[Data Day] The start-up of the day: TellMePlus, specialist in predictive analysis on mobile, raises 650,000 euros
Frenchweb invites you today to discover TellMePlus, a start-up founded in 2011 by Jean-Michel Cambot. It has built a “big data” platform, co-developed with the CNRS, focused on predictive technologies. More details with M. Cambot.
With Business Objects, I invented the product, and I ceded the exploitation rights against copyright… it was a very powerful experience. The product remains a benchmark and I am proud of it. But it was an independent adventure. While this time around I move on to a team story – JM Cambot.
Frenchweb: You founded TellMePlus and have just raised funds of 650,000 euros. Who are your new investors? How did the operation go?
Jean-Michel Cambot: The operation was rather long and was not done with the actors I was looking for at the start. My new investors are all industrialists and entrepreneurs who have invested personally, whether in terms of money or on a personal level.
Benoit Gourdon, from Neolane, Christophe Dumoulin, co-founder of Business et Décisions, took part in the tour.
Are new members coming to your board?
Benoit Gourdon and Christophe Dumoulin join me on the board. The three of us are an experienced committee. Entrepreneurs from Grenoble, including Lucien Lumbroso, and a venture capitalist Chinese are also expected to join us.
For my part, I have a technological profile, I am more about product development and vision. So I surrounded myself with people with more profiles business, specializing in business and international development.
They will bring me their knowledge in marketing and international. For example, the customers of Neolane – acquired by Adobe in 2013 – are potentially TellMePlus customers.
How are you going to use these new funds?
Much of the technology is already developed. We therefore want to use these funds for commercial take-off and the filing of patents worldwide. We are also aiming for another larger lift which I hope before the end of the year.
You market a “big data” platform, co-developed with the CNRS, focused on prediction. What kinds of predictions do you make?
We look at three major types of predictions:
- Marketing recommendation: this is the most obvious. All brands, retailers, chains or brands are potentially our customers;
- The sequential recommendation: this makes it possible to create, for example, sequences of actions for an operator who answers calls in a call center. Depending on the caller’s responses, a roadmap is adapted (eg: send an email within a week with a specific document, call back on a specific day, etc.). Technology delivers a living action plan that modulates according to people’s reactions. Exalead, from Dassault System, is one of our clients.
- Profile recommendation: We have very strong requests from stakeholders for the recommendation of profiles and matching (meeting, job platforms, etc.). I want to clarify that we are 100% anonymous. In other words, all the data is returned to a number, but no link is made with any nominative information. Only customer social networks keep names as part of the information that users have agreed to provide when submitting their information. This should also soon be imposed by European regulations.
What was one of the first issues in your development?
The first problem was to keep it concise. As soon as we started explaining what we were doing, our clients kept asking us for more (meeting the needs of smart cities, connected objects, etc.). But we had to stay very focused on the three use cases explained to offer a reliable and marketable product.
How is the “tracking” done?
Traditionally, there are two approaches. The first is statistical and data mining (looking at past history). It works well on large volumes of data, but the analysis is not very fine. We analyze mainly by broad categories of people.
The second approach, which is ours, is that of artificial intelligence where we can be finer and analyze a very precise user. The problem is, often, that it is difficult to use it on a large scale: this is what we managed to do, with a response time of less than 20 milliseconds, regardless of the volume.
How big is this market and who are the main players?
The market size is over $ 100 billion on the aspect business intelligence and marketing, and $ 35 billion for the marketing mobile applications part. You have to add the other use cases (the profile recommendation, etc.) that I cannot evaluate.
In the broad sense, Criteo competes with us. But not on technology, because we are on artificial intelligence, not on the data mining. TinyClues has developed a solution that can compete with us, but it is mostly designed for the web, less for mobile.
How do big data issues for predictive analytics vary from web to mobile?
At the level of the algorithm, nothing changes. The change is in the approach: on mobile, it must be instantaneous.
The conversion rate on mobile would be 18% against 2% on the Web according to a study. But to enter this market, several conditions: respect for private life, and be hyper relevant, because the mobile is a device much more personal than the computer. The results must be much more relevant and relevant.
You are at the origin of the Business Objetcs technology. How does one go from such a company to TellMePlus?
On a human level, that changes a lot. With Business Objects, I invented the product, and I ceded the exploitation rights against copyright… it was a very powerful experience. The product remains a benchmark and I am proud of it. But it was an independent adventure. Whereas this time I move on to a team story. I would point out in passing that Axeleo (an accelerator, editor’s note) was a very good accompanist.
Founders: Jean-Michel Cambot, joined by Philippe Courteaux
Creation date : january 2011
Effective : 6 employees, 11 by the end of the year, 20 in 2015 depending on the objectives.