Artificial intelligence and hyper customization of Marketing

During our first round table of the year that occurred January, the 30th in Hyatt Madeleine in Paris, we chose to evoke hyper customization theme. Why this choice of hyper customization helped by AI ? As Jean-Luc Marini and Julien Samarcq remembered it : “hyper customization is a necessity in order to maintain clients’ loyalty and to assure brands’ sustainability.”.

Among other arguments, Jean-Luc Marini remembered that:

  • 66% of French consumers are more susceptible to buy from companies who customize client’s experience
  • 50% of consumers wish a specific attention as good client
  • Half of them is ready to use automatic replenishment services
  • In return, more than 40% are afraid for their private life et worried about technology’s interpretation of their needs.

This consumer’s schizophrenia between injunction of a positive experience and worry linked to the intrusion of technology into their private life has of course been recorded. However, hyper customization is indeed CMOs’ objective, worried about maintain their market share and to conquer new clients. Same CMOs who are lost in a tools’ jungle, in front of technological complexity and under the double pressure of a fast result and of social injunction of being “data driven”. How to solve those equations in order to offer a positive client experience, how use AI in order to reach this objective ?

Here is the panel which is present in order to answer to these questions :

CIO of Target2Sell Adrien Coutarel, Commercial Director of Antvoice Ludovic Vannier, CEO of Visionnary Marketing Yann Gouvernec, speaker and former Jean-Philippe Cunniet, our AI Lab Director Jean-Luc Marini, and our Digital Marketing and Trade Director Julien Samarcq.

“CMOs are submitted to a real pressure created by the obligation of fast results and by the technological complexity.” Yann Gouvernnec.

What is hyper customization ?

“Hyper customization defines itself as pushed form of offers’ content’s customization, but also of user’s experience lived by the client, which has to be unique and at the individual scale in order to propose a tailored and in real time solution.”

“What clients want is value, discovery, and not an at-will hammered message.” Jean-Philippe Cunniet

According to Jean-Luc Marini, in order to reach regarding the amount of to-be-processed datum, it is difficult to do without AI. Especially since mature solutions are available. Maybe the ones used by the 39% of CMOs who use AI for their marketing targeting and 28% for recommendation.

Axys_soirée IAIt not remain that more than 60% of CMOs seem to be destitute against those stakes. As Yann Gouvernnec reminds it, “CMOs are often submitted to a real pressure created by the obligation of fast results and by the technological complexity, and finally by social pressure for “being into the data”. Today, 14% of companies are capable of realize a unique view of their clients.” The road will be long for an omni channel hyper customization approach.

All the more that as Jean-Philippe Cunnier remembers it, “What clients want is value, discovery, and not an at-will hammered message.” To the example of musical platforms’ algorithm, “proposing similar musicians to the ones I listen to has not value. Conversely, a record store could make me discover new musics, and try to open myself to new horizons”, completes Yann Gouvernnec.

In other words, not-well-think use of the couple made by data and AI creates a filter bubble by algorithm or a proposed products’ formatting. Examples are numerous into social media contents’ propositions, but also into some brands’ commercial propositions. “If I buy an apple, the fact of proposing me to buy another has not any sense anymore. Why not proposing me to discover different or affinity products, but based on my preferences”, pursues Jean-Philippe Cunniet. To the counterproductive approach of enhancement, the interest of customization is to open the consumer to other products in affinity with its preferences. In one word, to stimulate its curiosity in order to bring more value.

“Facing tools, CMO has absolutely to step back and to be aware of the fact that everyone is lost” Yann Gouvernnec

Which tools for the hyper customization : “Everyone is lost”

With which tools ? To make its choice into the 7,040 available MarTech tools (listed by chiefMartec.com) inevitably dive CMO in an abyss of perplexity and sideration. “Facing tools, CMO has absolutely to step back and to be aware of the fact that everyone is lost”, tempers Yann Gouvernnec who shares this anecdote of a DMP (Data Management Platform) ‘s editor who, in front of market’s hazards, has had upgraded its tool towards a CDP. During an interview from the CEO in order to explain what is a CDP, he had absolutely no idea of what it is.

“Do not forget to consider uninstall of the solution and the associated cost !” Jean-Philippe Cunniet

Axys_AICEO’s perplexity, largely shared with majority of CMOs should not prevent them to implement solutions and to initiate projects by “come back to fundamentals” and especially “think firstly to data quality and make use of good sense. Without data cleaning, nothing efficient can happen”, reminds Yann Gouvernnec. A view shared with Julien Samarcq who advises to focus on 1st party data and to “progress step by step in order to deliver a positive client’s experience.”

Let’s remember that 20% to 35% of turnover’s loss are due to a bad data exploitation according to PwC. Without counting unreliable data, 84% of companies’ common lot. Data quality indeed starts with the consolidation and capitalization of the existing.

“AI’s images and voice process capacity permits to propose scenarios of interaction with the consumer by finding the right time and in function of the context.” Jean-Luc Marini

Arriving to the choice of tools moment ? Jean-Philippe Cunniet gives us a little advice: “Do not forget to consider uninstall of the solution and the associated cost!” A condition sine qua non in order to the maintain the company’s agility and its capacity of evolution without friction from one tool to another.

Going step by step, by cleaning datum, hyper customization implementation operates well, affirms Jean-Luc Marini, “Artificial intelligence presents in many tools is very efficient for the predictive. As well as AI’s images and voice process capacity permits to propose scenarios of interaction with the consumer by finding the right time and in function of the context.” Words confirmed by Adrien Coutarel for who « each AI answers to a specific need. The stake is to identify the AI that will be adapted to your use case.”

Which datum for user’s experience ?

Having right AI does not exonerate CMO to capture data in order to understand the client. A challenge while Google (rejoining Apple and Firefox) renounces to use cookies into its navigators, with at result a certain difficulty to collect personal datum that better target consumers, according to Ludovic Vannier.

To bypass this abandonment which principaly impacts 3rd party datum, the alternative resides either in the Maketing permission -inbound marketing- and/or, more surprising, the use of the article number 20.6 of GDPR.

To Jean-Philippe Cunniet, « GDPR can have a positive aspect. But in a first tume it is important to explain to the user what we are going to do with its datum while the consent collection. Moreover, GDPR’s 20.6 article authorizes datum portability from one service to another. Rather than competitive portability, for example changing phone operator, why do not imagine a complementary portability. I can imagine to allow the exchange of my datum collected by Auchan to Marmiton. When I go shopping,my datum could be used by Marmiton for proposing me recipes in function of my basket.” A datum exchange creating value for the consumer and virtuous for operators. This synergy between operators would permit to go from one client’s experience to a more extended one. “If Air France informs an Uber driver that his passenger is ahead, it is in its role and provides an extended experience to its client. To the contrary, if the client waits too long its driver or does not find one, its global experience will be negative, no matter what the quality of Air France’s prestation.”, illustrates Jean-Philippe Cunniet.

“How far to know the consumer ?”

Hyper customization consists of knowing the consumer, its preferences, its purchases, context in which it is, what is it doing, where it is and so on. Conversely of this exhaustive profile’s quest, each speaker preached for a relevant information quest in order to bring value. With at result Netflix’s softcards example that tests spectators’ preferences in function of their degree of palatability to violence, action scenes, love, humour, etc. Strong of these tests, Netflix is even capable to propose adapted entertainment, and goes farer by adapting colorimetry for corresponding to spectator’s preferences. All of this in real time thanks to AI.

To this extreme example of customization linked to recommendation, others companies success to propose relevant offers like SEB or Yves Rocher. In the contrary, the opening of Fnac to some thirds partially makes him lose control of a part of datum that come from marketplace. Direct consequence of it, lived by speakers, is the reception of totally fanciful recommendations from the e-commerce platform.

No matter what the volume is, lesson is clear : “Data has to be extenso-controlled.” Whether for proposing three offers like an insurance product, or millions as Amazon does, control from start to finish of datum is essential for a relevant recommendation or targeting.

“For obtaining quick wins, CMO can start with predictive marketing of which deployment does not take many time for immediate results.” Ludovic Vannier

Hyper customization and AI : tests and explanations, key success factors

Waiting for this wish’s realization, how to initiate a hyper customization project ?

It is important to progress step by step for progressively understanding stakes that are linked to hyper customization. In order to obtain quick wins, CMO can start with predictive marketing of which deployment does not take many time for immediate results.” Confirmed words by Ludovic Vannier whose tool is operational after a first phase of audit and who advocates two month test campaigns for assessing first scenarios.

“You also have to think about what you are deploying. Everything has to be explainable and your project has to be understandable by collaborators. Paradoxically, project can continue, but if the company does not understand it, it is doomed to fail”, reminds Adrien Coutarel by evoking companies’ numeric transformation and necessity to exchange information and to acculturate collaborators.

Axys_AI

Hyper customization and transformation : a global project

Hyper customization is a complex subject. But it is today accompanied by mature and efficient products to some conditions, such as the data’s quality one. During the project’s elaboration, CMO should not manage it alone, but has to embark all stakeholders in order to deploy tools, clean and implement the process. Hyper customization is a structuring project for the company and it helps to create value. Not well managed, it is exactly the contrary that will happen : a loss of turnover and CMO’s disavowal.

Thank you to Frabrice Froissard who well anylyzed this event

Hyper customization and artificial intelligence : we must save CMO soldier!