Fighting fraud, the smarter way

Mobile Marketing - Sponsored by: Adjust

Mobile Marketing talks user acquisition and ad fraud with Christian Henschel, co-founder & CEO (left), and Paul H. Müller, co-founder and CTO (right), of Adjust.

Adjust co-founder & CEO Christian Henschel, and co-founder and CTO Paul H. MüllerIt’s well recognised that the app market is a tough one. Whether you’re trying to find new users for your app, or trying to keep those users loyal and engaged once you’ve onboarded them, there are thousands of other apps competing for consumers’ eyeballs and all use pretty much the same channels and tactics to try to win them over.

There’s no shortage of companies out there claiming to be able to help. Some are more credible than others, that credibility stemming from the number of years they have been doing what they do, and the calibre of clients they have managed to attract.

In both these respects, Adjust scores highly. The company was founded in 2012, principally to help app publishers with attribution and measurement, and these remain at the heart of the company’s offering today.

Client roster
From a client perspective, the company also has a very respectable roster. Its tech is used by companies such as NBCUniversal, Zynga, Robinhood, Pinterest and Procter & Gamble, along with around 25,000 others, in their apps.

Co-founder and CEO Christian Henschel says the company is driven by a passion to give its clients the edge, which means the search for other companies with smart tech that it can integrate into its platform is a never-ending one.

“Besides what’s on our roadmap, we are constantly looking at opportunities to acquire companies that fit our vision,” he says. In the past 10 months, it has made two such acquisitions. The first, in December 2018, was Acquired.io, a company which automates a lot of the time-consuming, manual, repetitive tasks involved in running hundreds of mobile ad campaigns across hundreds of channels. So how did the Acquired acquisition come about?

“The first discussions were actually initiated by a customer of ours who was working with them,” says Henschel. “The customer said they were a great company that was making their lives so much easier and that we really had to talk to them, so we did.” Henschel knew right away he was on to something.

“Through the Think Tank (our focus group, which brings marketers together to discuss innovations in the industry) and Mobile Spree events, Adjust’s own branded conferences that we host around the world, we meet around 1,000 customers every year and one thing we understood from talking to them was that they spend around 30 per cent of their time just adjusting and monitoring campaigns. This is a very time-consuming process, logging into multiple dashboards, and with some companies running as many as 1,000 campaigns at the same time, it’s a tall order for the human brain to constantly make adjustments to campaigns and creatives in a very manual way. Machines are better at this than humans, and that’s what the Acquired tech delivers.”

So how does it do it?

Secret sauce
“The secret sauce is that they took a step in a direction the market had not gone before,” says Adjust co-founder and CTO Paul H. Müller. “All the partners you buy from require you to constantly log in and out of their dashboards; it’s a very manual process that no-one had ever challenged before Acquired came along. Their tech automates a lot of the manual processes so that the human operator does not spend all their time entering their decisions into the machine. They just make the decision, and then the machine enacts it through simple automation.”

Adjust’s first step with Acquired was to take the tech and run it off Adjust’s robust and scalable infrastructure. Since then, the company has been developing the solution further, and the Acquired team has relocated to Adjust’s global headquarters in Berlin from Saint Petersburg and San Francisco.

A month after the Acquired acquisition, Adjust bought another company, Israeli start-up Unbotify. The cybersecurity and AI startup was named by Fast Company as 'Israel's most innovative company in 2017,' and listed as one of Gartner's '2018 Cool Vendors in Advertising'. Müller explains the background to this deal. He says:

“We started building the prototype of our Fraud Prevention Suite in 2015. Now the ideal scenario with ad fraud is to prevent it completely, but this is very difficult to do. The fraudsters are very clever and will do what they have to in order to protect their income. It’s a huge industry, worth hundreds of millions of dollars a year, and they are not willingly going to give that up, so they have developed some very sophisticated tech and types of fraud that are almost impossible to detect.

“For example, you would see what, for us, looked like a real user on a real device using a real app. In fact, it was a real device and it was the real app, but it was being used by a bot.”

This type of activity could be related to ad fraud, or, as Müller explains, other types of in-app activity. Up until recently, Adjust has mainly focused on preventing ad fraud. With the Unbotify acquisition, their capabilities now go beyond mobile performance ad fraud protection to protect an app’s business after an install.

In-app purchases
“If you take a game with in-app purchases, around one per cent of the people who play the game will buy stuff, and around one per cent of these will be the guys who spend thousands of dollars,” he says. “If you get someone like this playing a multiplayer online game and they see that they are constantly losing to what they know must be a bot, then sooner or later, they are going to get so frustrated that they will churn. It’s a massive problem for the games industry.

“Then we met Unbotify. They are a team of data scientists working with some of the biggest brands in the world, and they have developed technology that takes the phone’s on-board sensors – accelerometer, gyroscope, battery, light sensor or duration and pressure of touch events – and uses the data from those sensors to identify whether it was a human or a bot that was using the app. This gives us so much more data to play with, hundreds of data points, just from a single session. Importantly, this data is compliant with all data privacy regulations – and is only ever used to distinguish a bot from a human, not humans from each other. Once you can see the data behind the clicks and other actions in the app, it’s almost impossible for a bot to replicate how a human would use it. State-level actors might be capable of fooling this type of tech, but not the vast majority of regular fraudsters.”

In-app bot fraud is a big issue across all verticals, but eCommerce is one it’s hitting particularly hard. On 29 March, the number one paid-for app on the App Store was a $20 Supbot, which helps its users get their hands on the latest Supreme drops – the date when the fahsion brand’s new stock is unveiled – faster than other shoppers. For many of the bot’s users, it’s an opportunity to buy the goods in bulk and re-sell these limited edition items at a much higher price. For regular shoppers, it can ruin the user experience and cause long-lasting damage to brand reputation.

The Unbotify team continues to operate out of its Tel Aviv base, and the acquisition is unlikely to be the last, given that in June, Adjust secured $227m (£187m) in funding. “The money will help us to find and acquire more good companies,” says Henschel. You need to be able to move fast and having the funds in place will help us do that.”

The money will also enable Adjust to increase its footprint in its existing offices – it has 15 across the globe, in Berlin, New York, San Francisco, São Paulo, Paris, London, Moscow, Istanbul, Tel Aviv, Seoul, Shanghai, Beijing, Tokyo, Mumbai and Singapore. Headcount currently stands at 400, around 60 per cent of whom are engineers.

“We already have a large presence in south east Asia, but we will be adding more people in China, Japan and Korea,” says Henschel. “The same applies to our two US offices, India and South America.”

One platform
Looking ahead, Müller says that the company will continue to be driven by looking at what its customers are doing every day, and at how it can use technology to make their lives easier. “For us, it’s not about having the best tech around automation, AI or whatever,” he says. “It’s about having lots of different, complementary products in one platform that people can access easily with one login and one dataset, removing a lot of the manual work and leaving them free to work smarter.”

Henschel adds: “When apps launched around 2008, people said they were just a fad – but look at the industry now. Apps are still on the rise and will continue to evolve over the years. In the years to come, we want Adjust to be the backbone of this extensive and lucrative mobile ecosystem. Attribution and measurement will remain our main focus, but we’ll also continue to expand our services and fulfill customers’ growing demands – providing marketers with a single platform for all their needs."

Case Study: How Viber stopped mobile ad fraud with Adjust, saving 10 per cent of its marketing budget
Viber, the free instant messaging and calling app, now has over 1bn users worldwide. To increase its global footprint, the company decided to actively grow its user base across Europe, Asia and North America. But growth plans also mean apps often become targets for ad fraud. Viber turned to Adjust to help it get ahead of fraudsters and protect its marketing budget.

Working with Adjust’s Fraud Prevention Suite (FPS) saved Viber from wasting its budget on fraudulent traffic. On average, the company saved 10 per cent of its marketing budget, which could then be reinvested in acquiring real users.

Importantly, the tool also gave Viber more freedom to experiment with different networks. With FPS activated, it could be more lenient with testing new partners, as they knew the data it received was filtered for fraud. Since working with Adjust, Viber has now changed its internal evaluating logic for network quality. It no longer just looks at the top KPIs but also includes fraud rate as an evaluator. Now, if a partner reaches more than 20 per cent fraudulent traffic, Viber stops working with that partner.

A great example of effective fraud prevention in action. 

This article first appeared in the September 2019 print edition of Mobile Marketing. You can read the complete issue online here.