Procurement management : 7 digital challenges to overcome in 2021

Being a purchasing director in 2020 means being aware of the growing importance of its role in terms of leading innovation in the company. Purchasing has in fact a double impact, as a performance driver that is key to maintaining and improving margins, but also in being at the forefront of innovation.

What is the impact on the strategic organisation of the purchasing department? What are the new challenges arising from the digitalisation of the purchasing department?

The increase in the volume/speed/variety of data in recent years has increased the risks related to the strategic integration of data. It is therefore up to the purchasing department to ensure that it anticipates these risks in its daily operations.

—————————————-STRATEGIC INTEGRATION OF DATA———————————————-

CHALLENGE #1 - CHECKING DATA PROPERTY

More and more data is being collected by the purchasing department throughout the supply chain. This raises the question: when sensors are installed, for packaging reception for example, does the data belong to the company that manages the sensors, the manufacturer of the sensors, the company that had the sensors installed, etc.? Agreements need to be defined prior to this in order to be able to use this information correctly and transparently, as there is a real risk linked with data sovereignty. 

CHALLENGE #2 - SCRUBBING COLLECTED DATA

Collecting data is not enough to offer new opportunities. It is necessary to clean and standardise this data, which requires new skills... The purchasing department must therefore consolidate and clean its data thanks to skills developed internally, or externally with data specialists. One of the challenges will be to automate these two structural but redundant tasks

CHALLENGE #3 - PREDICTING RISKS THROUGH IA

Another challenge will be to use this data in machine learning simulations. Different solutions will be able to provide valuable input to purchasing departments, which would otherwise be impossible to get. AI offers the possibility of predicting trends (order peaks, stock shortages, cost increases or even identifying potential risk situations with suppliers (see this article).

————————————————————-SUPPLY CHAIN  ———————-————————————-

CHALLENGE #4 - SUPPLY CHAIN OPTIMISATION THROUGH CUSTOMER DATA

The increasing amount of data available on production lines represents an incredible opportunity for process optimisation. After integrating the data from the different management tools into a common IS, traceability can be clearly improved, as well as reducing storage costs. Close collaboration and data pooling between the purchasing and logistics departments will be essential for the purchasing department to meet this challenge.

CHALLENGE #5 - TACKLING INCREASING ELECTRONICS-RELATED EXPENSES

Regarding the evolution of the type of purchases within the company, purchases of electronic equipment has boomed. Between the performance needs of employees and the programmed obsolescence of suppliers, the purchasing department will have to anticipate the importance of this type of expenditure.

———————————————-—————-SOURCING—————-———————————————-

CHALLENGE #6 - FINDING INNOVATION: THE UNEXPECTED CHALLENGE

Unexpected but logical, as purchasing is nowadays at the forefront with external suppliers. As start-ups grow in size, the relationship is changing: large/small organisations, SaaS subscription... The trend at most companies to centralise purchases is driving the purchasing department to take on this new approach. But it also puts the purchasing department in direct contact with the market and its evolutions, and can become the number one source of information on innovation in the sector for the rest of the company.

———————————————- TRANSACTIONAL —————————————————-

CHALLENGE #7 - PROCESS IMPROVEMENT

This is not new to purchasing managers. However, digitalisation brings new opportunities, especially in Source-to-Pay or supplier reference management. For instance, digitalisation allows intelligent recording of the data to be integrated, or immediate access to all the elements of the transaction.

Automated Procurement-to-Pay and expense analytics accessible through a data-assistant capable of answering questions also represent a major innovation for procurement. Real-time and accurate decision making is a huge advantage for purchasing departments!

>> Discover askR.ai, the head of procurement data-assistant <<

A data-assistant to query Snowflake

Frustrated by the sluggishness of the major database providers they were working for at the time, 3 frenchmen funded Snowflake in 2013. Snowflake is the first cloud-native data warehouse. The team has raised $923 million since its foundation, for a current valuation of $3.5 billion. Thanks to a unique architecture, they are disrupting cloud standards, fostering innovations such as askR.ai.

Separating data processing from storage, a brilliant idea.

How did the company manage to win over competitors such as Amazon or Microsoft in just 6 years? Thibault Ceyrolle, president of Snowflake Europe, has an explanation “The cloud architecture has been designed to multiply the number of simultaneous users and the number of requests tenfold without compromising performance".

How? By separating storage from processing, i.e. by mobilising server resources on the go to respond to specific requests. Data is transferred only once, and the power of the servers is totally modular. This pay-per-use model, invoiced per second, is better suited to the needs of companies that are increasingly demanding on the scalability and cost control offered by cloud solutions.

>> Read the article "Cloud solutions are more secure than on-premise" <<

"No time to data" : a leitmotiv as important to Snowflake as it is to the data-assistant askR.ai

Snowflake's architecture significantly improves performance by reducing the time needed for queries. According to Gartner, only 35% of employees use traditional BI due to cumbersome setups and poor user experience. askR.ai and Snowflake, both are extremely simple to integrate with existing infrastructures, reducing setup time and increasing request speed. Snowflake is at the top of the data value chain, mainly used by data scientists who are able to grasp its full potential. askR.ai is further downstream, mainly used by employees who are not very familiar with the complexities of data architecture.

For Matthieu Chabeaud, CEO of askR.ai, the integration with Snowflake was obvious. "It is a new disruptive player in the Cloud industry. Companies interested in it can consider, as part of their global strategic vision, integrating a data-assistant into the daily lives of their operational staff. Snowflake is a multicloud solution, a strategy shared by askR.ai as it is also compatible with Microsoft Azure, Google Big Query and AWS. 

To make it short, these two solutions appear as essential as they complete each other when scoping the current data market innovations.

Big Data : how to lead your teams into going further with data ?

The common denominator of these meetings? Leading businesses further in the use of their data towards a "data-driven mindset".

SHARING DATA AND DEBUNKING MYTHS ABOUT AI

Cédric Le Saveant - VP Digital Sourcing Technicolor has setup Jim, a data-bot to handle purchasing data.

For Cédric Le Saveant, VP Digital Sourcing, teams need to be able to access data at all levels of the company. For almost 4 years now, he has been introducing a real change in Technicolor's purchasing culture. For him, it's very clear: having beautiful dashboards is very good, but having a tool as interactive as a bot capable of answering your questions about data is a concrete change. "It's the data that comes to you, and it changes everything! ", says the one who wants to make office tools as simple as the ones we use in our daily routine. Indeed, one of the biggest challenges when you want to put data at the heart of an organisation is the adoption rate of the solution, very much tied to its ease of use. William Marcy, Technical BI Director at TVH Consulting, says from experience: the same solution, presented with a more attractive portal, can be successful after a failure. He therefore distinguishes between two types of BIs: BI for experts, and BI for management teams. "We have very advanced but also very complex platforms that only correspond to the needs of the first category," adds Matthieu Chabeaud, CEO of askR.ai data-assistant. "However, it is the management BI teams who need to be involved in this project”.

"We must separate BI for management teams with the one for experts, and acknowledge that very complex platforms only meet the needs of experts." Matthieu Chabeaud, CEO askR.ai

Bringing natural speech to Business Intelligence is, in the end, a return to the most basic form of communication! AI technologies that allow us to communicate and get information in this way only bring back a basic and essential method of expression," reminds Matthieu Chabeaud. This is why it is particularly important to make users understand that AI is not a black box that acts like a human with unknown objectives, but rather as algorithms that can learn. For Thomas Binant of Géotrend, it is therefore necessary to give teams the means to learn and understand how to use AI tools.

FINDING INNOVATION STARTS WITH LEARNING HOW TO USE ITS DATA.

Salons Solutions Big Data - Leading businesses further in their data use.

Another challenge in adopting new tools is having IT/data teams too much in the operational side of things that they may neglect to look for new tools that can make their work easier. There are two ways of looking for innovation: either by looking for solutions themselves, or by getting support from a consulting firm. When you are an SME anchored in everyday life, you can dangerously get used to the discomfort of a solution that "does the job". Matthieu Chabeaud calls it the “fakir” syndrome, or how one no longer even feels the pain in a situation to which one has become accustomed!

It is essential to onboard BI tools in consultation with the IT department. If the IT department is excluded from the process, it is perhaps due to a lack of understanding of the pains, but also because management teams are thirsty for quick results. The IT department is not always in the loop on new decisions, but not including it from the start can significantly slow down projects. Rather, the IT department should be allowed to "restore its image" as it can bring real value to the business department. It's exactly what happened at Renault, where the project of an internal portal using a data-assistant (literally) triggered applause on its launch! "That was to tell you how great it matched between the issue and the new tool suggested for marketing data!" remembers Matthieu Chabeaud.

However, Xavier Bouteiller from Datasulting points out, that the success of data projects is based on the company's strategy: marketing, finance, information systems or sometimes the management itself. Whatever the case, it is necessary to ensure that users are as responsible as possible, by making sure that producers are also consumers. Users that regularly use the available data - whether it is to get accurate information on a contract or overall performance management over several years - are better able to see the benefits of AI solutions. The use of data opens up a path of global thinking: how can I make predictions in order to improve my decision making? Could certain tasks be automated with Machine Learning?

Autonomy and responsibility are therefore the two essential elements to lead users to new opportunities.

GO FROM SIMPLE TO ADVANCED ANALYTICS THANKS TO ASKR.AI'S DATA-ASSISTANT DISCOVER ASKR.AI

--------sum up--------

Conference « Value creation from data : nothing compares to BI, AI and data scien sinergy. What new gen plateforms are up to the challenge ?

with Pascal Minguet - Independant journalist

Salons Solutions talIntervenants :

• William MARCY, Directeur Technique BI, TVH CONSULTING

• Thomas BINANT, CEO, GEOTREND

• Matthieu CHABEAUD, Co-fondateur & CEO, ASKR.AI

• Cédric LE SAVÉANT, VP, Digital Office - Sourcing & Supply Chain Transformation , TECHNICOLOR

Table Ronde « Popularize analytics within SMBs : which methods and tools to make business users want to go further with self-service data ? “

with Pascal Caillerez - Journalist
Intervenants :

• Matthieu CHABEAUD, Co-fondateur & CEO, ASKR.AI

• Xavier BOUTEILLER, Directeur associé, DATASULTING

Data & Innovation : listen to the business users! #Datalab

Key accounts‘ opinion on DataLab (ENGIE, TF1, MAIF…)

Data Lab, Data Fab, Data Innovation Lab… Whatever expression is used, this kind of structure have been flourishing for 4 or 5 years. Paradoxically, Gartner still pointed out in 2017 that 85% of Big Data projects failed*. Wait, what ? A datalab would not be enough to properly gather and exploit data to make it available to every business users…? Read these major accounts testimonials to find out.

DATA projects : common difficulties.

One of the main reason for failure concerning data is the low adoption rate of new tools. The issue of adopting innovation is nothing new for change managers, who are confronted to change resistance on a daily basis. It is ridiculously obvious, and yet, leading countless initiatives to death.

When it comes to enterprise data, you will then be confronted to silos. Every business unit will go right to familiar context without a glance for general value creation. And when it comes to searching for relevant use cases, you can easily be blinded by the competition showing-off instead of focusing on your own business, because you are afraid of being left aside.

These 3 dead-end have in common the neglect of the innovation concept in its core definition: searching for a new, non-existing path. The best way to avoid these dead-ends? Taking into account data-specific technological needs and offering an environment favourable to new ideas to flourish.

Definition of a datalab: innovative workspace to explore, find and execute data driven improvements of business processes.

Different structures and one goal to bind them all: transversal transformation.

A datalab has to come within the scope of a company’s global transformation. It might be its spearhead or a simple side-project. Olivier Baes, MAIF dataLab manager, presents “the MAIF digital transformation support at every step through conferences, community management and internal communication” as its datalab main mission.

Various models are possible.

An independant datalab is possible, made of data specialists working on projects as an independent cell. At Generali France for instance, the board chooses projects to implement. Employees are trained from 3 to 6 months within the datalab, according to subjects. When they are sent back to their teams, they acquired new data skills they can use and share to the whole team, explains Hélène N'Diaye, member of Generali COMEX.

At GRTgaz, it is more about experimentation offered by its datalab structure. A place outside of the company premises was chosen on purpose to give a unique space for considering submitted business issues in a different way. According to Frederic Mours, head of Datalab of the industrial leader, accessing data is a strategic challenge as most of the teams are on the field.

On the other side, Swiss Life launched a more transversal organisation with its team members. DataLab members keep their original missions and work simultaneously on AI and data science matters. The Datalab depends on the COMEX, data strategy being directed by Cynthia Traoré. The SwissLife DataLab members meet regularly to exchange on various data issues.

At TF1, data issues are handled inside of the innovation center named MediaLab. Every 6-month start-up batches answer different issues of this TV channel core business. Every Business units own a start-up with whom a project has already been set. “ The first stake is always operational, which means data is always at the center of our business concerns”, underlines Florence Caghassi, Open Innovation Program Manager at TF1.

To conclude, a structure entirely dedicated to data promotion has to be a pilar of innovation strategies of big accounts. It can be a place, a team, transversal skills … But it has to help new solutions flourish, thanks to data-based exchanges .

“Not only is data exploitation about digital, but also a mindset that comes from more team interactions. “

Nadège Vignol, Head of Data Innovation Lab Engie

* source Nick Heudecker, Gartner 2017

Sources : Le datalab de Generali France // Faut-il un data lab pour innover ? Data Analytics post // Un data lab pour l’innovation, entretien avec Nadege Vignol

Data & Innovation : listen to the business users!

CIO-France Online Conference : Joining forces on Data democratization

Focus on Renault testimonial

Some feedback on the collaboration between Renault and the start-up askR.ai. >> original article (in french - Front commun sur la data : retour d’expérience de Renault <<

Top/down strategy request appropriate intermediaries

Since 2014, Renault has clearly announced its digital transformation ambition. A corporate vision that matches the “top/down strategy” definition of a company strategic orientation pushed from the top. The following years saw the creation of a new entity called “Renault Digital” to internally develop apps and softwares to answer business units’ needs. In this transformation process, data exploitation is key.

Jean Dumas, BtoB Customer Data Manager at Renault, started working on data issues in 2015. This is the reason behind his participation in the CIO France conference “Joining forces on data democratization”. As a CRM project manager, his experience feedback is all the more relevant as he is now responsible for making these positions more "data autonomous".

Choosing a manager from the field, in order to find new data tools and to make sure they are adopted, enabled a down-to-earth approach. It is about overcoming daily issues that face Renault distributors and that Jean Dumas has also faced.

Accepting feedback (or the absence of feedback)

Jean Dumas closely monitors the implementation of a new data portal for his teams from 2015: Fleetb@se, a dashboard helping operational teams accessing key information on their local market.  This tool was a first response to business users needs.

 However, our head of project quickly became aware that this new tool was not used enough.  He started looking for a solution that would make users more autonomous without entirely rebuilding an already operational system. It only requires a simple, instantaneous access that can be used by different types of users: business users tend to give up on a tool that is not fast enough.

Jean Dumas turned to askR.ai in 2017 to add our data bot to the portal: users could ask their questions and receive an answer instantaneously, without having to dig for the answer in FleetB@se. The number of connexions were multiplied by 7 in a few months !

Thanks to its relevance, the intuitive solution sparked off a “wow effect”. After all, every car dealer’s dream was simply to get a precise and instant answer to a question they had in mind when opening FleetB@se… Question that can now be asked to askR.ai, 24/7.

Develop data exploitation with a double approach

This CIO-Online conference testimonial proved that it is complex to offer business users a solution that meets both their daily issues and the company strategic orientation. When talking about data and users, different needs combine with multiple approaches and numerous tools. Last but not least, every piece of data has a different value according to its accuracy and its timing.

Decision making comes from data, which makes easy access to data for business users paramount. That is why IT and business teams have a common stake: IT is in charge of framing technical resources, and business teams, the only ones able to decide on their true needs. This, even within the framework of a top-down transformation strategy.