News of Methbot late last year brought the spotlight back on ad fraud and the urgency with which it needs to be dealt with. The Google-YouTube controversy earlier this year is another example of why the conversation about brand safety is something that every advertiser needs to have. In this context, GroupM’s John Montgomery speaks to BrandWagon’s Shinmin Bali about why automated buying is the logical way forward, the relevance of brand safety, being ahead of fraudsters and more. Edited excerpts:
What is your response to advertisers who ask why brand safety is important?
As digital has become more data driven, we have been able to understand more of what makes it work. We can target behaviourally, we can attribute to the right source and we can generally get better RoI. But along with that comes a lot of complexity which in turn brings with it an amount of risk. Any area of risk in the digital supply chain is something we will mitigate. A lot of our clients are big content producers like NBC and Universal Music which are losing billions to pirate sites. In fact, pirate sites are earning money from advertisers in a way because ads end up showing on pirate sites as that’s where they figure the audience is.
There are different markets at different stages of comfort with the digital environment. Is it then fair to say that brand safety is one of the teething troubles of the digital space?
A year from now, another new risk might emerge. In the privacy area, GDPR (general data protection regulation) in the EU, which is a regulation about anybody sharing data with somebody else, is suddenly going to become a new issue. Sure, contextual brand safety has always been there but after February, there is a level of hysteria around it now. It will only get more complex and more important because it is fairly nascent. Some of it is teething problems, true, but each of these issues will require mitigation and a strategy. It will get better. But just because we know it will get better does not mean we don’t have to keep watching it.
To what extent can digital ad fraud be minimised, say, in the Indian market? And how long can it take?
We can bring it down to the same numbers that we see in Europe. We can bring it down to a low single-digit figure. In the West, the fraud was high but the moment we started measuring it, it dropped like a stone. I don’t know how long it is going to take for us to bring it down to the same levels as Europe, but the Indian market is a few years behind in terms of measurement. Having said that, it won’t take India as long as it has taken for, say, Europe because now the technology exists to counter fraud. The technology we are using in the US is the technology that can work here as well.
Fraudsters usually work by looking into the KPIs advertisers are optimising. How can programmatic help deal with this?
Cheaters are always one step ahead of the chasers. Methbot came out in December, 2016 but it was caught in a few days. It did cause some damage in those few days but it was not huge. That’s why we rely on our technology partners to stay ahead of it. Programmatic has always been given a bad reputation. I think it is unfair. It is a little bit like blaming road accidents on motorcars. It’s about the operator of the vehicle and not the car. It is the same case with programmatic. Often, programmatic is asked to operate in an open exchange and find the lowest cost inventory. As a result, of course, the brand risk will be higher in that space.
But frankly, programmatic is just an optimisation technology. We have used programmatic in private exchanges and private marketplaces because it is as appropriate in a private market as it is in an open exchange. We have actually found in some developing markets that programmatic is safer because you have to measure. In countries like China, where there isn’t a lot of measurement, programmatic proves to be safer.
Which KPIs could be fraud proof? Are volume-based and pay-per-click models still the right approach for advertising?
All modes of buying (impressions, clicks, conversions, views, etc) are subject to fraud. SIVT (sophisticated in-valid traffic) is a highly evolved practice. It is important we use trusted third-party tools to block exposure to all kinds of fraudulent inventory sources.
India still has a fair number of advertisers who do things traditionally. How does a brand come to realise, in its first brush with digital, to consider brand safety measures?
For each day that a brand does not address safety, it is an opportunity lost where the brand could have been more viewable, with better quality inventory. Programmatic is about 10% or less but it will happen. It is one of those step change environments, like the move from analog to digital phones. Take for example YouTube; there are about 450 hours of video uploaded every minute and many hours of video viewed every day. So the only way for any advertiser, new or old, to make sense of this is to do it the automated way.
We have an expression on our side of the business, ‘everything that can be bought programmatically, will be’. If that is true, advertisers are going to need to learn to do it extremely well. In the US, after being talked about for years, programmatic finally realistically took two years to catch on. It will happen in India too. Publishers are probably worried about the yield from programmatic and so they hang on to what reserved media can offer them, but there will be a momentum that will finally tip the scale.
In hindsight, should players like Google and YouTube be penalised for failing to develop thorough systems to ensure the best possible brand-safe conditions?
It was a case of free market mechanism. What was found is many advertisers pausing their advertising because they weren’t sure what was happening. Unless publishers respond to brand safety, they will get less business from big brands that are increasingly concerned about brand safety. So when you say penalised, it will happen through a free market environment.