Generating data insight: it’s a (supply) chain reaction (#1 Data Maturity Series)

Photo by Mitchell Luo on Unsplash

It’s not the sort of statistic you’ll ever see in a sales deck, but the dirty secret of the data business is that most projects fail.

In this new series of blogs, I want to take a look at where organisations go wrong and try to provide a better model for achieving the business outcomes they expect.

We rarely acknowledge it, but the uncomfortable truth is that our industry still only delivers a fraction of what it promises. And a pretty small fraction – some estimates suggest less than 15 percent of big data projects deliver business value.


 To begin with, we need to think of data-driven outcomes as a supply chain.

A chain is only as strong as its weakest link

Producing meaningful business outcomes depends on correct execution at a number of stages or links along the chain, and a failure at any will result in the failure of the project as a whole.

Yet many businesses, probably the majority, still view data as a discrete entity, something that can be assigned under a single function; most commonly I.T., but sometimes marketing, finance or a particular business unit.

 This often results in things functioning well for a while in one part of the chain, perhaps giving a false impression of success. But then nothing moves beyond it – no actual business decisions are ever taken or actioned – because of a breakage in the other links.

The ultimate result is that projects are eventually marginalised or even abandoned without ever producing any worthwhile outcomes.

The data supply chain model

Data literally needs to flow right through an organisation in order to generate those unexpected, bottom-line insights that make it all worthwhile, the unknown unknowns buried deep in the numbers that would not otherwise have surfaced.

The supply chain concept can be summarised as follows: 

  • Capture – Data first needs to be produced, captured and stored, and is generally done by the I.T. team.
  • Organise – The raw data needs to be organised and enriched, transforming the raw data into Information, which is generally stored in a data warehouse.
  • Analysis – The information is then explored and analysed by analysts and data scientists who create Insights.
  • Decisions – Business management then need to make data-based Decisions and avoid the dreaded HiPPO scenario, where gut-instinct overrules any evidence presented by the data.
  • Actions – Finally, the business need to be able to Execute the decisions in an efficient and timely manner.

 It is critical to note that real value is created only if the business is able to action and make use of these insights.

An organisation may build high-performance capability in their analytics teams, but if the decision-making process in the business does not take advantage of this, then the ultimate impact is minimal. Alternatively, if the analytics teams do not have sufficient data or processing capabilities available to them, then they will be hamstrung in their capacity to deliver meaningful insights.

Scaling data excellence

 Even when all of the elements are put together, it generally starts off with specific use-cases that do not cover the breadth of the organisation. For example, an organisation may successfully execute various marketing initiatives, but there are many other opportunities in logistics, pricing, merchandising, etc, that remain unexplored.

The next step here is to scale that initial success throughout the rest of the organisation.

Often, these initial projects are driven by the head of a function, and scaling that excellence will usually require more senior sponsorship – someone who can align the other functions to adopt the now proven practices. At best, scaling leads to analytics and data-driven decision-making transforming the organisation; at worst, the lack of scaling leads to the business focusing on suboptimal outcomes, where outputs are optimised for a function, but not the organisation as a whole.

Furthermore, it may be that business processes may need to be redesigned, or functions restructured, and this is another reason why sponsorship at the most senior, C-suite level is so important (as we’ve alluded to in previous blogs).

In my future posts, I’ll be exploring the topic of data maturity and a model by which organisations can use to progress in their journey.

Popular posts like this

Offense and Defense

Data governance: on the offensive (#2 Data Maturity Series)

Server Side Tracking – What’s Old is New Again

What is data quality? Part 1