In-app video advertising can be a risky format for advertisers; the ease with which anyone can build an app and distribute it via Google’s Play Store or Apple’s App Store means a lot of inventory might be problematic from a brand safety perspective, or may not run ads properly. Here Valtteri Pukander, CEO of Finnish in-app advertising specialist Thirdpresence explains how his company uses AI to reduce this risk and match advertisers with the right audiences, and why Finland has so many mobile-centric tech companies.
Could you provide an overview of Thirdpresence’s business today?
In 2007 the company started as a provider of Telecom Carrier mobile-TV solutions. At the time the whole system was very server-centric. The technology and service was built to support the first Nokia TV -mobile phone and on DVB-H (digital video broadcasting – handheld) technology. We promoted it together with Nokia-Siemens-Network to Telecom Carriers. Eventually, DVB-H didn’t succeed and we had to close down that business.
Around 2010 we moved everything on to the cloud and supported OTT (over-the-top) companies, e.g. we built the first mobile catch-up TV platform for Finnish Media Company Sanoma. Soon we realised that the most scalable part of the business was mobile video advertising and we released our mobile video advertising SAAS (software as a service) platform in 2013. It was quite well received and we managed to acquire over 100 clients globally by 2015.
After having worked for seven years as a tech provider, at the end of 2015 we took the decision to build an in-app video offering for programmatic brand advertisers. Today we curate billions of daily ad requests and use AI to improve ad relevancy within the in-app space.
When you say ‘AI’ – what are you referring to and how do you use it?
Over the years we have built our own tech stack and are now implementing more intelligence in contextually recognising the right audiences to match with the right advertisers. By doing so, we aim at reducing ad waste for the demand side and improve the relevancy for in-app audiences. We only pass some 10 percent of ad requests we receive on to our advertisers, with the main reasons for rejection being fraud, non-performing ad placements and brand-safety issues.
We have applied machine learning algorithms for the programmatic space, but the end goal is to have purpose-built AI that controls the full scope from contextually categorising the ever growing in-app space and feeding the relevant impressions to buyers.
Why are you focusing only on mobile in-app inventory?
Initially we had two key reasons: in-app being still relatively new for brands and us being located in Finland. There is a large mobile gaming ecosystem in Finland, with over 400 mobile games companies based here. Today we cover other verticals as well, such as news, utilities, directories and entertainment. As such, we think that in-app provides a good environment for brand advertising, e.g. our MOAT-score in 2017 was 96 percent.
What type of advertisers do you work with?
We are a pure programmatic player working with brands directly, agency trading desks and platforms.
So many of the most successful mobile gaming companies have come out of Finland – Supercell (now owned by Tencent), Rovio Nextgames and Fingersoft to name just a few. Why has mobile gaming flourished so much in Finland?
I think it has a lot to do with Nokia. Even if Nokia was not a games company, many people entered in the “mobile mindset”. Then came Rovio with Angry Birds, and after that lots of new gaming companies were established to follow Rovio’s success.
What’s your view on rewarded video advertising, where gamers are given incentives to view ads?
A couple of years ago there was some stigma among brand advertisers with the concept of “rewards”. It has a somewhat negative tone and was compared to incentivised traffic. Nowadays I would say it is the opposite: brand advertisers are getting single-view, audio-on ads with almost 100 percent completion rates. Also a number of studies show that players value rewarded ads.