Let's Stop the "use it or lose it" Habit in Advertising Marketing

When I said "it's White Christmas for the bad guys" every Q4 in relation to ad fraud surging in Q4, everyone should realize it's also "White Christmas for agencies, ad tech middlemen, and every other vendor" holding out buckets to catch the money falling like rain from big advertisers. Why? Because of the pervasive "use it or lose it"?mentality amongst advertisers and marketers, and the policies that incentivize this illogical behavior.

When I worked at a big advertiser, I was told to just come up with some more campaigns so we can use up the digital budgets before the end of the year, after all "use it or lose it," right? When I worked at the big agencies, I was told to help clients think of new campaigns in December so we can finish spending their budgets before year-end, otherwise they won't spend as much with us next year. A friend who works at an exchange tells me every year, customers come to them with sizable budgets that need to be spent before the end of the year, begging them to help them find some supply (of ad impressions) to use it all up. How do you use up $10 million within a month? Right, "let's spin up some more BS" (which in this case means "bot supply"). This bad habit incentivizes ad fraud. After all, there's not a whole bunch of humans sitting around with nothing to do but to go to the websites you tell them to visit in large quantities.

Ad fraud didn't happen in other media channels at the scale it does in digital because of the laws of physics. In the physical world, we can only stick so many billboards by the side of the highway (left side of slide above), only so many pages of ads in a magazine, or only so many minutes of advertising on a TV show (e.g. 15 mins out of 30 minute show). It's finite and limited. In digital, however, since everything is just bits and bytes, we removed those physical constraints and bad guys took advantage. They created literal millions of fake websites, stuck literal hundreds of ads on each page (right side of slide above), and used literally limitless bot traffic to manufacture as many ad impressions as needed to absorb all the budgets that marketers had to spend.

The slide above is an example I use in "fraud hunting workshops" I've run over the years. The top part of the slide shows an example of a traffic surge on the last day of the month (green bars, much higher on Feb 29). That is what we call "end of month traffic fulfillment" -- and as you can see it's the red line (bots) that shot up and back down. In the chart on the lower half of the slide, you see surges in volume (green bars) corresponded to upticks in the blue line. This was a news site and those surges were from human visitors when certain pieces of news hit.

Further, if you're a clever criminal, you no longer even need to set up fake websites or pay for real bots; advanced fraudsters these days just fabricate the excel spreadsheets and the log level data to make it appear that the ads ran, when no ads were actually ever served. This is how criminals ripped off Uber in 2017 -- they just falsified the analytics and fabricated the transparency reports [1 ], [2 ]. It's literally some 1's and 0's in a database and a few summaries on spreadsheets. They can even rest assured the client won't look too closely or ask any hard questions -- after all, the advertiser WANTED to buy it, because they had to spend it -- i.e. the "use it or lose it" habit. Agencies, hucksters, ad tech leeches, and fraudsters alike, all cheer for this "use it or lose it" policy. It's been a goldmine for them for at least the last decade (when programmatic ad buying took hold); and every Q4 it can literally help them "make their number" too -- i.e. even their stretch-goals for revenue.

Final reminder that since programmatic ad buying took hold in 2012-13 (chart above), you can see that the blue line (digital ad spend) diverged more and more from the yellow and green lines (humans' usage of the internet, social, and mobile). What can account for that dramatic and continued increase in ad impressions, if it's not due to human activity and time spent with media? Right, bot activity. Network data providers have documented that automated (i.e. non-human) traffic on the internet has increased to nearly 3/4 of all traffic. Automated traffic doesn't necessarily mean fraud bots, there are useful bots like search crawlers that perform a useful service. Between this data point and the observation that vast botnets can generate enormous traffic for DDoS (distributed denial of service) attacks, it is easy to see how straightforward it is for fraudsters to generate trillions of ad impressions per year out of thin air. Programmatic ad buying and technologies facilitate fraudsters, ad tech intermediaries, and media agencies to feast on ad budgets supplied by the largest advertisers.

So What?

Hopefully the data above illustrates how marketers' "use it or lose it" mentality is incentivizing more ad fraud (and is an illogical waste of money). Like I said before, ad fraud is not a tech problem; it's an incentives problem. And this is just one example of it. Marketers should run their own turn-off experiments, like P&G, Chase, Uber, eBay, and Airbnb. They turned off digital ad spend and there was no change to business outcomes and activity.

CFOs and CEOs should also change company policies so they don't inadvertently perpetuate this "use it or lose it" bad habit. If marketers knew it was OK to save the money, if that is the better option, they should save the money and let that flow directly to the bottom line, rather than waste it, cause more ad fraud, and line the pockets of cybercriminals. Small businesses and DTC (direct to consumer) marketers already do this. You can too, big corporate advertisers.

Further reading:

https://www.dhirubhai.net/pulse/positive-story-digital-marketing-stopped-spending

https://www.dhirubhai.net/pulse/marketers-wake-up-sold-you-manufactured-scale-cost