Predictive marketing is a science that aims to convert buyers’ interests into revenue by supporting companies to engage with them with greater personalization. It anticipates customers’ behaviors and needs through statistic data-based projections, such as predictive machine learning or AI models. It also sends personalized communications, at the right time, to the right person, with tailored products and recommendations. Data is the cornerstone of this science.

Fortunately, the emergence of new technologies and new actors has contributed to significantly increase availability of customer data at firms’ level. As a matter of fact, 60% of companies even consider that they collect too much data from too many sources [1] which is one of the biggest hurdles preventing them from learning from customer data and acting accordingly.

However, with the support of predictive marketing, companies can now do both in order to get the most out of acquired data. That is why 91% of top marketers are either fully committed to or are already implementing predictive marketing [2].

Nevertheless, in an increasingly complex marketing and data ecosystem, made up of new technologies, with strong customer expectations and overwhelming information... how can companies unleash the full power of predictive marketing?

After several years of experimentation, we are now able to identify the key success factors of predictive marketing involving 3 crucial conditions that companies struggle to meet:

  1. An adapted organization,
  2. A human-centered project,
  3. A strong project management.


In order to unleash the full potential of predictive marketing, the key step – as in most data science projects – is to set up the right ecosystem with the right data partners and define a strong, agile and replicable project framework.

The premise is to have enough reliable quality data to feed and run this project. Once the data is available, specific skills are required to retrieve, process, analyze and present it through data engineering, data science, data analysis and data visualization.

As in all projects, a predictive marketing project has to clearly define the RACI for each step and set a strong framework with clear deliverables at the end of each of the following steps: data collection, data quality audit, exploratory analysis, data-driven marketing plan, offer definition and reporting.


Today, the agile methodology for project management is fundamental, notably in its test and learn approach. While there may be a fad behind this notion, using it for predictive marketing is completely relevant. Indeed, it facilitates to:

  1. Set up the framework:
    • Define the vision/goal of the project,
    • Highlight customer needs addressed by the marketing messages,
    • Outline the project’s success,
  2. Empower a lean marketing team:
    • Pick a team and set them free,
    • Ensure cross-functional / end-to-end delivery capabilities,
    • Ensure top-management sponsorship to remove impediments,
  3. Start agile operations and scale:
    • Have a transparent backlog and prioritize continuously and ruthlessly,
    • Conduct the first pilots with an iterative and always-on marketing approach,
    • Test and validate.

An agile mindset to involve all stakeholders, to ensure the completion of deliverables, and to build a strong foundation in order to industrialize and automate the project will maximize its chances of success.

It’s worth noting that there is no one-size-fits-all model for data science projects. Indeed, if a company thrives on a model, another may flourish on an entirely different one. Every company should find the right ecosystem with the right balance between automation, industrialization and personalization to meet its objectives, taking into consideration time, budget and resources.


Beyond the organizational aspect of the project, the most important is the human one. Since it introduces a new way of thinking – data-driven – within the organization, the management must present the project as human-centered: it is not a program which will decide what is best to send, but rather a recommendation for the people.

For example, the decision to send out communications will not solely be based on the marketing campaign calendar, but also on data-based recommendations (e.g. customer journey, propensity score) for a seamless and unique customer experience.

As a consequence, companies should involve from the very beginning all stakeholders taking part to the data-driven project development where algorithms support human decision-making processes.

A strong executive sponsor should evangelize the role of data within all departments and at all levels of the organization by:

  • Eliminating organizational silos: directly linked to the organization’s ability to store data for a complete overview of the customer information,
  • Popularizing data mechanisms among management and operational teams,
  • Communicating on the project’s output.

Once companies have spread out data culture, they need to set up a change management strategy to educate the organization and adopt data-driven projects. Doing so will foster acceptance and commitment as well as reduce resistance to change. Besides, it will improve the existing program, enable new ideas and facilitate future data initiatives from each department.

As a matter of fact, companies that are successful in predictive marketing projects are, most of the time, those who invest internally on developing their staff’s skills instead of fully relying on external providers. Indeed, they develop required skills through recruitment –  data scientist, data architect – and employee trainings – data analysis, data visualization.


As mentioned previously, predictive marketing results into campaigns that anticipate customers’ behavior and/or forthcoming needs. Predictions are based on scores – probability of success – calculated from empirical data and analyzed by marketing teams to develop pre-established campaign scenarios. Marketing actions are then triggered automatically by events, so-called "triggers".

With 76% of consumers expecting companies to understand their needs and expectations [3], exploiting these data sets offers a considerable competitive advantage.However, achieving hyper personalization, the holy grail of customer activation, requires a strong project management and willingness to change.

Traditional project management involves producing realistic project plans, budgets, estimating time and effort while keeping the work organized and the stakeholders informed. But for a successful predictive marketing project, management capabilities should go beyond, especially on the hyper personalization aspect:

  1. Personalization is a complex field with a gap between expectations and the ability to translate them in tools, actions, processes... Thus, management needs to ensure business and technical teams work closely to identify and built the use cases,
  2. Although predictive marketing is a science, it is an imprecise one. To make it sharper, organizations need to test and learn from the campaign scenarios in an iterative approach (ideally in an agile method),
  3. Data is the foundation of the project. It is then key to fully exploit it:
    1. With 48% of data stored on a multitude of different systems [4], ensure organizational – especially technical – silos elimination to prevent systems unable to speak to one another,
    2. With 56% of companies unable to process the data fast enough to act on time [4], make the data quickly available to act accordingly for a more relevant and effective project.

Secure the hyper personalization part with a strong project management for a successful outcome is an observation that is backed-up by experience. Indeed, one of our clients, a car manufacturer who used to perform up to 4 static and manual campaigns a year, is now reaping the benefits of setting up the first pilot and is planning to deploy to other countries [5]:

  • up to 200 000 automatic campaign variation by month,
  • +1M additional turnover in value,
  • +4% marketing activation.

However, even if companies believe in personalization and see its benefits (improve customer retention, increase revenue per visit, lower churn rate …), 61% of them are not willing to invest the required resources [6]. The lack of financial and human resources puts additional pressure on teams in charge of predictive marketing projects. Furthermore, at some point, it may jeopardize the company’s success, with results less thrilling and programs unable to reach their full potential.

In this context, it is key to strongly manage the project with the available and dedicated resources to maximize its chances of success. The road to deep personalization should be considered by companies as a long-term vision and investment instead of a short-term one.

To unleash its power, predictive marketing requires the right talents, culture, processes and the company’s continuous support to educate the organization. Thus, regardless of maturity level, these recommendations will allow to lay the foundations for an optimized and successful predictive marketing project. Marketers will be able to achieve a new level of growth by using the power of data.

These projects will in time to reshape companies, helping them evolve in their processes and their culture, towards a more « customer centric » organization.


Frédéric Gigant is Partner in the Digital & Strategy team from Paris' office.

Chakir El Messaoudi is Manager in the Digital & Strategy team from Paris' office.


[1] Adobe Symposium Paris 2018



[4] Adobe, 2018, Tout est question de contexte

[5] After 9-month period (from March to December) for one country

[6] Dynamic Yield, 2019, Dynamic Yield’s Personalization Maturity Assessment

Data Marketing by BearingPoint

BearingPoint capabilities in Data 'n' Analytics

Turn valuable insights from your data into strategic assets! Data is a valuable raw material for companies when it is correctly processed and evaluated. In these times of big data and digital transformation, the flood of data in varying formats and from different sources poses enormous challenges for companies, their business models and processes.

En savoir plus

The opportunity for fashion retailers

Everywhere we turn there are proclamations of retail doom, high street downfall and talk that things have never been so tough. But turning the tables, could this be a great opportunity for fashion retailers to shine?

En savoir plus

Artificial Intelligence in Retail: Store Footprint Optimisation

Retail Customer Journeys are becoming more complex. Sounds like an obvious statement, yet there are few Retailers that understand the true complexity that this drives into their businesses, and even fewer that are actively planning to exploit the opportunity it creates.

En savoir plus
Toggle location