Das Hamburger Targeting-Unternehmen Semasio weist ein paar interessante Faktoren auf, um es einmal genauer zu betrachten. Innovative Idee, der Gründer ist Skandinavier (wenn auch kein Schwede) und hat mit , dem High-Tech Gründerfonds (www.high-tech-gruenderfonds.de), Bertelsmann Digital Media Investments (www.bdmifund.com) und Project A Ventures (www.project-a.com) erfahrene Geldgeber an Bord. Gründer Kasper Skou im Interview mit Gründerszene.
Kasper, please introduce yourself.
I am a Dane, Master of information science from the University of Aarhus and moved to Hamburg to work for ), the veteran of behavioral targeting in Europe, which I had done a project for while finishing my MBA at INSEAD south of Paris. After working three years for Wunderloop, I realized that we needed a completely new approach to profiling and targeting, which eliminates the subjective factor, i.e. the human interpretive element, from the digital marketing process.
I contacted the CTO I worked with at Speed of Mind, a semantic search engine company I was the CEO for before going to INSEAD, and together we developed a completely statistical approach to profiling and targeting utilizing the semantic methodology we had perfected at Speed of Mind.
In November of 2010, High-Tech Gründerfonds together with minor investors completed the seed round, and we began development of the platform.
We chose to do our more mundane development in Kiev, Ukraine for cost reasons and the very short lead time to get highly qualified employees, while we kept the really advanced development of the semantic and statistical engines in Denmark. The commercial headquarter is Hamburg, as this is still a major media city with many online advertising players.
„Instead of associating each page a user consumes with a category in a pre-defined category system, we extract all the significant terms and phrases in the page and store those in the user profile.“
What does Semasio do?
Semasio (www.semasio.com) takes a completely new approach to profiling and targeting using semantic technology to make sense of the content the user consumes. Instead of associating each page a user consumes with a category in a pre-defined category system (which is created by humans and is thus subjective), we extract all the significant terms and phrases in the page and store those in the user profile. That way we do not pre-select the information to store concerning the user’s contact with a given page, but instead store all the information in that contact.
Once we have a goal to achieve (finding users that will sign up for a test drive in a BMW 5-series, users that are male, users that have a great interest in a lifestyle of health and sustainability etc.), we statistically determine which information is relevant and discriminatory for setting the users in the target group apart from those outside it.
This service we offer to buyers and sellers of online advertising, who want to do this more intelligently. For the buyers of online advertising, we offer them the possibility to create the exactly right target groups for their customers, the advertisers, without having to share sensitive and valuable information of these advertisers with the sell-side. The buyers can then purchase these target groups over Semasio’s DSP, which contains advanced optimization based e.g. on user, place and time, or via any 3rd-party DSP they currently use.
Sellers of online advertising get the possibility to not only set up the standard target groups the market expects, but also – via our semantic profiling technology – to set up any target the buyers might require in a matter of minutes. Additionally, our optimization technology brings together user, place and time to help the seller optimize the yield from non-premium inventory.
In a previous interview you said that old targeting costs where equally as high as their benefits. What do you mean by that and how does Semasio’s targeting differ from previous paradigms?
I was quoting a venture capitalist I had recently pitched to, who summed up the last decade of targeting technologies in this way. The way I understand his comment is that until now, targeting companies have not created more economic value than they have tried to capture. The increase in RoI has not been able to cover the costs from running and developing these technologies.
„Semasio is collecting up to 1.000 times more data per user without increasing the total cost.“
Semasio is changing this by collecting up to 1.000 times more data per user per month without increasing the total cost per user per month. This is because new types of hardware have become production ready, and Semasio has deployed them to effect a completely new level of information processing without increasing the cost base.
By preserving much more information about the user’s behavior, Semasio has the potential to increase the RoI for buyers and sellers without increasing the cost base and thus the value it needs to capture.
How do you think this will change the display advertising eco system in general?
I am a Scandinavian at heart and have a social streak. Thus I believe the buy- and sell-side of the market should come together and realize that if they intelligently pool the information they have about the user in moment of the ad impression, the economic value creation will be maximized – along with the user’s experience of relevance – and this will benefit all parties.
My vision for Semasio is to become the platform on which the advertiser, the agency, the marketer and the publisher can all safely put their user data into the ad impression, knowing that their data stays their property and are not shared with anyone while at the same time driving maximum value creation in the ad impression.
Our responsibility to is to effect this maximization, and that is why we have created this fundamentally new approach to profiling and targeting.
You attracted HTGF, BDMI, Holtzbrinck Ventures and Project A as investors. How did you convince these big four of investing in Semasio?
Everyone would like targeting technology to work, but if we are to believe my VC friend it hasn’t for the last decade. I told my investors why I think that is and why our approach will change that, and it made a lot of sense to them.
They all have ample industry experience to realize what it would mean to really ‘crack the targeting code’ – the areas of application and value creation are almost infinite in an online advertising value chain that is becoming increasingly digital, data-driven and real-time.
We are extremely happy we were able to attract these investors, as we will not only benefit from their experience and insights, but also directly from their ability to help us bring this new technology to market.
What are your plans for the next twelve months? More products or more detail to your existing ones?
We recently signed contracts with a trading arm of a global media agency group, and premium Vermarkter with several more in the pipeline. This means we have a very healthy customer pressure to realize the full vision of the User Intelligence Platform, utilizing the full potential of the semantic paradigm.
The next 12 months will be focused on a very customer-centric process to perfect the main functionalities of the system: pre-defined targeting (on age, gender, interests etc.), performance targeting (custom target groups created automatically by the system to achieve a specific marketing goal), audience insight (giving clients analysis insights about users, sites, pages, campaigns etc.) and offline translation (taking the empirical research from the system and bringing it to other marketing channels such as TV, print, outdoor etc.).
In parallel we will constantly be developing the statistical engines that process the massive amount of information gathered from the semantic profiling approach. This is why we are expanding in every area of our development organization – we are at the end of the day a technology company.