Today Google announced it took down over 3.2 billion “bad” ads from its platform last year, equivalent to over 100 ads removed every second. However, two SSP’s working with Google’s DoubleClick Bid Manager (DBM) told VAN the search giant has created a ‘blank cheque’ for itself as it awards itself refunds for invalid traffic (described in reports as ‘IVT’) without disclosing how or why it detected the fruad.
In most instances buyers and sellers resolve fraud issues together, usually via an agreed independent intermediary (i.e. a third party such as Moat, Integral Ad Science or one of the other fraud detection vendors) who detect the fraud based on an agreed set of standards. However, Google recently changed their terms and conditions so that Google — and only Google — can decide what is or isn’t fraud.
Here’s what the new clause in Google’s contract says:
One of the SSP’s VAN spoke to claimed to have investigated the traffic themselves and couldn’t detect any fraud whatsoever. “Based on their size, DBM are simply able to bully people into signing,” one SSP executive told VAN. “We’d be more than happy to refund Google for genuinely fraudulent traffic – it’s an industry scourge – but Google have effectively given themselves free rein to pluck numbers out of the air to issue themselves refunds. This isn’t just about our revenue either. It’s going to hit publishers, who we have been told to pass the revenue hit on to, but it’s ridiculous that we can’t explain how Google are supposedly detecting the alleged fraud.”
Another executive from a leading SSP said, “Google are exploiting their strength in the market to bully us and other SSPs into paying back fees. They’re using fraud to effectively write themselves a blank cheque to refund themselves seemingly arbitrary amounts that we have no way of verifying. These are precisely the type of opaque business practices that the industry is supposed to be trying to get away from. Once again we have a situation where Google are marking their own homework.”
Other SSPs chose not to comment, although one said it was a known issue but they were concerned about any possible fall-out that might arise from damaging their relationship with Google. A number of executives said they rely so heavily on demand from DBM advertisers (one said it’s often around the 50 percent mark), they felt compelled to sign up to the new terms.
A Google spokesperson issued the following statement to VAN earlier today: “We have a longstanding policy of refunding advertisers for invalid traffic. As we announced last year, this is currently being expanded to include ads purchased via DoubleClick Bid Manager. Most invalid traffic is filtered from our systems before our advertisers are ever impacted or charged. Maintaining a healthy ads ecosystem and improving transparency in the supply chain for our advertisers is an area we continue to invest in significantly.”
Google also said that over the past 15 years they have developed an extensive internal system to filter out invalid traffic – from simple filters to large-scale machine learning systems – and have launched more than 180 separate invalid traffic filters. The company also said they have steadily introduced new defences across their ad systems, including protection against ad hiding, ad injection, clickjacking, botnets, falsely represented inventory, and ‘much more’.
Over the last six months Google has introduced several new initiatives which they say helps to “simplify this complicated threat” for their clients for and partners. While Google has a longstanding policy” of refunding advertisers for invalid traffic, in September of last year the company expanded this policy to include 3rd party inventory (ads purchased from non-Google SSP’s via DoubleClick Bid Manager).
Google Say They’re Fighting the Good Fight Against Invalid Traffic
Earlier today Google reported on actions it took to fight harmful and malicious advertising, showing that ad takedowns almost doubled from 1.7 million in 2016, and outlining emerging forms of undesirable ads that have popped up over the past year.
Google removes any ads it serves that appear on websites which violate its AdSense publisher policies, and also ads which themselves go against its AdWords advertiser policies. While some of these standards are controversial (such as those set by the Coalition for Better Ads, which Google began enforcing last month), Google says others are geared towards combating blatantly malicious behaviour.
Some of these bad behaviours seem to be on the rise, as Google is blocking certain types of bad ads more and more frequently. Over 130 million ads were removed last year for trying to abuse Google’s ad network through malicious activity, or for attempting to circumvent its ad review process. The company took down a furthers 79 million ads for sending people to malware-ridden sites, and removed 400,000 of these websites.
Tabloid cloaking, a tactic whereby ads disguise themselves as news stories to lure users into clicking, continued to rise, with 7,000 AdWords accounts suspended for tabloid cloaking violations last year, up from 1,400 suspension in 2016. “Trick to click” ads on the other hand, which use deceptive messages to trick users into downloading harmful software, fell compared to 2016, with 66 million trick to click ad removals down from 112 million in 2015.
Google also continue to block large number of publishers, disabling AdSense accounts of those it deems harmful. Last year it removed 320,000 publishers from its ad network, and blocked just under 90,000 websites and 700,000 mobile apps for policy violations. Of these, 12,000 were blocked due to “scraping” (duplicating and copying content from other websites), up from 10,000 in 2016.
When it comes to gauging the amount of fraud in the industry, Google’s numbers should probably be interpreted carefully. Firstly, it isn’t clear whether rising ad removals are a sign of an increase in ‘bad ads’, or are simply a symptom of Google’s methods becoming more effective. More importantly though, not all ad takedowns were necessarily legitimate. Nonetheless, Google is continuing to expand its scope for ad removal, having introduced 28 new advertiser policies and 20 new publisher policies last year to target trends it deemed problematic. Significant numbers of advertisers and publishers were blocked as a result of these rule changes. Ads were removed from just under 9,000 pages for violating new rules on intolerance and hate speech introduced in April, while 90 publishers were struck from Google’s ad network for hosting misrepresentative content, part of the company’s crackdown on fake news.
The report also outlined new areas which Google will be tackling over the coming year. Ad policies will be updated to cover “unregulated, overly complex, or speculative financial products like binary options, cryptocurrency, foreign exchange markets and contracts for difference”.
Cryptocurrency ads caused problems for Google earlier this year as it was found that YouTube ads were covertly leaching off users’ CPUs in order to mine cryptocurrency. However this update seems more related to blocking ads which encourage people to buy highly risky financial products which they may well not understand.