sensor structure
  • IT staff who can connect sensors to networks and data storage, and develop applications for the plant floor
  • Data scientists and data engineers with the knowhow to extract and interpret insights using analytics tools

These teams typically live in their own specialized worlds. To develop a full picture of your digital requirements and how to assemble them, these diverse and specialized groups need to be brought together in the same room to contribute their particular skills and expertise.

Working in concert, this cross-functional team can undertake a full audit of your existing production data and processes. OT can specify the problems they need to solve and the benefits they’re looking to achieve. IT in turn can identify relevant data. Collectively, the full team can determine if these inputs are available and if not, how to obtain them.

Change Management

Sometimes project leaders will invest in IT systems and services but neglect to fund the operational side. This is why cross-functional governance is essential to managing your digital effort. Common goals need to be set by IT and OT leadership in terms of staffing, resources, and changes to processes and workflows.

Existing metrics, incentives, and procedures need to be re-evaluated in light of the new capabilities a unified data platform will deliver.

Put in the time to develop and articulate a long-term digital vision, and obtain upfront support from corporate, plant, and IT leadership.

Chapter 3:
Technical Readiness

What aspects of technical infrastructure do we need to assess?

Very few manufacturers have all their technical requirements in place at the outset of a digital initiative. It’s important to get an accurate picture of what you’ve already got and what you’ll need to build or buy. Machine connectivity is the primary must-have but there are other technical considerations as well.

Connectivity and Accessibility

The three areas to evaluate are:

Connectivity and Accessibility

At the foundational level, ascertain whether data is being collected from sensors on your key production machines and is flowing into a system of record. Also explore whether you’re measuring and gathering the right kind of data to address your operational goals.

  • Are all machines equipped with sensors?
  • Are sensors network-accessible?
  • Is data streaming into a system of record or database?
  • Are you capturing key attributes such as machine downtime, defect data, and part serial numbers or batch numbers with time stamps?

Cloud and Security

Cloud and Security

With huge data volumes and massive computing power the norm for AI-driven manufacturing analytics, the cloud is far and away the most practical, affordable venue. How prepared are you to migrate production data to cloud-based infrastructure? Security is a key concern, especially in manufacturing sectors governed by regulations requiring special treatment for sensitive information such as classified, ITAR, or HIPAA data.

  • Are your system of record (i.e. historian) and other key data sources (ERP, MES, MRO, supplier data) securely accessible from outside the company?
  • Do you have policy guidelines for working with cloud providers?
  • Are data segregation requirements clearly defined and do you have a system for cordoning off sensitive information?

Data Awareness

Data Awareness

The project inception stage is not too early to assess your capabilities for working with the data you generate. Do you have the in-house expertise to interpret data streams from machine sensors and how they map to your physical production process? Much of this knowhow will come from blending the skills and understanding of IT with those of your operational engineers. The two sides must work in synergy to define project goals and develop processes to achieve them. You’ll also likely need to source advanced data analysis capabilities through consultants and by creating and nurturing a data-focused specialty within IT.

These requirements trace back to the premise of this tutorial: adopting a “data first” digital strategy to collect and unify your entire universe of production-related data to address a wide range of use cases… rather than pursuing a succession of disjointed one-off pilots that don’t scale to multiple objectives and plants.

Ensure success by focusing on organizational readiness as described earlier. Build a solid foundation by articulating a long-term digital vision, validating how a unified data platform enables it, and securing early and strong support from leadership across corporate, operational, and IT domains.

A clear picture of what you are — and aren’t — ready for will make the difference between a smooth project lift-off and getting nowhere fast.

Chapter 4:
Assessing Your Data: A Deep Dive

A detailed review of what it takes to make data accessible and usable for digital transformation

Data readiness is a unique challenge in the manufacturing industry, because production-related data arrives from a multitude of sources and systems, is stored in diverse networks, and comes in a myriad of different formats and structures.

Data readiness

Manufacturers often embark on projects in the mistaken belief that their data is accessible, usable, and relevant, only to discover that most of it is not even close to being ready to support a digital initiative.

The time to ascertain the condition of your data, and take proper steps, is before your project begins. Over the course of numerous engagements, Sight Machine has developed a process for assessing production data against the key attributes required for successful digital manufacturing initiatives.

In this 3-part video tutorial, an expert data engineer walks you through the 3 stages of data assessment, providing illuminating explanations and guidance throughout.

The key to project success is understanding your ability to access and use production data for real-time analysis.

Chapter 5:
Evaluating Use Cases

Evaluating and choosing the right projects for our level of digital readiness

Now that you understand what data readiness is about, what would you like to use your data for? What production-related improvements are you eager to make? Which uses cases, issues, or problems do you want to apply digital techniques to solve?

Source: McKinsey

Source: McKinsey

As you can see above, digital capabilities open up a universe of possibilities, once you have them in place… that last phrase being the key. A common misstep manufacturers make is pursuing projects they’re not yet ready for. Another problematic scenario is having various groups pursue diverse use cases, each using different data. For digital projects to scale — and for techniques like AI to be brought to bear — it takes a single source of data that digitally represents your full production process. This is the all-important advantage of building a unified data platform.

At the outset, align the digital goals you’d like to achieve with what you’re ready for. Assess your current capabilities, the elements you can build out in the near term, and the kind of use cases you can support across functional teams so that everyone moves forward in unison.

Exactly How Ready Are You?

From our experience with numerous manufacturers, we’ve identified 5 categories of readiness, with different use cases feasible for each.

Organizational Readiness

  1. Connection Zone. At this stage, you have sensors on some production machines but not all the important ones, and are not yet capturing all relevant data. Some or all of your data streams are not networked, and there is no central repository. Cloud and data security strategies have yet to be defined.
  2. Visibility Zone. Here, most of your machines have sensors and are networked. IT leadership is on board and providing support and resources. You have operational people who understand how sensor data maps to production processes. Some collection processes remain to be automated, and data aggregation and harmonization capabilities need to be deployed. Change management expertise does not yet exist at the plant level.
  3. Efficiency Zone. Companies at this stage are capable of taking on projects that deliver significant operational and profitability impact. The technical connectivity is in place to capture machine downtime, defect, and batch/part serial numbers data, and information is flowing to a system of record. While plant leadership is committed to harnessing data to drive operational execution, analytics and application development resources are not in place at the plant level.
  4. Advanced Analytics Zone. Data scientists and engineers are available to extract and interpret insights using analytic tools, and IT staff has the knowhow to develop custom applications for the factory. Business process change resources are on hand but there is only limited support for multi-plant and supplier transformation.
  5. Transformation. At this most advanced stage, comprehensive technical excellence is present across the extended organization, including suppliers. Corporate, plant, and IT leadership are committed and aligned for transformation. Factory-level and corporate change management functions — including communications, business process, training, and analysts — are established and aligned.

Determine your Readiness Zone

Determine your Readiness Zone

Our Digital Readiness Index tool enables you to assess your stage of preparedness, facility by facility. Take the questionnaire, and the tool will use your responses to calculate your readiness zone.

Once you’ve ascertained what you’re ready for, take a look at these zone-specific suggested use cases.

Quick-Win Digital Projects for Each Zone

Connection Ready:

The first order of business is to get data flowing: connecting key sensors to the network and capturing all the data produced. At this level, plant personnel need to increase their understanding of what the data means, mapping out what each sensor describes about machine and product conditions at each step in the manufacturing process.

Use cases:
  • Build offline data acumen
  • Develop foundational data visibility
  • Automate data capture
  • Create a digital twin

Visibility Ready

Visibility Ready:

At this level, it’s important to aggregate data into a global view of real-time production operations, so performance can be monitored and controlled to reduce scrap rates.

Use cases:
  • Create a global operations view
  • Access plant information
  • Monitor OEE and KPIs
  • Compare KPIs across the enterprise
  • Benchmark plant, line, and machine performance
  • Measure and trend quality and performance
  • Monitor out-of-control processes and failures
  • Automate collection of regulatory production data
  • Set up alerting and notification of production challenges
  • Monitor actual output vs. planned
  • Improve machine performance
  • Statistical process control
  • Parameter relationships
Efficiency Ready:

Manufacturers at this readiness stage are able to move beyond visibility and apply analytics to improve efficiency and quality. You can increase profitability by addressing stubborn scrap and quality problems, and by optimizing processes to boost productivity.

Use cases:
  • Track machine performance using the OEE metric
  • Set up part traceability
  • Perform high-level defect analysis
  • Drive and measure continuous improvement efforts
  • Analyze out-of-control processes and failures
  • Measure and increase capacity utilization
  • Analyze and reduce scrap
  • Determine root causes
  • Increase first-pass yield
Advanced Analytics Ready:

At this level you have the technical and organizational readiness to deploy projects with a high impact on operations and profitability. Apply predictive analysis to deliver advanced notification of impending downtime or defects, and use advanced statistical techniques (such as multivariate regression, factor analysis, decision trees, and clustering) to identify root causes of persistent problems. Potential benefits include increased production, reduced cycle times, and lower scrap rates.

Use cases:
  • Deploy advanced statistical techniques to identify root causes
  • Predict out-of-control processes and failures
  • Implement track and trace
  • Monitor and optimize energy use in production
  • Perform predictive analysis
  • Implement extensible analytics
Transformation Ready:

Companies with these sophisticated capabilities are ready to utilize analytics learnings to develop transformational business models such as capacity-based pricing. Analysis can also be extended across supply chains, using third-party data to transform relationships with suppliers and customers.

Use cases:
  • Establish cross-system analysis
  • Perform machine/line level costing analysis
  • Set up data-driven design to value
  • Create supplier optimization programs
  • Develop real-time capacity-based pricing

Choose and Vet Candidate Use Cases

Choose and Vet Candidate Use Cases

Knowing your readiness zone doesn’t mean you’re limited to only those projects we suggest. However, the zone attributes help you understand the prerequisites to work on — understanding that’s critical for establishing budget and timeline expectations.

Choose one or more of these or similar use cases that line up with your readiness zone. Discuss them with your teams to see which inspire the greatest support and promise to deliver the highest value.

In the next section we’ll provide guidance in prioritizing your selected projects: determining and comparing the business value of each, and developing a roadmap for implementation.

It’s critical to choose use cases that align with both your readiness zone and your overall goals.

Chapter 6:
Project Prioritization

How do we decide which projects to proceed with, and in what order?

Now that you’ve chosen some candidate digital projects, the next task is to decide which to take on first.

Tech Readiness

We suggest evaluating them according to these three criteria:

1. Tech Readiness

It’s worth reiterating that success will be elusive if you select projects that greatly exceed the technical foundation you have in place. The timeframe, resources, and investment required for preparation may likely be too great to sustain, and the initiative can collapse under its own weight well before implementation. Or else critical technical gaps may only become evident during rollout, which is too late. Hopefully, the previous steps of this tutorial will have helped you avoid any such scenarios.


2. Momentum

In the beginning, we highly recommend choosing use cases that will generate real value in a short timeframe. Every organization has doubters, and we find that when initial efforts move slowly, skeptics only become more skeptical. Select early-stage projects for rapid deployment. Quick wins strengthen confidence, shore up the initiative, and build sponsorship and support for the substantial investments you’ll need later on.

Financial impact

3. Financial impact

Sight MachineTo know which projects will deliver the most bang for the buck — and how soon — you have to quantify the impact of manufacturing production improvements on corporate profitability. Plant managers naturally focus on productivity, but conventional productivity metrics don’t readily convert to dollars. What’s required is a universal and consistent method to do that.

Sight Machine has developed a productivity metric, the Manufacturing Performance Index (MPI), that lets you translate performance improvements into output gains and dollars generated.

We recommend using the MPI, and our tool for calculating it, to evaluate and compare the profitability impacts of a range of possible digital projects. Our white paper will show you how.

In the early stages of your initiative, build momentum and reinforce buy-in by choosing use cases that produce clear wins.

Chapter 7:
What to Buy vs. Build

What is the most effective way to combine market solutions with our own in-house capabilities?

We recommend utilizing the power of the market for underlying digital capabilities you don’t already have and can’t develop quickly in-house. It’s critical to assess these realistically. Bear in mind that manufacturing data comes in huge volumes from thousands of sources and in diverse formats, and needs to be harmonized and integrated into usable form. Because of these unique data complexities, packaged solutions created for multiple industries — that is, not purpose- built for manufacturing — are of nominal usefulness.

manufacturing analytics solution

Let’s examine what’s available in the market for the essential components of an end-to-end manufacturing analytics solution:

Machine /IoT Data Acquisition

The marketplace offers capable tools from multiple vendors.

Data Preparation and Blending

Commercial solutions are available but must be evaluated carefully, as some are not a good fit for the dynamics of manufacturing data, as explained above.

Contextualization/Data Modeling of Production Interdependencies and Relationships (creating a digital twin of your entire manufacturing process)

To our knowledge, the market offers a single manufacturing-specific solution designed to handle the job comprehensively and at scale, powered by AI, machine learning, and sophisticated analytics techniques. This platform was built and refined over a period of years, drawing on highly specialized software development capabilities that would be impractical for most manufacturers to acquire in-house.

Analytics Algorithms

This is the domain in which to leverage your in-house expertise and understanding of your own production dynamics. Building on the data modeling platform alluded to above, your in-house team can integrate custom analytics and algorithms targeted against your specific objectives. This unique blend of unified data platform plus internal knowhow will deliver rapid, unprecedented value and impact.

Analytics Algorithms


As with data acquisition, the market offers a selection of capable solutions in this area.

Operationalization/Integrating Output into Workflows

Here too, a number of serviceable packages are available for these functions.

Take advantage of manufacturing-specific market innovations that are flexible enough to seamlessly incorporate your in-house expertise.

Chapter 8:
Selecting Solution Providers

What are the essential capabilities to look for in solutions and vendors?

As we mentioned in the previous step, the market offers a number of solutions for data collection from production sensors and other instrumentation. Well-known names in Industrial IoT are represented, such as IBM, SAP, PTC/ThingWorx, Oracle, and others. These products do a good job of aggregating data from a variety of sources. However, that’s as far as they go.

None of these IoT leaders offer a scalable platform for the all-important second phase of manufacturing analytics: contextualizing and modeling the relationships between data and production processes. So while it’s great to have a data lake, if you don’t also have an engine to turn all that information into actionable insight, you’re not gaining much business value.

Selecting Solution Providers

To build a unified data platform for manufacturing analytics, it’s necessary to explore beyond big names and basic data acquisition products.

Here are some key questions to consider when evaluating solutions and vendors:

  • Is the platform comprehensive enough to factor in multiple types of manufacturing processes and machines? (Even continuous process manufacturers have discrete process needs for boxes or other containers. You need an analytics platform that can handle that as well.)
  • Can the platform incorporate real-time streaming data?
  • Does it scale to a full range of use cases, enterprise-level production volumes, and multiple plants?
  • Is the platform designed to create integrated data models from thousands of inputs and multiple data types?
  • Does the solution readily integrate with data acquisition, analytics, and visualization tools you may already have?
  • Are data and analytics output exportable so they can be utilized elsewhere?
  • Does the platform offer workflows, alerting, and integration with production-related systems?
  • Does it provide visibility and viewing tools relevant to all levels of the organization, so everyone from operations engineers to supervisors and corporate leaders can draw on the same source of truth?
  • Is the solution scalable and software-centric? Or is it a services-led approach that requires repeated consulting engagements every time you need something done?
  • Does the platform come with self-service tools that let your engineers build their own applications on top of the system foundation?
  • Does the vendor have configuration and support specialists who understand the relationships between specific data inputs and production processes?

Getting Started with Industry 4.0 and IIoT More and more companies and CIO’s  have started thinking about getting started with Industry 4.0. But the big  question is how do I get started? What is really needed for me to get started?  Read our 3 tips to get started with an industry 4.0 project for your company  Read More

Look beyond well-known names and multiple-industry products for capabilities and expertise focused solely on manufacturing analytics at scale.

Authors and

Find out about the people behind the conception and design of this book

Meet the people behind this book

If you’re interested in speaking with any of our team about digital transformation, get in touch!

Picture of Sudhir Arni

Sudhir Arni

VP of Mfg. Transformation

Sudhir Arni​ is Sight Machine’s VP of Manufacturing Transformation. Prior to joining Sight Machine, Sudhir was an engagement manager at McKinsey & Co., where he designed and led manufacturing transformation programs for pharmaceutical and chemical manufacturers. He received joint MBA and Master of Science degrees from the Kellogg School of Management and McCormick School of Engineering at Northwestern University.

Picture of Beth Crane

Beth Crane

SVP, Delivery and Transformation

Beth Crane, PhD is Vice President of Data for Sight Machine, the category leader in manufacturing analytics. In this role she focuses on helping manufacturers understand how advanced analytical techniques can solve complex problems in production and operations.

Prior to Sight Machine, Beth has worked in both academia and industry and has led the development of analytical and reporting tools used for continuous process improvement.

She received her PhD and Masters of Science degrees from the University of Michigan and was awarded a National Science Foundation postdoctoral fellowship to explore the development of statistical methods for predicting dysfunction in multi-dimensional time series data.

Picture of Kanna Potharlanka

Kanna Potharlanka

VP, Digital Transformation

Kanna Potharlanka is VP of Digital Transformation at Sight Machine. As part of his role, Kanna leads our large enterprise customer engagements at both strategic and tactical levels in enabling their transformation journey and realizing growth opportunities. Previously, Kanna spent 8 years at Apple and led the iPhone Planning and Sustaining Operations organizations. As part of the core leadership for iPhone Operations, Kanna played a key role in scaling and supporting the steep growth iPhone experienced (revenue grew from $40B to > $200B).

Prior to Apple, Kanna was one of the first engineers at Sierra Logic, a networking startup that went through a successful exit in 5 years. Kanna earned his M.B.A. from the Kellogg School of Management at Northwestern University, and holds an M.S. degree from University of California, Davis, and a B.E. (Honors) degree from Osmania University, India.

Picture of Jon Sobel

Jon Sobel

CEO & Co-Founder

Jon has served on the management teams of several companies in pioneering industries, including Tesla Motors, SourceForge, and in its early years, Yahoo! Jon holds an BA from Princeton, a JD from the University of Michigan, and an MBA from Wharton.

Picture of Nate Oostendorp

Nate Oostendorp

Founder & CTO

Nathan Oostendorp is the CTO of Sight Machine, he co-founder the company in 2011. Nathan started his career as a controls engineer at Donnelly Corporation (now Magna Mirror) where he worked on PLC programming, computer vision, data acquisition, and robotics for a major automotive supplier.

In 1996 he co-founded, a major tech news blog which was the center of the Linux and Open Source Software movement. During this period he spun off several other successful open online communities including (an early precursor to Wikipedia) and, the central hub for the Perl programming Language. He also created the first Open Source advertising and analytics platform. He then joined as the site architect and ushered it through a period of growth where it became a top 100 website globally, and hosted several hundred thousand software projects.

He holds a BS in Computer Science from Hope College in Holland Michigan, and an MSI in Information Science from the University of Michigan.

Picture of Matt Smith

Matt Smith

Sr. Vice President, Digital Transformation

Matt Smith is the Sr. Vice President of Digital Transformation at Sight Machine.

Curious about how we can help? 
Schedule a chat about your data and transformation needs.