How Twitter hopes to assure marketers that clicks on its network are real

Measuring human traffic

With advertisers estimated to lose over $7bn this year (in terms of media budgets siphoned off by fraudsters generating false impressions on ads) Twitter is one of the latest tier-one online media owners to move to tackle the issue.

Earlier this month, Twitter’s ad exchange MoPub struck a deal with anti-fraud outfit Forensiq in a bid to reassure advertisers that clicks on its mobile ad exchange are generated by human beings and not bots.

Twitter’s MoPub marketplace sits within the social network’s ad tech stack – an increasingly important part of its business, especially as growth of its overall user numbers begins to plateau – and connects media buyers and sellers to trade in-app mobile inventory using real-time bidding (RTB) technology.

Through the partnership between MoPub and Forensiq, Twitter hopes to improve the “supply quality” across the more than 45,000 mobile apps, and over 450 billion monthly ad requests on the marketplace.

This is a growing sector of the business for Twitter, with the social network provider claiming that over 115 per cent more advertisers accessed the service between 2014 and last year.

The partnership also means that advertisers on the social network can more confidently claim their media investment with Twitter’s ad tech offering is paying dividends, and that their ads are being seen, and acted upon, by human beings not bots – commonly referred to as non-human traffic (NHT).

It is also an example of MoPub’s three-prong approach to supply quality, which combines human moderation, its own technology (plus that of third parties) to help ensure high-quality inventory for buyers in our exchange.

Commenting on the partnership, Janae McDonough, vice president of MoPub at Twitter, added: “We want to give advertisers the confidence that mobile in-app is comprised of high quality supply — and we’re thrilled to have this partnership in place to ensure this.”

Ad tech's increasing importance to Twitter

With the growth of its user numbers starting to stall, and advertising revenues coming in at the lower end of expectations ($595m in the last quarter, numbers that failed to satisfy investors) the company’s board was eager to point out the role of ad tech during its last earnings call.

Twitter chief executive Jack Dorsey penned a letter highlighting the developments of Twitter Fabric (a suite of tools it makes available to advertisers, including MoPub), with the letter also containing a strong emphasis on measurement and demonstrating ROI to advertisers.

Notable in the letter were mentions of tie-ups with Google’s DoubleClick Bid Manager, Nielsen, as well as a similar tie-up with Moat to measure the quality of video ad impressions on the social network.

A wider look at fraud

The recently announced partnership also tops a host of new clients wins for Forensiq, which specialises in transparency issues such as fraud and viewability, etc. Such issues are increasingly on the agenda for brands, particularly those attempting to generate app downloads.

Julia Smith, director of communications at Forensiq, explained that a lot of performance clients (i.e. advertisers that pay their ad tech partners based on whether or not their activity generated a sales lead, or direct download, etc.) are increasingly interested in anti-fraud products.

“It’s an easier sell to them, as you can say: ‘we’re saving you money’,” she said, adding that Forensiq has observed much more fraud on campaigns where clients (typically gaming clients) optimise to generate app downloads.

Given the vast swathes of traffic involved in the ad tech space it can be different to detect, isolate, and root out fraudulent traffic – this is often what attracts such criminals to prey on this sector of the industry – but Forensiq has developed a number of strategies to detect it, particularly on mobile.

Is in-app traffic susceptible to fraud?

Hannah Beanland, Forensiq’s director of technical operations, explained some of the methods the outfit uses to signpost suspicious activity, which it then uses as a signal that may warrant further scrutiny.

For instance, if a bid on an ad impression takes place on a different IP address as the eventual conversion, then this would warrant additional inspection. “Ad fraud is becoming increasingly prevalent in the mobile app space,” she said.

“It’s easier tp commit fraud in the in-app space, as you can’t use Java all over it. So with in-app and mobile it’s easier [for fraudsters] to put a few proxies in place [which oftentimes blur the lines, in order to avoid detection],” added Beanland.

Beanland further explained the different measures that can be taken to detect mobile ad fraud on the in-app space. These include examining the speed at which ads in an app refresh.

For instance, a typical ad refresh rate in an app is every 30 seconds, but fraudsters often programme their bots to generate as much clicks (i.e. cash) in as short as period as possible, so the ad refresh rates generated by bots will more likely be every five or 10 seconds. So when Forensiq detects such activity it makes a deeper inspection, this method is often referred to as inspecting the ‘refresh rate velocity’, according to Beanland.

Wider moves to tackle ad fraud

Tackling the issue of NHT is increasingly an issue for advertisers, a fact underlined in a study by US-based trade body the Association of National Advertisers (ANA) in a report entitled: “Sourced Traffic: Buyer Beware”.

Only 61 per cent of advertisers are either "slightly familiar", or "not familiar at all" with where their traffic is sourced, according to the trade body which earlier forecast that advertisers would lose $7.2 dollars to fraudsters this year.

Meanwhile, only five per cent of advertisers surveyed said they are "extremely familiar" with the practice, according to the study which quizzed hundreds of marketers.

“The concept of sourced traffic or traffic sourcing (any method by which digital media sellers acquire visitors through third parties) is unique to digital advertising. Marketers investing in digital advertising need to understand sourced traffic, the issues associated with it — including lack of transparency and high levels of bot fraud,” according to the report.

The industry units under TAG, but key names are missing

This was further followed up by the fellow US-based outfit Trustworthy Accountability Group (TAG) which this week announced the launch of its cornerstone anti-fraud certification program, which will award TAG “Certified Against Fraud” seals to buyers, sellers, and intermediaries in the digital advertising supply chain.

More than two dozen leading digital advertising companies and ad agencies have already agreed to participate in the programme and undergo TAG anti-fraud certification, among the 24 companies named in the announcement were: AppNexus; BrightRoll (Yahoo’s ad tech stack); ComScore; Forensiq; Integral Ad Science; MediaMath; Moat; OpenX; Rubicon Project; Sociomantic; Sovrn, and SpotX.

Mike Zaneis, chief executive of TAG, added: “Participants in the digital ad supply chain can now ask a simple question to tell if their counterparties have taken the necessary steps to fight ad fraud: ‘Are you TAG Certified Against Fraud?’ As more TAG anti-fraud seals are awarded, the cracks in our industry exploited by bad actors will also be sealed against their criminal endeavours.”

Also named in the TAG announcement were some of the leading agency networks including Omnicom Media Group, Publicis Groupe, and WPP. But notable by their absence were some of the industry’s biggest names - Facebook and Google*.

Sources in the industry claim that while Google does offer some sort of support for wider industry moves to address issues such as fraud, it is inherent in the DNA of the company not to be beholden to standards from outside parties.

Do UK advertisers lag their US counterparts in tackling the fraud issue?

Earlier this month, JICWEBS – a UK cross-industry trade body with a similar remit to TAG - issued further guidance to its constituents on how to tackle ad fraud and that it will issue seals of approval to ad tech companies that adhere to its guidelines later in the year, similar to the move just made by its U.S. counterpart.

However, speaking at a recent Brand Safety Summit, hosted by 614 Group in London, the IAB’s Steve Chester, told attendees the industry’s fight against those committing fraud is simply not backed well enough in terms of finance.

Unless the advertising industry is prepared to better fund units of the police dedicated to disrupting fraudsters - similar to other industries, such as the financial sector- their activities are unlikely to be seriously affected he told attendees.

“We’re going to have to get the industry to collectively fund this thing. To make a blunt point: we as an industry will have to stump up a lot more money. At present we virtually put in nothing in terms of protecting this industry,” he warned attendees.

*A representative of TAG has since contacted The Drum and highlighted that while Facebook and Google are currently not certified under the latest scheme, both outfits do provide support for its efforts by participating in its various councils. etc.

Join us, it's free.

Become a member to get access to:

  • Exclusive Content
  • Daily and specialised newsletters
  • Research and analysis

Join us, it’s free.

Want to read this article and others just like it? All you need to do is become a member of The Drum. Basic membership is quick, free and you will be able to receive daily news updates.