the inSights Value Chain

"Why Analytics-driven Insight Initiatives succeed or fail in your business"

The intent of writing this Article is to share with people, a framework that I have developed over many years and across hundreds of initiatives to help ensure that Analytics-driven Insight Initiatives/projects are successfully implemented in Commercial Organisations.

For almost 20 years, I have built, advised, and overseen the delivery of hundreds of Commercial Insights projects using Data & Analytics. In that time I have worked with many companies, built, managed, and operated Data, Analytics & Insights practices. I have also worked in private equity-owned businesses (where it is all about measured value creation in a rapid cycle), built and deployed AI & Machine Learning enabled products, and now run my own Insights company. Across that journey, I have appreciated both successes as well as the odd failure, and it is through these experiences that I provide my perspective on where I have seen value created and value eroded through Analytics-driven Insight projects in commercial organisations.


The Promise of Value

While many organizations have made investments in adopting different forms of Advanced Analytics, few commercial organisations (to-date) have seen quantifiable Value Creation, ROI or the promised levels of a performance step change from Analytics-driven Insights. In this article I will talk extensively about Commercial Value Creation, so, I want to be very clear in my definition of commercial value:

"Commercial value creation is a quantifiable value defined in commercial terms (i.e. profit, revenue, margin, cost, risk, multiple) that is directly attributable to the outputs of your Analytics-driven Insight initiative"

The reality is that driving business outcomes in a traditional organisation is difficult. Many organisations have and continue to struggle to drive enhanced decision-making and optimised business outcomes from analytics-driven insight.

A number of years ago, in the spirit of continuous improvement, I started doing full-360 reviews of all of my projects (with all project stakeholders), to understand why some projects were wildly successful and why some were abject failures. I have continued this process for many years and this now includes insight from a number of different organisations and many different initiatives. Today, I continue to work with different organisations and a selection of leaders in the Analytics and Insights space to broaden my understanding of value creation through analytics-driven insight.

Summarising my Insights

The reality is that failure to deliver a business outcome, in most cases has little to do with technical capabilities or the quality of modelling, but a lot to do with right-sizing scope, understanding business needs, working effectively with a portfolio of business stakeholders (including IT) and a better capability to facilitate, support and measure change. I have seen simple dashboards create seven-figure returns and I have seven figure-projects simply fail!

Through my analysis and continued research, the common themes around why initiatives fail to achieve significant value creation include:

  • The problem statements are too vague and value propositions are unclear, teams are not really clear on what problem they are solving or what defines a successful business outcome

  • The Initiatives outputs never get sustainably deployed, i.e. never reach Production so sustained insight to support key decisions is not achieved. Most analytics-driven insights have a required level of currency & relevance, they need to be continually scored to be useful.

  • The Initiatives never get implemented into business processes. Teams don't apply enough consideration to managing process change beyond the model's output i.e. How would a decision-maker actually use these outputs sustainably and consistently in their existing roles? see my article on 4 questions most data scientists can't answer

  • The outputs of the initiative don't make decision-impact. Often the outputs (i.e. a scored label) of the model are embedded at the start of a long chain of human decision making, which means the insight from the model is often lost in translation and doesn't materially impact the critical decision-making points. In this case, the output adds little to the business outcome.

  • The Initiatives don't get appropriately measured. Teams don't robustly and scientifically measure the incremental impact attributable to the initiative. Many teams are not able to transparently link value creation to implementation of their initiative.

  • The Initiatives just aren't needed or wanted by the business in its current strategic position, under its current business model, or in its existing decision making maturity. If a change to the status-quo (i.e. the way we do things today) is not an executive priority, then the impetus to change behaviour within the business is very limited. However, when change is defined as a strategic priority and that impetus comes from the top-down, change happens!

  • Swinging for the fences, not hitting singles - Many projects in the Analytics space are just too ambitious in the intended business outcome they strive to achieve. While models can be built to provide insight into a multitude of decisions, the corresponding level of augmentation to the business process (beyond the model) is often more than the business is willing/able to accept and too often Change management is decoupled from the project and seen as somebody else's responsibility.

What I learned is that driving business outcomes through Analytics-driven Insights in a Commercial organisation is much like a value chain. Like any Value Chain, it has a sequence of dependencies and failure to comply upstream causes impact and value erosion downstream (See my article on Broken Links in the Insights Value Chain).

It is through this value chain that we deliver a finished product (Insight) that is consumed by an end-user (decision maker) in a way that creates benefit for the consumer (better decision making) and creates value for the company (commercial value creation).


I have spent the large part of my career trying to form a perspective on how to navigate this complex value chain from end-to-end. To summarise my efforts I present the Insights Value Chain and Insight Management Methodology. I am sharing what has worked successfully for me & my teams over the years.

What I present here is the actual process that I use today and have used to significant effect in commercial organizations to deliver real, measured benefits. I hope that by sharing this, people can borrow bits and pieces that may help you on your journey to value creation.


The Insights Value Chain

"Achieving Insight-to-Business Fit"

The objectives of the Insight Value Chain is to ensure that Analytics-driven Insight projects/initiatives are:

  • Fit-for-the-business - Aligned to Strategic Priorities, Principle Risk or Competitive Threats of the business

  • Fit-for-purpose - Implementable and actionable in the business

  • Measured in commercial terms - have clearly defined commercial outcomes

  • Sustainably Deployed - They reach Production and can be monitored and maintained as required

  • Integrated into a decision-making process - The insights are integrated into decision-chains in a way that they impact real business outcomes

  • Used by audiences for decision support - They become seen as essential decision-support tools to decision-makers

  • Directly attributable to real business outcomes - we can attribute value creation directly to the initiative.

To help achieve these objectives, we introduce the Insights Value Chain (IVC). The IVC consists of 8 core domains, and each domain plays a vital role in ensuring a successful outcome. Through the IVC, we ensure initiatives are coordinated, supported, strategically aligned, implementable, get deployed, used, measured, enhance internal decision making, and ultimately deliver quantifiable value creation in clearly defined commercial terms.

Call-Out: One of the call-outs I want to make about this diagram is the importance of the red-green colour coding. Red refers to the cost accumulation of your initiative, and Green refers to commercial value creation. It is not until our outputs are deployed and used that we start creating value.

It is essential to understand that the lifecycle of the Insights Value Chain takes a single initiative from end-to-end.

The Domains of the Insights Value Chain include:

  • Domain 1: Align – In the Align Domain, we ensure that any proposed initiative is strategically, tactically, and operationally aligned and that we frame the Value Proposition in commercial terms. The Align Domain is designed to ensure that your initiative is fit-for-the-business, has a clear definition of value, and that key stakeholders are aligned, supportive, and aware of their responsibilities/accountabilities.

What is passed from Domain 1: Align to Domain 2:Define are Aligned Value Propositions

  • Domain 2: Define – In the Define Domain, your Value Proposition is abstracted to a level of detail such that we can understand if it is implementable within the existing business model. We evaluate the value proposition so that any resulting Value-Creating Initiatives have supported arguments, supporting analyses (as required to support arguments) and assessment of available data, people, technology, and processes.

What is passed from Domain 2:Define to Domain 3:Plan & Design is a Value-Creating Initiative ready for Planning & Design

  • Domain 3: Plan & Design – In the Plan & Design Domain, your proposed Value-Creating Initiative goes through a rigorous planning and design process. We start by mapping decision chains and performing data discovery. We then ideate on analytical approaches and consider interfaces. We work with IT to define the required infrastructure for both production & consumption. We also perform an initial definition of value measurement. We then undertake project planning for execution, ensuring we have a fit-for-purpose plan. While Machine Learning is inherently iterative and exploratory, these projects can't have open-ended timeframes. One of the most important things you need to do at this stage is to understand and define fit-for-purpose-model-performance. We need a target level of accuracy to guide how long we should spend on model building.

What gets passed from Domain 3: Plan & Design to Domain 4: Develop, is a robust plan of attack for building a fit-for-the-business, fit-for-purpose, measured initiatives within commercially palatable time frames.

  • Domain 4: Develop – In the Develop Domain, it is all about laser focus on execution. Our Agile for Analytics approach looks to ensure that the iterative/exploratory nature of the model building is coordinated and aligned with the supporting and dependant processes. Supporting and dependant processes include data engineering, process design, domain & subject matter input, infrastructure design & planning, testing, and measurement.

What gets passed from Domain 4 to Domain 5 are MVP's that get tested for value creation, usability and performance

What gets passed from Domain 5 back to Domain 4 are refinements for further development.

  • Domain 5: Test & Refine – The Test Domain is an iterative Domain with the Develop Domain. Each initiative will likely cycle through several Develop and Test phases. Learnings and feedback drive refinements that will enhance value creation, usability, and achieve required performance levels. What we are testing in this domain are three specific things. The first is the ability of the initiative to achieve value creation as defined in commercial terms (often controlled testing). The second test determines if the output usable by the intended audience for the decisions they need to make. The third test looks at two components, can IT support this at the scale and how will we maintain and control the performance of the model over time.

What gets passed from Domain 5 to Domain 6 is an initiative that has achieved its target value, usability, and performance thresholds.

  • Domain 6: Deploy – In the Deploy Domain, we look to deploy our tested and refined initiative as the final solution. We implement optimized processes, train users, operationalize data & analytic pipelines, release interfaces & insight consumption points, and turn on value & performance measurement. We must think of deployment as far more than periodic model scoring. Deployment is the integration of your outputs into decision-chains in a way that sustainably impacts business outcomes.

  • Domain 6(b): Agents of Measurement: You will see a box on the diagram called the Agents of Measurement. I am quite emphatic about the importance of scientifically robust measurement in organizations, not just for the Analytics initiatives that we deploy, but more broadly for the Key Performance Indicators and investments of the organization. Years ago, when I set up my first analytic practice in a commercial organization, I established the "Culture of Measurement," and the purpose was to provide an independent and scientifically robust measurement of performance. Not only was this critical in communicating the benefits of our team's initiatives, but it became a vital source of new analytically identified value-creating initiatives for the organization. I believe that the panacea of an analytically driven organization will be born from a "culture of measurement,". When we are analytical in our thinking of performance, we can start to understand what creates value and kills values. We can then use that insight to drive new initiatives and projects.

What gets passed from Domain 6 to Domain 7 is continuous measurement of value, usage, and performance.

  • Domain 7: Improve – In the Improve Domain, we leverage measured performance metrics to understand how to improve the business outcome further analytically. When we analytically measure performance, we begin to understand drivers of growth, decline, risk, and more, and it through these insights that we prioritize future enhancement & development of the deployed solution.

At the highest level, the domains provide a foundational overview. However, to help practitioners get from end-to-end, we also offer a detailed approach for each step, the methodology we use for getting from start to finish of the Insights Value Chain is called Insights Management.

Insight Management

"End-to-end value creation across the Insights Value Chain"

Layered on top of the Insight Value Chain, we apply a methodology for helping us navigate our way through each Domain. I have spent many years refining this over many iterations. The structure has been used successfully to deliver millions of dollars in commercial value across different organizations.

I refer to the process of getting from the start to finish of the Insights Value Chain as Insight Management.

Over the course of this series, I will share insight into each of the Domains of the Insight Management methodology in the hope that it will help Insights professionals think more expansively about how to drive value creation through Analytics & Insights within their organizations.

Next Post: Achieving Strategic Alignment

If you have questions or would like to know more then, please contact me at or visit