Skip to content

Startups: How do you get along with machine learning?

Find interesting and promising startups in any market. Automatically and within seconds. The Leadspicker project of friends and graduates of the Brno University of Technology Vlastimil Vodička [on the introductory photo on the left] and Jan Skácel uses a machine-based algorithm to detect even beginning local startups.

Both large international companies and investment funds are interested in Leadspicker services. With over two hundred and fifty clients in thirty-six countries, the company is now expanding to America. In addition to specific search, it also offers clients pre-paid access to its databases and its own platform for addressing and evaluating the searched companies.

Save unnecessary work

The effort to save unnecessary work. This was, according to Vlastimil Vodička, the initial motivation that eventually led to the founding of Leadspicker. The company ranked seventh in the Czech Republic and Deloitte CE Technology Fast 50 in the Rising Stars category last week. The idea originated four years ago in Vienna, when Vodička was tired of doing the same thing every day.

“For one company I was looking for new companies and technologies around the world. Especially in markets where they had no idea what startup was. The procedure was simple. I typed keywords into Google, searched through search results, databases, articles, and searched for mentions of the right businesses to match. I wrote them down in a table, addressed them, and prepared a list for the R&D department. It was a process that lasted for several weeks, it was repeated and only keywords changed”, Vodička said, wondering if some parts of his work could be automated.

He turned to his long-time friend and graduate of the Faculty of Information Technology of BUT Jan Skácel who worked for Seznam at that time. “Honza focused on both automation and artificial intelligence. I have described what I do in each step. He subsequently identified those parts that could be automated. We tried to gradually integrate technology into the whole process and I eventually reached a stage where I managed many more projects in less time and with better results. We thought that something like that could be of interest between accelerators and larger companies. So we became independent and started offering the service, ”Vodička explained the beginning of Leadspicker.

Since then, nearly twenty new colleagues have joined Vodička and Skácel, and Leadspicker services are regularly used by some 200 companies in thirty-six countries around the world. “Our services are divided into the search and automation of certain processes. We wondered how best to find a new business that no one else knows yet. When you start up a startup, you usually create a website for it. On LinkedIn he says he’s CEO, CTO or something like that. Facebook will be added to certain groups and will follow a certain type of content. We decided to search people with similar profiles all over the Internet. Our tool crawls websites, social networks and startup databases and compares their content with our database of existing companies. When he discovers something that resembles another company from the database, he points out that it is probably a new startup, ”explained the main principle of the Vodička algorithm.

Leadspicker analysts then review the information and indicate to what extent the search was correct and useful. Thus, the neural network is constantly improving its computational model and refining searches and recommendations for searched startups in a specific industry.

Success will not estimate the best algorithm

So far, Leadspicker has only used the tools internally and has already delivered a complete service to clients such as PwC, Thomson Reuters, Singapore Airlines and Axel Springer. Companies have been given access to a platform where they can review the final results and evaluate selected startups.

“This service still accounts for most of our revenue. At the same time, however, we are gradually moving towards greater automation, so we managed to launch a Scouting platform last year, which clients can use themselves, ”added Vodička. .

“Originally we started with startups because it was an area on which I focused professionally for a long time. However, our tool can be used to search for any business and technology in any industry. Therefore, we are increasingly turning to us to find new clients or business partners”, Vodička pointed out.

Improving the look and feel of the new data portal

The company now focuses primarily on improving the look and feel of the new data portal. And also to the expansion into the United States. “We have managed to attract several American clients from Europe. At the same time, it is a very specific market and we have said that we need to be right here”, said Vodička, who is now mostly in the United States, looking for new partners and an office in addition to new clients.

They do not rule out that they would start investing in promising startups that nobody knows yet. According to Vodička, however, nobody can estimate success at the beginning. Not even the best algorithm. “In the beginning you never know what will be good. I am very careful from experience. We have seen great technologies, but they have failed. The team did not make it, the market was not ready. And we also saw the biggest banalities that we believed wouldn’t work. Many of them celebrate international success today. Like Skinners. Years ago we discovered them and thought it was a simple sock without sensors and software. In the finals they are absolutely unreal. Therefore, in the near future, we will measure the so-called growth factor for startups on our platform. As a result, we can see when a business starts growing faster, or that it’s unique in something and growing interest in it. But we can not predict who will be the new Google or Facebook. The only thing is certain that the new Google will no longer be similar to the existing one and the new Facebook will be completely different. So far, we can only imagine successful companies of the future”, concluded Vodička.

Also published on Medium.

Published inStartups
%d bloggers like this: