Building a Metric Based Data Driven Business Model - Seven Habits of Highly Effective Chief Data & Analytics Officers

I have seen lots of people try to define what it means to be Data Driven Business, and most have failed to encapsulate it.? Too many companies only focus on the output data, and develop insight from the ‘exhaust data’ of processes aligned to a focus on the ‘output’ metrics associated Revenue, Profit, Free Cash Flow etc.?These measures are obviously critical, but they only tell you how you have done – they are ‘backward looking’ by Design.?

A Data Driven Business is what I have called an Intelligent Business, which I define as

• An organisation that utilises predictive and adaptive insight to dynamically reconfigure itself, in response to the expected needs of its customers, and simultaneously anticipate and respond to changes and events in the external environment.

?The key for me here is that Data is both an Input and an Output for the operating model of the business - and data also describes the Value Drivers of the Business.?Furthermore, you can’t dynamically reconfigure your business unless you have the capability to change processes and ways of working in a truly agile way taking weeks, not months or years!

To consider a business that is truly data (and value) driven, we should consider the case of Amazon.? Their business strategy was famously written on a napkin by Jeff Bezos in 2001, using the ‘Flywheel’ concept from Good to Great.? Jeff created Virtuous Cycle is a strategy that leverages on customer experience to drive traffic to the platform and third-party sellers. That improves the selections of goods, and Amazon further improves its cost structure so it can decrease prices which spins the flywheel. (https://businesschronicler.com/business-strategy/amazon-flywheel-explained).

?The key here was also that Input Metrics driven by working backwards from the Customer Experience were (and are) used to help Amazon drive its business forwards.

1. Better Customer Experience leads to more traffic
2. More traffic attracts more sellers seeking those buyers
3. More sellers leads to wider selection
4. Wider selection enhances customer experience, completing the circle
5. The cycle drives growth, which in turn lowers cost structure
6. Lower costs lead to lower prices, improving customer experience, and the flywheel spins faster

?The leading businesses, look at Input Metrics as well as Output Metrics, after all Data is the fuel of the business and therefore drives the processes that operate the business.? Input Metrics could include Page Views, Stock Availability, Price, Discounts, Convenience.?

?The ‘Intelligent Business’ model is founded on the principles that put People at the top and data at the heart of the business model, with Data mastering Process and technology.? However, when Technology dominates and the purview of the Data, Analytics and AI Leader starts at the ‘exhaust’ of the big Enterprise Systems which are used to ‘run the business’, the importance of the Data Model is often downplayed, but for me it’s a hugely important aspect and is where the rubber hits the road in the relationship between the DAI, IT and the business.

I still cry (often) when people say they can’t measure the benefits of Data investments… Why, because the discipline of Data, Analytics & AI, includes Business Intelligence and Reporting!? In other words, the smart Data Leader is also accountable for providing the key reporting of business performance, so you have the absolute levers at your disposal to measure whether the initiative you’re accountable for has made a difference to the cost, revenue and/or profitability of the business!

??

The Enterprise Data Model – a Tech History

?Traditionally, an Enterprise Data Model (EDM) is an integrated view of the data produced and consumed across an entire organisation. Critically, the EDM unites, formalises, and represents the things important to an organisation, as well as the rules governing them.

An EDM represents a single integrated definition of data, unbiased of any system or application. The EDM is independent of “how” the data is physically sourced, stored, processed, or accessed.

?Despite the criticality of an EDM, Data Models and Data Architecture is often treated with limited value, and skills sit somewhere in IT, and less often in the CDO organisation.

• When SAP, Oracle and Salesforce came to dominate the enterprise systems landscape they all rely on a Data Model to operate.? Typically, in an IT centric world we adopt the ERP or CRM vendor data model.? The vendor will tell us it incorporates an industry specific model, but the reality is that it’s therefore the same model that your nearest competitors also use.
• Integrating systems, is basically a reality when we need to move data from one system to another.? ERP vendors tried to minimise this by insisting you could do everything in their single package or suite.? If your business wasn’t dynamic and/or didn’t change this was logically true, but these days businesses that are static typically go out of business quite quickly.
• Therefore, your company will almost always rely on multiple enterprise systems platforms, and in this case do you want to be beholden to multiple data models or should you have one firm Data Model that acts as the master across the entire business.

?When the CDO only cared about what came out of the many Enterprise Systems, they would get the data dropped (from one or more systems) into a Data Warehouse, where it was normalised, and mapped against a Data Model created for Reporting purposes!

• Oracle were quite happy with this, as they were often the Database vendor and therefore a default provider of Data Warehouse capabilities.? Other ERP vendors, SAP, didn’t like this and created their own Data Warehouses HANA in the case of SAP.

?With the advent of Predictive/Prescriptive and Adaptive Analytics including Machine Learning, the Data, Analytics and AI (DAI) Leader ceased being the person responsible for the ‘exhaust’ data and increasingly provided data and algorithms that became the fuel of the ERP and CRM platforms.?

• the ERP/CRM vendors responded to this trend and tried to repeat their closed loop architecture and flipped the Data Warehouse from being a reporting platform to be the source of Data that drives their ERP/CRM platforms.?
• Again, if your business runs exclusively on SAP, Salesforce (or Netcracker if you’re a Telco) then this is a reasonable approach.? The reality is most companies don’t, can’t and/or won’t!!?? Which brings us back to the importance of your Enterprise Data Model

The future technical operating model is Cloud based, and the agile business moves away from monolithic enterprise systems to micro services, which are connected by APIs to your Data platforms.

• Machine learning and AI models drive data to and from the micro-services via APIs and continually improves performance and accuracy in real-time, in response and anticipation of the needs of your customers.
• Your Data resides in secure data platform(s) and is structured using your Enterprise Data Model
• Your business is able to dynamically (re)configure itself in anticipation and in response to events, and your Data Model, coupled with algorithms is the heart of orchestrating your business!

?Many Architects will tell you that an EDM is a legacy concept, never complete and with today's modern architectures like Data Mesh, you don't need a monolithic model. That might be true, technically, but id you want your Data Model to represent the Value Driver model of the business and actions taken in the business to roll up to the Enterprise KPIs, my advice is to still focus on having a clear EDM!

?

Building a Data Driven Business Model

The new value driven Enterprise Data Model focuses on putting the metrics and hence the data to drive the key decisions for the business at the heart.? The old ‘process-centric’ operating models could be made more efficient with systems and (robotic) process automation, but crucially reinforced ‘the way we used to work’ into tomorrows systems.?

When Michael Hammer launched the Business Reengineering craze of the 1990s he talked about ‘obliterating processes’.? Instead, most chose to merely automate what they had (with some minor changes).? In the world of today, those legacy processes do need to be obliterated and the data driven model will allow us to dynamically reconfigure those processes and technology towards the way we need to work tomorrow and again for the day after!? Of course, there needs to be a business case to implement this!

The Data model should be built by looking at the Value Drivers of the organisation, which allows the creation of a KPI /Metric Model which gives uninterrupted line of sight between your Level 1 Key Performance/Results Indicators (like Revenue, EBITDA, Free Cash Flow) you report to Shareholders, and the underlying operational metrics of the business (Customer Lifetime Value, Cost of Goods sold, Employee Costs etc etc); as well as those critical Sustainability and ESG metrics.? BUT it must also include those critical leading metrics that will roll-up to the Level 1 performance metrics and allow you to drive the business to success

The Data Model can be built around that core performance model and needs to reflect the criticality of both leading (input) and lagging KPIs.?

• Companies that put all their efforts into lagging KPIs (Revenue, Sales, Profit, Free Cash Flow) will always be backward looking (descriptive) by design.?
• Input (leading KPIs) are often found at level 3 or level 4 in a KPI tree but are often at the heart of predicting company performance.? For example, in Consumer, and particularly eCommerce type businesses, the level of investment in ‘performance marketing’, which aims to ensure the right adverts are placed in the right channels at the right time, to drive customer buying behaviour, and is hence a predictor of future Sales.

The Enterprise Data Model provides the overarching framework, and then sub-sets of that framework can be implemented into systems as you upgrade and/or replace them, thereby designing for change and the future.

?

Changing The Way We Work

?This approach institutes a new way of working where Data takes pre-eminence over Process and Technology.? The DAI Leader has a critical role in not just providing insights to the ExCo and Board but increasing advising the CEO in how best to run the business to drive optimum performance. ?In Banks and organisations that have a Chief Operation/Operations Officer, this means the DAI Leader becoming the natural incumbent of that role, with Chief Technology, Information and Process leaders reporting into them.? CDOs/CDAOs need to be able to fully grip Process and Technology to fulfil their role, now they need the true Leadership skills to lead over C-Suite Executives and be a genuine left or right hand to the CEO, with the CFO as their partner in crime.