Digital Marketing Based on Business Outcomes, Not Vanity Metrics

Listen: https://www.youtube.com/watch?v=-cD2Srf4nj0

Let me put it plainly, as I do. You just bought a billion ad impressions; what do you have to show for it? Excel spreadsheets or fancy dashboards that show how many impressions you bought, how many clicks you got, and what average CPM price you paid for those ads? Cool, cool. But did those ads drive any business outcomes (i.e. sales, leads, subscriptions, etc.) for your company? Did you even measure for business outcomes? How? As a marketer, your duty is to drive business outcomes and revenues for your company, not to "spend it all, so I can get more budget next year." You can also drive the bottom line (profits) by NOT spending ad budget on useless ads; every dollar saved is a dollar more of profits for your company.

"every dollar saved is a dollar more of profits for your company"

The largest of advertisers that have gigantic digital ad budgets to spend before the end of the year are by far the most at risk of ad fraud; in fact, they practically incentivize ad fraud because there's not enough real ads from real human activity to absorb all that budget. The largest of advertisers are also the parties who are most willing to delude themselves with data, of course with the help of trade associations and verification vendors that tell them fraud is low and to keep spending like maniacs. Small and medium businesses can immediately tell if digital campaigns are not driving any new customers or sales for them, and will adjust accordingly (like stop spending). Here's why it is easy for the largest advertisers to delude themselves with the data they are given -- they see what they want to see.

What follows is what you can do about this, so you can do better digital marketing.

"But, but, programmatic is driving so much traffic"

Just yesterday, I heard a marketer lament "but, but, programmatic is driving so much traffic [to our site]" implying it'd be a shame to turn it off. But Google Analytics and Adobe Analytics don't tell the advertiser how many of those clicks from programmatic channels are not humans. Wait till they install FouAnalytics on the site and look at the clicks coming from "utm_source=programmatic." Below are donut charts of clicks arriving from 15 different programmatic campaigns.

How many more of these donut charts do I need to show you, before you realize that getting a lot of clicks does not mean "performance" and certainly does not mean more business outcomes? If you squint, you may be able to see the tiny slivers of dark blue in the examples above. The percentage of real human clicks (dark blue) when measured by FouAnalytics on the landing pages is in the single-digit percents. That means whatever ads you bought upstream from these clicks were crappy or entirely useless. Look at the chart below, with "bid floor buckets." It shows that the vast majority of bid floors are from $0.00 - $0.15 CPMs. 15 cent CPMs is a bid floor that NO legitimate publisher can set, because they have actual costs of creating content with real journalists and editors. Fake sites and mobile apps can take 5 - 15 cent CPMs and still be profitable because they have no cost of content -- 100% plagiarized or auto-generated by AI algorithms.

Of course, you didn't know the fraud above because for years, the legacy fraud verification companies you paid for didn't tell you anything useful. In fact they told you something that was harmful instead -- that fraud and other issues were less than 1%. That's why you did nothing about it. If fraud were that low, why would you do anything differently? See the following hilarious screen shot from a legacy fraud verification vendor dashboard -- everything is >99% "brand suitable, fraud free and in geo" yay! No need to do anything differently. But sadly, these numbers are the result of the legacy fraud verification vendors not being able to detect most of the fraud (i.e. fraudsters are easily able to avoid their detection). I won't go into the details here, but review the evidence yourself in this other article: Legacy Fraud Verification Vendors Are No Longer Fit for Purpose

The trade associations happily parrot this crap from legacy fraud verification vendors because they want to claim their certifications caused everything to be so good and clean. The following table is from the 6th year of press releases from TAG, the lapdog of the ANA. This is why I have said North American advertisers are far behind the rest of world because they have been consistently misled by the ANA, with the grand speeches on the main stage about transparency, viewability, brand suitability, etc. but no real action and no real knowledge about the problems at hand.

How are advertisers breaking free of this narrative from the ANA?

FouAnalytics -- analytics for your digital media

Just like you have Google Analytics for your website, you can add FouAnalytics to your website to get new insights that GA could not give you. See these 2 articles for screen shots and examples: 1) How Site-Owners Use FouAnalytics to Troubleshoot Bot Traffic , 2) How to use FouAnalytics to Scrutinize Clicks from Programmatic Campaigns

Even though you've paid for legacy fraud verification for years, they have done nothing useful for you -- they haven't reduced fraud for you because they couldn't detect most of it in the first place. When was the last time they gave you a list of sites and apps to add to a block list? Time to upgrade your tools. Advertisers are adding FouAnalytics in-ad measurement tags to their programmatic ads, one step upstream from the clicks they are receiving on their sites. In-ad measurement shows them where their ads went -- e.g. breitbart.com -- even if placement reports don't show that. FouAnalytics also tells them which sites and apps to consider turning off due to fraud of all kinds, not just bot traffic -- e.g. ad stacking, pixel stuffing, auto-refreshing of ad slots every 1 second, auto-reloading of webpages, hidden ads, mobile apps loading ads continuously even when not in use, pop-unders, etc. etc.

If you've read any of my articles over the years, you will already realize that there's an endless number of sites and apps that are committing fraud. And block lists are only so long (10,000 rows in most cases). You realize that a block list approach is an endless game of "whack-a-mole." Savvy advertisers have started adding new campaign lines that use an inclusion list approach -- adding only quality sites and apps to the inclusion list. There's only a small number of these. Let common sense be your guide here. If you have never heard of the sites and apps before, it is likely other humans have not heard of them either. Add the sites and apps that you have heard of to the inclusion list and keep adding a few more, based on what the data tells you. Good clicks come from good publishers (and real devices). Remember the slide below? When we focus on real clicks (blue and dark blue in FouAnalytics) we can see they come from real publishers that you've heard of, and real devices. Humans use iPhones, a lot of them, especially here in North America.

Keep adding good sites to your inclusion list when you have sufficient confirmation that you are receiving good clicks, human clicks from them. (If you have questions about how to do this, ask me; I will show you how).

Optimize towards humans, not just away from bots and fraud

Once you have the right analytics in place, you can not only optimize away from all forms of fraud, not just invalid traffic ("IVT"). You can also compare the relative quality of different paid channels. In the examples below you can see the difference in dark blue (humans) and dark red (bots and fraud). For example, "Google paid search" has far more dark blue than "Google with search partners." TikTok without their audience extension network Pangle is far better than with Pangle, and so on.

In the column charts above, you can compare 3 different "paid display" sources and easily understand that you should allocate more budget to C (because it has more dark blue) and reduce budget to A (because it is more dark red). You can also compare across types of media. For example, "paid social" F is working really well -- more dark blue and less dark red than all three "paid display" sources. That means you can allocate more budget to "paid social" F and reduce budget to all of the paid display sources.

With legacy fraud verification vendors, you only got a number -- a percent IVT or fraud. The legacy fraud vendors don't measure for humans. By upgrading your tools, you can see not only the levels of fraud, but also the levels of humans (dark blue). With these new-found details you can not only optimize away from fraud, but also optimize towards humans. You can add sites and apps to block list and you can also create campaigns based on inclusion lists (where are you getting the most human clicks). You've never been able to do this before. Now you can.

"showing ads to humans in the first place is necessary for digital campaigns to have any impact on business outcomes"

So what?

Now to bring this home, you can finally start to optimize for business outcomes, once you have taken care of the above. I have said over the years that "showing ads to humans in the first place is necessary for digital campaigns to have any impact on business outcomes."

Previously, you didn't know your ads were not shown to humans, because legacy fraud verification vendors only told you about invalid traffic -- that there was very low IVT. Low IVT does not mean high humans, obviously. It just means those vendors failed to detect anything wrong or invalid. By upgrading your tools, both on your landing pages and in the ads themselves, you can not only see more kinds of fraud, you can also optimize towards showing ads to humans. Once you do that, you can focus on measuring for business outcomes. The high clicks you were getting before, were NOT from better targeted ads but from bots clicking on your ads (remember the 15 donut charts above with mostly orange and red?). Even if you got a lot of these kinds of clicks it doesn't do anything for you business outcomes.

When you have the ability to measure for humans, you can see that you can buy far fewer ad impressions, pay higher CPM prices (to good publishers) and still save money, get your ads in front of humans, and drive more business outcomes for your company. That is how you become better stewards of the digital investments you were entrusted with.

Are you with me?

Happy Sunday, Y'all!

Further reading: https://www.dhirubhai.net/today/author/augustinefou