Mopsos

I went this morning to a conference organized by l'Atelier in Paris on behalf of Temis, a text-mining software company.

Temis's presentation was about its software and services offering: semantic technologies for extraction, clustering and categorizing of information. It was brilliant, and Temis undoubtedly has something to say in this area, as their track record suggests.

I was however a little uneasy about their marketing strategy, for two reasons:

1- Temis said nothing about the limitations of text mining technologies in general, and theirs in particular.

2- With prices ranging from 100 to 500 KEuros for a Temis solution, the target clients are clearly the corporate staff of large companies.

1- I had to ask Anne Bonnet-Ligeon from Total - a very satisfied customer of Temis apparently-, who was presenting a customer case with Temis, her opinion on this. Her answer was that Total made a comparison between "manual" extraction and automated extraction using Temis Extractor. 92% of the results were the same, and the 8% that Temis did not find were "poorly structured documents", which is not a big surprise. If Temis people don't clearly state in their white papers that "poorly structured documents" are unlikely to be extracted with their technology, they are selling snake oil, which, in the long run, will not do them justice.

2- As everyone likes to repeat: "information is power". Thus, in the vast majority of large corporations, where mutual trust is definitely not a core value especially among staff people, the guys who control information today will want to buy systems that will help them stay in control over information in the future. They will use the "time saved" argument for lack of a better one because they are too remote from the actual business. And ultimatey, because nobody really buys the "time saved" argument, as Carla O'Dell points out, those guys are driven by internal politics. They still have budgets today, and that's why companies like Temis are talking to them. But, in such a highly political environment where such tools are used to maintain the organizational status quo and do not provide an answer to a burning business need, lead times can be quite long. There was almost two years between the first visit of temis at Total and the closure of the first order.

Temis, like many other KM software companies, might be a 21st century company by the products they make, but they are still a 20th century company by the way they go to market. If "information is power", as everyone claims, then a lot of people are craving for good information that helps them manage their career. In the corporate world, I can think of many text mining applications that employees would be ready to pay for as individuals if they provide answers to some of their daily concerns: "What partnerships is my company currently negotiating?", "What is the image of my company in my personal professional field?", "How's the job market for people like me?" or even, in a (not so distant) future: "What is the world's opinion on me?". The same holds for groups of people like project teams or communities of practice. The trouble is that individuals and small groups don't have this kind of money to spend. Typically, what an individual is ready to pay for yearly is what he can very easily be reimbursed for by his company, so that's in the $10 to $500 range; what a community or a team can pay for is likely to be higher, in the $ 500 to $ 5000 range. So how come companies like Temis don't team up with news syndicators like Factiva or Lexis-Nexis to offer an online paying text-mining service targetting individuals and groups and giving them access to valuable information? How come they don't provide a free text mining service on the web using publicly available databases to evangelize on text mining and grow their user base?

What Temis, and other software vendors, fail to understand, is that large established corporations have a vested interest in maintaining the status quo. Even the most advanced in terms of KM thinking do not want them to rock the boat too much by introducing technologies à la Google, which everybody can use for free and which will give more visibility to "sensitive information". Never will large corporations help Temis grow to the point when an IPO will be possible. If Google people are talking about an initial public offering that could value the company in the range of $15 billion and $25 billion, it's because tens or maybe hundreds of million people use it daily, not because some high-profile multi-national companies bought so many licences per seat.

They don't have enough cash, so they look for deep pocketed customers. Right. This was probably the only thing to do in the dark years after the internet bubble collapsed. But this is very dangerous for the company's future. Now is the time to spend more effort broadening the user base and looking for investors, even if it means emigration to the West Coast, because nothing serious will happen in France any time soon, unfortunately.

No-one has commented on this entry (yet).

Thanks for signing in, . Now you can comment. (sign out)