Marketing, customer satisfaction and loyalty
Satisfied customers will follow you everywhere
What if Bard (Google) was trained on data from Gmail?
Apr07

What if Bard (Google) was trained on data from Gmail?

The scandal is potentially huge. Kate Crawford posted on Twitter (see screenshot below) a response by Bard, Google’s conversational agent using generative AI, to a seemingly innocuous question, “Where does Bard’s data come from?” Although Bard’s answer may be a hallucination, it is important to remember that Google has demonstrated its ability to disregard the most basic rules of privacy and personal data protection in the past. I would not be surprised if there were some truth in this...

Read More
chatGPT plugins: disruptive to the future of tech
Apr05

chatGPT plugins: disruptive to the future of tech

Expedia just announced on Twitter the launch of a plugin powered by chatGPT’s generative AI. At the same time, a “hacker” revealed the “secret” list of plugins available on chatGPT. Contrary to popular opinion, this will not challenge Google search as we know it. However, it could change the whole tech ecosystem. The chatGPT plugin ecosystem may well map out the future of LLM (Large Language Models) usage. The danger for Google does not come from chatGPT itself but...

Read More
ChatGPT: 4 sources of risk for inbound marketing
Mar23

ChatGPT: 4 sources of risk for inbound marketing

Companies that rely on SEO to find customers may need help with llead generation. After a few weeks of using ChatGPT and the new Bing, I look at the dangers of these conversational agents for companies using inbound marketing. Based on the search context, I have identified 4 issues likely to impact users negatively. Contact our agency for your inbound marketing projects Summary The opacity of the data used for training Wrong but very convincing answers Selection bias Risk to serendipity Conclusion The...

Read More
ChatGPT and Bing won’t replace Google. Here’s why.
Mar07

ChatGPT and Bing won’t replace Google. Here’s why.

According to Microsoft, 50% of online searches could be carried out via a conversational agent (Chat BOT) based on generative AI. But what if Internet users are not interested in this value proposition? Everything depends on Internet users’ willingness to change how they search for information online. But the user experience proposed by Google has been ingrained in us for almost 20 years. So, I wonder if Internet users will change their habits and massively adopt the new...

Read More
ChatGPT: 1000 texts analyzed and up to 75,3% similarity
Feb16

ChatGPT: 1000 texts analyzed and up to 75,3% similarity

Is every answer generated by ChatGPT unique? Or are we overestimating its ability to produce different texts? This is the question I asked myself after analyzing 1000 texts produced by ChatGPT. I compared the similarity between 1,000,000 pairs of texts generated by this artificial intelligence to answer this question. The results are surprising and challenge the promise of OpenAI. Find out in this article the results of our in-depth analysis and what it means for the...

Read More
6 algorithms that have triggered disaster
Oct12

6 algorithms that have triggered disaster

Algorithms are increasingly used to help humans in their decisions. A poorly programmed algorithm, or humans lacking critical thinking skills, are recurrent causes of algorithmic errors. These errors sometimes trigger users’ anger or incite them to hack them. This article analyzes 6 algorithmic errors that have had harmful consequences on humans. The 4th example is particularly serious because it impacted thousands of people and caused irreparable damage (suicides, separations, unjustified debts). Only for our subscribers: exclusive analyses and marketing advice Email address  *...

Read More
ETL: Anatella in web version with high performance
Jun17

ETL: Anatella in web version with high performance

If you are interested in data science, you know that data preparation is a very time-consuming step. That’s why choosing the right ETL software is so important. My favorite solution is called Anatella. I appreciate its qualities in terms of speed and the richness of the proposed transformations. So, of course, I am always enthusiastic when a new product arrives on the market. In this case, it is the web version of Anatella which allows the publisher Timi to propose a very disruptive price. Anatella...

Read More
Patek Philippe Nautilus: how have prices changed in 4 years?
Jun15

Patek Philippe Nautilus: how have prices changed in 4 years?

How much does a Patek Philippe Nautilus cost? How will Nautilus prices evolve? Answering these questions, as relevant as they are, has become extremely difficult following the bubble formed in the luxury men’s watch market. Some models have seen their prices explode. The Nautilus from Patek Philippe is one of them. This article analyzes the price history of 31 Patek Philippe Nautilus models. We reveal the models that have gained an absolute value, the most dramatic relative increases, and...

Read More
LinkedIn’s algorithm changed again in 2022: what impact?
May16

LinkedIn’s algorithm changed again in 2022: what impact?

LinkedIn’s algorithm underwent some pretty significant changes in May 2022. LinkedIn introduces constraints in its algorithm and takes into account more explicit variables. This has had a substantial impact on the visibility of posts. Although the basic functioning of the LinkedIn algorithm is not changed, we analyze these announcements for you. The goals of LinkedIn’s algorithm changes The rules for highlighting posts have not changed. Engagement is still the number one criterion for the algorithm. However, LinkedIn introduces mechanisms to penalize certain...

Read More
57 unpublished Linkedin statistics
Apr15

57 unpublished Linkedin statistics

The LinkedIn statistics you find on the internet are often the same: the number of users, geographical distribution, and age … In this article, we offer you 57 new LinkedIn statistics. They are unpublished simply because they are calculated from a dataset that only we have. This is the same dataset that allowed us to unravel some of the mysteries of the LinkedIn algorithm (based on dwell time since 2020) and highlight the 3 factors that explain 55% of the virality of posts. Check...

Read More
Men – Women: the battle for influence is unequal on LinkedIn
Apr04

Men – Women: the battle for influence is unequal on LinkedIn

In a first analysis, I showed that most female influencers on LinkedIn were young (18-30). Could age be a factor in the virality of LinkedIn posts? To be sure, I had to check if my first conclusions applied to men too. If you only have 30 seconds Half of female LinkedIn influencers are under 30 years old. That’s twice as many as their male counterparts Across all languages, men and women get about the same ratio of comments: 1...

Read More
On LinkedIn, women are 17.3% more viral than men
Mar28

On LinkedIn, women are 17.3% more viral than men

Is it better to be a woman to be a LinkedIn influencer? At least, that’s what a preliminary study we conducted suggested. Today we extend this study to nearly 5 million LinkedIn posts. The results show that beyond 500 reactions, women’s posts get 17.3% more reactions. However, there are significant differences by language. If you only have 30 seconds Globally, only 0.5% of posts on LinkedIn get more than 500 reactions. About half (0.3%) are authored by a LinkedIn user with fewer than...

Read More
On LinkedIn, to create a Buzz, it is better to be a young woman…
Mar14

On LinkedIn, to create a Buzz, it is better to be a young woman…

Analysis of 1118 viral LinkedIn posts shows the domination of women in terms of reactions (likes and comments). Among these influencers, there is also apparent domination (nearly 50%) of women aged 18 to 30. This domination is exercised in terms of the number of profiles in this age group and the number of reactions collected. Although men are more numerous, they collect fewer reactions on average. Lenna Jouot dominates the ranking in the French-speaking world thanks to 6...

Read More
CBD online: French and English, all addicted! [SEO Research]
Feb07

CBD online: French and English, all addicted! [SEO Research]

SEO research on online searches related to CBD shows that 2 countries share the first place. France and the UK have the highest number of searches per capita. The SEO research we realized in 31 countries also shows significant differences in the keywords used. Follow us into the beautiful world of CBD online; it may surprise you! Summary Introduction Methodology What are the most searched CBD-related keywords per country? Volume/difficulty ratio: ranking of European countries  ...

Read More
ETL: comparison, selection criteria, advice [guide 2022]
Jan26

ETL: comparison, selection criteria, advice [guide 2022]

ETL: Extract – Transform – Load. These 3 words alone summarize the functionalities that software solutions belonging to this family must fulfil. THEREFORE, an ETL tool is a software that simplifies data preparation, local or on the cloud. The latter focuses on realizing 3 types of operations on data: (1) extracting them from files of different natures; (2) transforming and enriching them to prepare them for exploitation; (3) reloading them into another system so that they can be...

Read More
Netflix ratings: tough competition for the most popular movies and series
Dec13

Netflix ratings: tough competition for the most popular movies and series

Netflix now publishes its weekly ratings in the form of a TOP 10 of the number of hours watched worldwide. Although Netflix only publishes the total number of hours watched, this data is still rich in information for those who want to understand how the ratings of the top Netflix movies and series evolve. In this article, we analyze the data of the most popular Netflix movies and series by distinguishing 4 categories: films in English foreign language...

Read More
Talking to the dead, the future of artificial intelligence?
Dec01

Talking to the dead, the future of artificial intelligence?

Deadbots are a new application of artificial intelligence. Personally, it scares me. Deadbots are chatbots that allow you to converse with dead people by imitating their responses. Their development is part of a more global context of virtualizing relationships and blurring the boundaries between the physical and digital worlds. Joshua Barbeau spoke with his girlfriend, who had died 8 years ago In August 2021, the San Francisco Chronicles published the story of Joshua Barbeau, a 33-year-old Canadian. Eight years earlier, the...

Read More
We forgive mistakes made by algorithms more easily than by humans
Nov26

We forgive mistakes made by algorithms more easily than by humans

Algorithms have taken over. They are in charge of making more and more decisions. More and more complex decisions affect all aspects of our lives. But algorithms are not perfect. They also make mistakes. So, a question arises: do we forgive algorithms more easily than human beings for the same mistake? Research from 2021 sheds new light on this question and provides essential insights into the ideal characteristics of algorithms and chatbots. Contact us for your data project If you only...

Read More
Cloud Act and GDPR: can we host our data in the Cloud?
Nov17

Cloud Act and GDPR: can we host our data in the Cloud?

Does your use of cloud services lead you to violate the GDPR? Can you store your data on AWS, Azure, … and still comply with the GDPR? The American Cloud Act sows doubt and creates a legal uncertainty which we discussed with Me. Jérôme Tassi, an attorney at the Paris Bar. Introduction The starting point for this article (and the accompanying video) is a thought brought to me by an entrepreneur friend. He was worried that his company’s...

Read More
The panda is the most rewarding emoji on LinkedIn
Jun28

The panda is the most rewarding emoji on LinkedIn

An analysis of 2.9 million emojis on LinkedIn delivers a surprising piece of information: the ? emoji elicits the most reactions. This research offers us an opportunity to critically look at the data and understand the origin of a polluting LinkedIn phenomenon. Follow me for an exciting data mining exercise. To learn more about the factors that impact virality on LinkedIn, I invite you to read my complete analysis here and send me a request to connect on LinkedIn. If you...

Read More
The most used emojis on LinkedIn [analysis and statistics]
Jun21

The most used emojis on LinkedIn [analysis and statistics]

What are the most used emojis on LinkedIn? Do the most used emojis on LinkedIn vary by language and country? In this analysis, discover exclusive statistics on the use of emojis on LinkedIn. If you’re passionate about LinkedIn statistics, don’t miss my other posts on the subject: hashtag statistics, the 3 factors that explain 55% of the virality of a LinkedIn post and the sentiments expressed in LinkedIn posts. Summary Introduction Methodology The most used emojis on...

Read More
Which are the most popular hashtags on LinkedIn?
Jun16

Which are the most popular hashtags on LinkedIn?

Hashtags on LinkedIn are probably one of the most popular topics about this social network. So naturally, there are assumptions about how LinkedIn’s algorithm uses hashtags. However, I have shown statistically here that hashtags do not significantly impact the virality of posts. Anyway, the dataset I have at my disposal (4.6 million posts) provides an exciting overview of the most used hashtags globally and the differences that exist depending on the language. This analysis is, therefore, fundamentally different from anything else...

Read More
Tableau tutorial: a radial diagram and a pie chart
Jun09

Tableau tutorial: a radial diagram and a pie chart

How to make a radial diagram using Tableau? How to integrate a pie chart? In this article, I guide you step by step in the realization of this data visualization. Introduction For my subsequent research on fundraising in Europe, I wanted to test a data visualization in the form of a radial diagram. To get an even more exciting result, I integrated an interactive pie chart in its center. You will find 2 parts in this Tableau...

Read More
The 11 challenges of data preparation and data wrangling
Jun07

The 11 challenges of data preparation and data wrangling

The events of 2020 have accelerated the shift to telecommuting and digital relationships. With digitalization, another transformation is also underway: the analytics transformation. To face this data revolution, I find that businesses do not necessarily have the right tools for preparing and analyzing data (data preparation or data wrangling). In this article, I wanted to put into perspective the role of these tools in the functioning of the business. In particular, I think they can contribute to efficiency...

Read More
Data preparation: how to reduce the processing time by 85%
Jun02

Data preparation: how to reduce the processing time by 85%

In a previous article, I realized using a benchmark of 4 ETL solutions to process a file of one billion lines. Today I test the effect of SSD and proprietary file formats on processing speed in Alteryx, Tableau Prep, Talend, and Anatella. The results are quite unexpected. Introduction and review In my previous analysis, I compared the processing speed of 4 data preparation solutions: Alteryx, Talend, Tableau Prep, and Anatella. After its publication on social networks, which raised several voices to criticize the...

Read More
Here is the most crucial factor for the virality of your LinkedIn posts
May28

Here is the most crucial factor for the virality of your LinkedIn posts

After discussing the effect of emojis and the word count on the virality of LinkedIn posts, it’s time to reveal THE factor that crushes all others. This factor accounts for 1/3 of the success of your LinkedIn posts. So, you should take the time to read the research results I just conducted on 4.6 million LinkedIn posts in 193 countries. I obtained the unpublished results presented here thanks to data made available by Linkalyze. They were prepared and analyzed using the...

Read More
LinkedIn: engagement statistics by country and language [2021]
May26

LinkedIn: engagement statistics by country and language [2021]

How does engagement on LinkedIn vary by language and country? Globally, 71.06% of LinkedIn posts receive no comments, and 24.39% receive no likes. The analysis of 4.599 million LinkedIn posts gives us an objective view of the essential differences that exist. The data comes from a dataset made available by Linkalyze. The data preparation was realized using Anatella (Timi) and the visualization using Tableau. If you are interested in these results, do not hesitate to subscribe to my...

Read More
Signs of the Times : an artistic project on algorithmic recommendations
Sep08

Signs of the Times : an artistic project on algorithmic recommendations

Recommendations are everywhere online but offline recommendations have never been heard of. Until New-Zealand based artists Scott Kelly and Ben Polkinghorne chose to use this widespread technology in an offline context. The result is an eye-opening experiment called “Signs of the Times” that forces us to look at the predominant place recommendations have taken in our online lives. What would we do without recommendaiton engines ? While the fear of filter bubbles exists (and we tried...

Read More