Digital advertising technology enthusiast Frank Verbist has twenty years of experience in technology development and startups. Six of these he worked at IgnitionOne, leading the team that built one of the first User profiling and Real Time Bidding platforms. Frank is co-founder of early-stage technology investment fund Strike4, and member of the board for mobile marketing automation company TapCrowd. At Argus Labs, he serves as technology advisor, and helps push the company forward, and entering new industries.
Hi Frank, thank you for taking the time for this short interview. What is your current focus regarding the Argus Labs technology?
I am looking into how we best deploy Argus Labs’ mood and moment platform in the advertising industry, taking into account the requirements of the publishers and app developers, demand-side platforms and, of course, the end consumer.
Mood and moment, both sound like they are open for interpretation; hard to quantify. How do you make such information usable for a such a diverse sector?
Mood and Moments are two attributes that allow us to provide better and new insights in the customers journey, and that provide promising guidance for timing the delivery of ads and messages.
Nowadays, most of mobile advertising campaign optimisation is based on location tracking and geo fencing. There are at least two problems with the current approach. The first is about the value of – and in exchange for – location information. Getting an opt-in for location tracking will become very difficult when people start to realise the only thing they get in return for 24/7 GPS logs of their daily activity is randomly timed targeted ads. The second problem is that location is just one data point in a customer’s journey, and probably not even the most important one – but definitely a sensitive one. In order to get a holistic view on a customer’s journey more data layers need to be combined. The fact that an ad is about a shop nearby, does not necessarily make it relevant to you at this moment.
Once the EU data protection reform passes, there will probably be a third problem with going full-on location; legislative compliance. But for now, where you are is at the base – or even the sole data point – of most self-proclaimed ‘context-aware’ advertising solutions – because it is that easy to obtain and use given current technology and standards?
It is true that location is today probably the easiest type of data to collect, from IP, from GPS, localisation settings, .. . Phones are virtually broadcasting location – with or without the user being aware of it – and real time bidders are equipped to receive and interpret this data. But location does not help you in deciding what type of content to deliver and when. It is not because someone is at 51.20 latitude, 4.39 longitude – with a 0.32 accuracy – that he or she is in the mood to consume your 2 minute video ad.
The opportunity lies in having a flexible platform that will allow us to generate the right level of abstraction for specific campaign optimisation problems using the available data, and in having an almost effortless way to get the needed data to our platform, our mobile sensing SDK.
You have always been involved with ‘next generation technology’. Do you feel that the advertisement industry is currently using all technical advances to its full potential?
I believe digital advertising is on the verge of big changes, during the last years a lot of energy went into ‘real time bidding’ and ‘programmatic buying’ for direct marketing and brand campaign optimisation. The data points used consist mostly out of simple pageviews, conversion tracking and some social network data. This already produces massive amounts of data and requires hefty processing infrastructure, but the complexity is not comparable to what happens when we throw mobile sensor data into the mix. This is where I strongly believe Argus Labs is able to make a difference.
You say ‘mobile sensor data’. Does this mean usage of the moods & moments is limited to mobile channels?
Absolutely not, one example is when you start to aggregate this data across audiences, locations or even families a new dimension of insights becomes available. Imagine what you can do in the area of recommending music, movies or shows to watch. Your evening Netflix recommendation might be very different after a busy day at work compared to one for the whole family after a relaxed shopping afternoon.
Across channels, how up to date is your information? I’m feeling much better than I did this morning, could that be reflected in the content I get see now?
As everything today in digital advertising, updates need to be real-time. Mood is something that evolves over time, it might lag a bit due to its complicated nature and the multiple reference points you need to have a significant detection, but our models take into account what is the most likely way for it to evolve. This means it might not always be completely real-time, but it will be useful. And then, the underlying platform is built to handle events in real-time right down to advertisers RTB infrastructure.
As a person, I don’t feel static either. Last year I was interested in running, currently I’m allocating more effort – and budget – to my garden and travel plans. I have a really cute godson now, so became interested in baby stuff. I no longer have a functional iDevice, which means iGadgets are no longer relevant to me. How do you keep the profiles up to date? Do you have a hundred behavioural scientists hidden away in a basement somewhere?
This is where current machine learning techniques fall short, multiple layers of data need to be correlated to even begin to get an understanding what is going on in real life.
We bet on deep learning – self-learning – techniques to come to the rescue and to provide us with better insights into a users profile and finding non-obvious correlations between events and effects.
Are people OK with companies holding that much data about them?
On privacy, as a user, I will stop sharing data if they give nothing in return. Towards advertising this means I expect them to deliver a better service. They should become more like an emphatic advisor, don’t just throw anything at me at any possible moment in- my – time. They need to understand my interests and habits and timing without me having to carry a webcam on top of my head.
My smartphone holds more then enough information about me to know about what I am up to at any moment in time. Mind you, this does not mean he should broadcast it in great detail to the rest of the world – if somebody makes sense of it and provide me a meaningful service in return, I am in.
What would your dream context-aware ad say?
I don’t care that much, as long as it is relevant. Let’s keep it simple: 10 relevant ads per day instead of the 1000 cluttering my screen every second of the day. Let the highest bidder win!
That sounds like a digital dream come true!
Argus Labs will be introducing its expanded advertising industry product offering at Ad:Tech London, October 21st-22nd. Presenting at the Engagement & Experience summit, we will be bringing you up to speed on some innovative and non-intrusive, privacy-compliant sensing methods you can use for delivering the right messages with respect for your customer’s life rhythm.
If you are attending (the Ad:Tech event is free, register here) and want to discuss the data-rich future of mobile marketing and context-aware advertising , please do reach out, via email@example.com, or @arguslabs on Twitter.