Does programmatic encourage clickbait and fake news?

Scott: Print, TV and radio numbers are independently audited, but ad networks such as Facebook and Google largely don't let that happen. We saw a new lawsuit in the US last week alleging that Facebook committed fraud by not disclosing bad video ad view data for more than a year. We have seen other data corrections in the past. Should we trust Facebook's numbers ever again?

Beumala: Facebook has given bad viewability data, but my question is, “who has given good viewability data?” Viewability and measurement in the advertisement world are screwed up all over the place – not only with Facebook.

Indeck: I think people aren't sceptical enough about data. We do have to use our minds while we're looking at things and think, “does this make sense?” As advertisers, marketers and publishers, we have to have a healthy dose of scepticism when we're using any data.

Bidel: As an industry, if we really want to mature, we need to have standards that everybody abides by so that everyone is comparing apples to apples and working on the same playing field. Without that, I think it's just never going to mature into the kind of business that marketers and publishers need it to be.

Scott: But who sets the standards, and who enforces them?

Beumala: So, we have the IAB, which has been consistently trying to get a definition of what “viewability” means. It's just now for the last two or three months that we've got Google coming into an agreement with Twitter, defining viewability pretty much in the same way. Facebook now looks that it's kind of getting closer to this, but viewability is something strange. I think that it's not only a tech decision, it has to come as well from marketers. Let’s take the desktop. So you have a big screen, and they say that as far as there is three seconds of the full video that can be seen on the screen, that is viewable, right? That is one of the official definitions of viewability. But compare a very tiny video on the bottom-right corner compared to a massive screen – this really means that it's viewable? There are different descriptions all over the place – they need to sit down and agree.

Scott: In 2016, Facebook's VP for Europe, Nicola Mendelsohn, said “we’re seeing a year-on-year decline in text". Mark Zuckerberg also said "five years to all video". Now, we see that Facebook allegedly inflated video ad views by up to 900%. So, for publishers, was the “pivot to video” a big mistake? Should they now pivot from video?

Indeck: I still think one of the big growth areas for publishers is video. We didn't ever really have a big play on video on Facebook. We have a presence there because you have to – it's Facebook – but that's not the place we're looking to monetise or even get big reach because you can't. They limit it. But I still think video is very much going to be a part of publishers’ businesses moving forward – you can't ignore it.

Beumala: It's easy to say that Facebook was inflating the numbers, but I know lots of publishers that have made a ton of money with those videos. The market collapsed a little bit because everyone was producing video, so things got worse, but people have made a lot of money with video.

Scott: Are are there certain sectors or verticals that might be more interested in video than others?

Beumala: Two very concrete cases with millennials and young generations – media was all over the place, producing viral videos as well as fake videos and all this kind of content. It's really catchy. The cases that I know that have been producing money are those cases.

Scott: BuzzFeed News reported a multi-million dollar ad fraud scheme involving more than 125 websites and Android apps. Whenever machines make advertising or monetisation decisions, there will always be some level of fraudulent activity. Today, what percentage of visits, impressions, clicks and purchases are from bots or other bad actors?

Chimeno: In 2016, the amount [of lost revenue in the US] was $14bn, and it was 16% of the digital ad spend in the country. The forecast for 2022 was an increase up to $44bn – 30% of digital ad spend.

Scott: So, in 2019, how should publishers minimise the effects of that fraud?

Chimeno: The most important decision so should come from the advertisers because they decide the kind of campaigns they would like to deliver. You can buy really cheap, but you are at risk of being a victim of ad fraud. Advertisers should also take care with their teams and technology because if you don't know the tools you can use to fight against ad fraud and keep good brand safety [and] if you don't have a team who knows these tools, [you can also be a victim].

Bidel: Publishers need to be honest with marketers about what they can organically deliver. If you are buying an audience, you lose control – and that's when you compromise your own environment.

Indeck: You really have to vet your partners – what networks and what PMPs and who you're working with to ensure that they're quality and up to snuff. If you have a good partnership and you have a level of trust with them, that really helps eliminate a lot of it. But agencies really have to step up in their commitment to ad tagging so that it limits the ability for fraud to happen.

Beumala: There is something else as well – having a good pricing strategy on the website and avoid selling extremely cheap impressions. That eliminates a lot of the noise.

This has nothing to do with the publishing side, but I think that there is something about the platform. There is a big difference in terms of ad fraud on the native application ecosystem compared to the web ecosystem, and the reason for that is the browser and the main difference is that native applications have access to a unique identifier that the browser doesn't have.

It's strange because Google would have a massive interest in that. Technically, it's easy. But there is like a bias towards making it happen or not. I think they should be answering these kind of questions.

Scott: Countless studies show that your average person – not anyone working in marketing – does not want to be tracked and targeted by ads. They don’t want their personal data collected. As a result, the use of adblockers, script blockers and VPNs increases worldwide every year. In such an environment, how can publishers monetise more effectively?

Beumala: There are some anti-adblocking tools you can use. When this tool detects a user using an adblocker, it asks the user to deactivate the adblocker.

Scott: But how many people do that?

Beumala: I suppose it will depend on the quality of the content. If you are really interested in reading the content, you will deactivate.

Chimeno: The conclusion that we made is that it depends on the aggressiveness of the ad setup that you put as a publisher. As a user, if I'm not getting inundated by ads, I'm fine with [deactivating blockers]. I can understand that publishers need to make money.

Scott: Is there a point for publishers that after a certain number of ad units per page, you see a drop off? Can you recommend that publishers have no more than a certain number of units of ad units on a page?

Beumala: At some point, it's not worth it economically to keep making impressions. If you keep making impressions, the user will get annoyed with you because it's going to be low-quality ads, not targeted ads.

Indeck: I think it's almost going back to what publishing was to begin with. It's offering quality, relevant content and not inundating the users with advertising. Give them a quality experience. I think we forget that sometimes.

Scott: So, would you say that websites today should have nothing but banner ads like a newspaper?

Indeck: No, because I think there are some really nice integrations that are not traditional and not intrusive. They're still relevant to the experience, but they're not cumbersome and to the actual user experience.

Beumala: There is a big challenge in the span of attention that we get. You get a visit from Twitter, you know that this guy is gonna go in 30 seconds. The challenge is to make that visit worth it in terms of economics.

Scott: Is there a place in programmatic for ads that don't track people or collect their personal data?

Beumala: Lots of times, it's not personal data – it's data that we infer statistically. I know that with a probability of 80%, you're a man. Statistically, that data might be relevant or not. I prefer to be targeted, but again that is a user preference – and as far as I know there is no way to stop that from happening.

Scott: Susan, one thing that you said earlier was that digital ad spend has now surpassed TV. I've seen data showing that most digital spend is going specifically to direct response. It's not going to brand advertising. Is there a place for brand advertising online, or is it all direct response?

Bidel: I think that the old distinctions between brand and direct response are morphing in the online environment. The way that marketers think of the audience is that there are consumers that they want a response from now or at some point in the future. There's definitely a place online for brand advertising. For publishers, there better be because that's where the biggest portions of marketing budgets really are.

Scott: Have you seen any good examples of brands being built online?

Bidel: Warby Parker. There are a whole bunch of brands that have built themselves from the ground up online.

Indeck: Airbnb.

Bidel: Uber.

Scott: In 2018, P&G cut digital spend by 50% and saw a 2% increase in sales Taco Bell moved spend from digital to TV. So did Turner and several other large companies. Why do you think this is happening, and is this a trend?

Beumala: This is something really notorious in the media. I think digital transformation is a fact in the world, and it's also a fact in advertising.

Indeck: I think there definitely was a race to the bottom of people trying to get the most volume for the cheapest CPMs, and that's not how way we create our product. We believe in our product. We create a high-quality product so I think it's a good trend if people are reevaluating and marketers are saying they want better quality because we want the same thing.

Bidel: I think P&G had been experimenting as close as they could get to one-to-one targeting. They don't have the kind of data to be able to identify on a one-to-one basis who their consumers are. Second of all, for a lot of their products, basically the consumer is anyone with a pulse. So they did that experiment and they pulled back, and they're trying to go back to classic CPG marketing strategies, which is to go for reach.

Scott: One of the selling points of programmatic is that you ignore publications and target individuals to get the cheapest results. If John Smith goes to the BBC and then he goes to some fake news website, programmatic may target him only on the fake news website because it will be cheaper. Does programmatic encourage clickbait and the worst the internet has to offer because it pushes prices down?

Indeck: I think it does if it's not managed well. If it's a programmatic setup that's managed well there are things that you can put in place that will eliminate verticals or sites. That takes manpower. You can't just rely on a computer or AI to do all the work. You have to set up your environment in the right way so that you're eliminating environments or types of advertisers that you don't want. It takes a lot takes work. You can't just rely on a computer to do it for you.

Bidel: I mean the pipes certainly facilitate monetising fake news, and I think that that has to be monitored and eliminated at the exchange level and at the SSP level. But they have to be discriminating about the inventory they offer and be known for offering quality and being partners that are to be trusted.

Scott will publish the video of the full hour-long conversation later.

The Promotion Fix is an exclusive biweekly column for The Drum contributed by global marketing keynote speaker and workshop facilitator Samuel Scott, a former journalist, newspaper editor and director of marketing in the high-tech industry. Follow him on Twitter. Scott is based out of Tel Aviv, Israel.