Publications February 11th 2019

10 steps to transition from an intuitive strategy to a predictive strategy

In corporations, traditional practices tend to favour experience, business conventions, uses and habits. Sometimes even political reflexes at the expense of effective decision-making. An analytical culture challenges these established mechanisms by being more pragmatic and qualifying what has never been qualified before. This creates a culture shock between old and new ways by challenging established ideas, expert opinions and old habits, which can be unsettling, to say the least, for certain managers. But the results are well worth it. A systematic data-driven approach can lead to new areas of innovation and significantly improve performance.

Ten steps are key for transforming an intuitive strategy into a predictive strategy.

Step 1 - Clarifying decision-making processes

Launching a data-driven project does not mean throwing the baby out with the bathwater! It requires drawing up a detailed inventory of beliefs, good and bad decision-making practices, and flawed processes but also an inventory of your resources and skills.

This step is all about preparing the ground to ensure that is fertile. All problems must be listed and analysed to identify the exact need, clarify objectives, implement the right strategy and check that sufficient resources are available.

Step 2 - Aligning expectations with the business strategy

‘Start with the why’; this golden rule focuses attention on defining needs and finding consensus in terms of the company’s key areas e.g. improving profitability, rationalising investments, etc. The business strategy must inform each step in order to avoid only focusing on technology or data. Goals and resources, but also the business lines and processes, must be in synch to ensure your project remains meaningful.

Step 3 - Understanding the past thanks to a sound data base

Understanding the past helps us anticipate the future. It also helps us identify growth drivers and factors of success and failure, as well as focusing and optimising our efforts.

This step involves creating a database by optimising data collection and standardising in accordance with the objective: processing structured/unstructured data, data that is scattered throughout the company or only available in disparate formats, etc. Undeniably a thankless task but one of major importance in terms of guaranteeing the accuracy of the analyses and the soundness of recommendations, as well as minimising errors. The key to success? The relevancy of the selected path and managing all stakeholders’ expectations.

Step 4 - Identifying levers of performance

Explanatory data helps us understand ‘what to do’ and ‘how to do it’ e.g. the customer journey, growth drivers, measuring impact, etc. The Marketing Mix Modelling (MMM) approach, employed by many CAC40 and Fortune 500 marketing execs, is the perfect response. This approach makes it possible to measure the impact of marketing investments, isolate the ROI, quantify intuitions for the first time, etc.

Step 5 - Continuous learning and optimisation

Data is only a decision-making tool. It has no value unless it is accompanied by flexibility, creativity and human intelligence. Test & Learn, iterations, learning loops, etc., it is important to establish methods and tools that can be adapted to different contexts, to varying customer demands and constraints, to extend an offer or develop new markets, etc. The ability to adapt over time.

Step 6 - Being in the game thanks to predictive models

Prospective data relies on the past to better understand the future. Using advanced statistical methods, machine learning and AI, we can predict our customers’ or teams’ needs, and thus offer the right product at the right time.

In terms of prediction, data is totally impartial. It can be used to confirm or refute assumptions, quantify risks, gives us figures for new areas, and combat feeling of resignation (‘we'll never know’). While some talk about storytelling, we are talking about a proactive mechanism for storymaking.

Step 7 - Constructing a clear and shareable measurement ecosystem

The race to have the latest technological solution can be counter-productive as it can produce results that are contradictory, illegible and out of context. The multitude of tools and platforms can result in managers relying on their intuition as they struggle to make sense of the results. This is not helpful. Avoid the error of a ‘Tools First’ approach: step back and take a holistic view that is always focused on the customer.

This rule also applies to measurement systems. Each department has its own set of performance indicators that - very often - can be hard to compare. Working in isolation does not encourage the creation of a common vocabulary or links between business units. Everyone remains in their ivory tower. Therefore, it is necessary to create a global and standardised ecosystem to facilitate how performance is understood thus creating virtuous collaborations.

Step 8 - Having an organisational approach that improves decision-making

Shake things up! Creating a data-driven culture requires teams and individuals to be agile. This paradigm shift often upsets organisations making it harder to implement. Yet there is a very human dimension in a data-driven approach as it tends to reinvent business lines, processes and bring teams closer together whether they like it or not (e.g. IT and marketing), tackle political blockages (recovering data from several departments), and coming up with new language elements (e.g. between online and offline).

Step 9 - Using technology as a facilitator

Once we have finished identifying the use and business goal then tools become very powerful allies. However, if you prioritise tools above the business goal, you risk being disappointed. A tool is not a silver bullet. The human dimension, support from the business lines, and the company’s vision are key to making the right decisions.

Step 10 - Enhancing your intuition with data analysis

Real value is created by synergies between data and intuition. It feels risky switching to autopilot in favour of a 100% instinctive approach or a 100% data approach. So why not implement a classic 80/20 approach i.e. 80% for data-driven decision-making leaving 20% for creativity?

Today’s managers must invent new decision-making mechanisms, and encourage the creation of winning combos; questioning historical hypotheses via exploratory analyses, and analysing weak signals are key to creating an innovative and enlightened business approach.

To sum up, enlightened and effective decisions rely on a combined intuitive and data-driven approach. But the Holy Grail will never be found without data literacy. Perfectly combining various practices, the prospective approach - unlike explanatory approaches - is already proof of significant data maturity. Organisations must take a step-by-step approach to learning this.