Many marketers have looked upon digital marketing as the holy grail or a shiny object that had to be had. So they shifted mucho dollars into digital. Digital does have the advantage of far more data; you can literally “measure everything.” But does more data mean better marketing? Obvs, no. Another article already discussed how many shiny digital objects are no more than digital snake oil. This article will focus more on the practical matters of waste, fraud, and inefficiencies that can be improved right away with a healthy dose of skepticism and a pinch of common sense (once you have the right data and know what to look for)
Profitable middlemen pockets – The association representing British Advertisers (ISBA) published a study of the digital supply chain in May 2020 that showed only about half of the dollars spent by advertisers were getting to publishers, for the purpose of showing ads. The rest of the dollars were absorbed by the middlemen – agencies, exchanges, ad tech vendors, etc. Of course some vendors are adding value. But of course others are not. Do you know how much incremental sales you got for every dollar spent on a digital shiny object seller? If you can’t answer this question right off the top of your head right now, you probably shouldn’t be paying for it.
Mysterious “unknown delta” triangle – The IBSA report also showed something curiously, or euphemistically, named “unknown delta.” This was 15% of the overall dollar moving through the programmatic supply again, and roughly 1/3 of the portion that went to middlemen. A THIRD of the dollars could not be accounted for? That’d be like putting $3 into a vending machine, and it showing you only put in $2. You’re definitely not getting that $3 candy bar.
Oops, accidentally “kept” the money – And finally, have you considered what happens when you plan to buy $10 million of digital media, but you’ve only bought $6 million worth of ads, by the end of the campaign? What happens to the $4 million that was not spent? Did the agency give it back to the advertiser? Did the advertiser want it back in the first place? Or did the agency keep it secretly or the advertiser tell them to keep it, because it would look bad if they didn’t spend it all? Wait till CFOs and shareholders start asking questions.
CMOs can reduce each of these forms of waste right away. Review every bucket of cost that went to shiny digital objects, and actually calculate whether the returns were greater than the costs. Shorten the supply chain and buy more direct from publishers – this way, more of your dollar goes towards showing ads, rather than into middlemen pockets. And get receipts for the digital ads purchased on your behalf (what you got) and cross reference that to what you paid. If those numbers don’t match, ask where the money went.
Digital ad fraud is a multi-faceted thing; and I’ve written extensively about it elsewhere. I will only mention three examples here that don’t require specialized tech to improve, just a tablespoon of skepticism and a pinch of common sense.
Domain spoofing is where fake sites pretend to be legitimate sites simply by declaring themselves to be esquire.com — i.e. falsifying the data. In the placement reports, the fake ads show up in the same row as the real ones, because all the impressions are labeled esquire.com. A tell-tale sign that some fake ads are mixed in is when the prices for the mainstream site appear to be too good to be true. The real esquire.com can’t afford to sell ads at very low prices, but fake sites can. When the prices seem too good to be true, ask more questions, like how is that possible? Also, ask for placement reports so you can check if your ads ended up on sites and apps like the ones in the slide below.
Geolocation fraud is a perennial favorite of fraudsters because they can just pass a fake location, any location, and make more money. Studies have shown that ads with location data appended get higher prices than ones that don’t. With so much money being spent in “mobile” you should definitely get more detailed data so you can check for things like teleporting devices (same device showing up in 50 different locations at the same time) or device swarming (huge numbers of random device IDs appearing to be in the same place, like the middle of a lake).
Phantom ads are where you paid for the bids won and the ad server may have even served the ad. But the ad never arrived in the device and rendered on screen. This type of fraud is particularly prevalent in mobile apps, especially the ones designed for ad fraud. Those can load more ads if they don’t waste time and bandwidth waiting for ads to load and fully render. If you buy such ad impressions, ask what portion of the ads actually rendered on screen. If they don’t know or can’t tell you, should you keep buying?
Finally, you may not be aware of other aspects of digital campaigns that are suboptimal, but can be easily improved if you know what to look for.
In-Geo and On-Target – Did your ads actually serve in the geography you were targeting? Sometimes when a campaign targets the U.S. other countries show up in the data. It could be a small problem (single digits) but it could also be larger than that. Do you have the data to know? If other geos are showing up in the data, you should add negative targeting to reduce those.
Pacing – Did all of your ad impressions get blown out by 4a? If you don’t have hourly charts, how would you know? The many cases of the entire quota of ads being spent before human waking hours can be resolved by ensuring pacing is turned on. This spreads the ad impressions more evenly throughout the course of the day, and ensures that there are ads left to be shown when humans are awake and online. Was pacing turned on in your campaigns? Note that even if it were turned on, you should still get hourly charts to verify, because sometimes the tech doesn’t work.
Frequency Caps – how many times do you want to annoy — oops, show — your ad to the same user every day, week, or month? Do you have frequency caps set? Did they work? How would you know? Do you have detailed enough reports so you can check for yourself? The charts and data were collected using #FouAnalytics.
Reach – You thought buying through programmatic would help you reach users on thousands of long tail sites? What if the data shows that 90% of your ads were shown on only 5 – 10 sites (most or all of which were fake ones). You didn’t get the reach you expected, right?
CMOs and marketers can reduce waste, lower fraud, and improve on inefficiencies seen in their digital campaigns, without using any fancy tech. But you must have the right data to look at. Ask your vendors for more details. If they don’t have it or won’t give it to you, that should tell you something. Find alternatives so you can measure it yourself and take the actions above.