Buy Digital Directly. No, I’m Not Crazy!

Dr. Augustine Fou just published an interesting presentation, “You’re Not Getting What You Paid For,” which I encourage all marketers to read. The presentation lays out a number of the ways you–yes you–are being screwed in the digital ecosystem. In short:

  • You’re not running where you think you are
  • Your fraud verification tools aren’t catching the problems
  • The ad networks you’re buying on are serving ads to bots
  • The system is so complicated you don’t know where to start.

But there’s a glimmer of hope and Dr. Fou gives you a clue to the solution.

In slides 15-18 he gives examples of marketers (JPMorgan Chase, Restoration Hardware, and Uber) cutting back on the number of sites they purchase on. In all cases, the performance didn’t suffer. I’ve seen the exact same results with my clients.

Thinking back to June, I read this article about P&G and Unilever and reduced it to the following graph:

Screen Shot 2017-09-19 at 3.10.40 PM

You can see the raw data on this Google sheet.

Four things should jump off the graph above:

  • Google and Unilever sell products to everybody
  • Unless you work for a massive CPG company, you don’t
  • They don’t run on many sites, yet sell billions of dollars of stuff
  • They are running on fewer, not more, sites as time goes on

So how come P&G and Unilever can be successful running ads on <1,000 sites when you “have to” run on thousands?

The answer is they are looking ruthlessly at the data, calling BS on what they see, and you’re not. The bonus is that once you can get down to <1,000 sites, you can buy direct. No more going through sketchy ad networks selling ads to “audiences” of who-knows-what.

Yes, it’s manageable. Heck, it would even work in Excel. I’ve done the analysis on hundreds of thousands of sites by hand, by myself, in Excel. It didn’t even take that long.

So why aren’t you buying direct and only on a few sites?

Takeaway: Hard whitelist. Buy direct. Cut out the ad tech middlemen. Act like a performance marketer. And win.


Posted in Data, digital marketing, Direct Response, fraud, Media, Strategy | Tagged , , , , | 2 Comments

Profit Maximization, Humanity Minimization

Facebook stepped in it again, this time by allowing users of their advertising platform to target users via vile anti-Semitic terms. Do I think the people or the company is anti-Semitic? Of course not.

I think the company* is anti-human. We get in the way of making money. We refuse to be upsold predictably, consume on demand, and take predictable “customer journeys” that conveniently spew forth the most profit.

I think Facebook’s problem was the blind faith that algorithms can replace what people do. But the algorithms can’t be written to demonstrate empathy. You can’t score common sense. Sure, algorithms can maximize Facebook’s advertising revenue. A side effect–as we see–is that they can minimize humanity when humans aren’t empowered to deploy common sense.

There’s an article at The Atlantic that asks if Facebook could have caught the problem, and I expect you’ll see more articles along those lines. That line of thinking is ridiculous. Of course they could have caught the problem. But nobody was looking.

People instantly saw the abhorrent nature of that type of targeting and correctly shut it down. But I believe the people inside the company have been minimized as well. The techbro coder culture takes away people’s decision-making ability: The algorithm built by the expensive guy in the hoodie, living on Soylent, can’t possibly be wrong.

Silicon Valley algorithms are just the work of humans, riddled with their own biases, layered with the inevitable software bugs, and rushed out rapidly to maximize profit. To fix human problems, you need to empower humans and place them, not the algorithms, at the top of the heap.

Takeaway: Build the best tech you can. Make sure humans, with their ability to be empathetic, nuanced, and unpredictable are always allowed to use judgement. Build up humanity. And win.


* I lump Facebook, Google, Amazon and other Silicon Valley companies that practice the destruction of value and human lives through “disruption” together.


Posted in Leadership, Organization, Philosophy | Tagged , , , | Leave a comment

AdTech: The Worst Server Room Ever

Screen Shot 2017-09-14 at 7.51.57 AM

“Do Not Touch Any Of These Wires”

We get a kick out of images like this one at the left. How in the world could somebody let a system get so out of control?

When Alan Kay shows this picture, he points out that every wire was put there for a reason, to solve some particular problem. The excessive complication of the wires probably masks some moderate underlying level of complexity. As Kay points out, when we hit the “Yikes!” moment, we’ve got to figure out what caused the “Yikes!” moment. Was it complexity or complications?

Complications are noise that can turn a complex situation into “Yikes!”

The problem with our server room? If you start pulling wires, you’re screwed.

Now tell me how the ad tech ecosystem below is any different than our server room? I get the same answer when I ask about pulling ad tech and mar tech wires: “If you do it, you’re screwed.”

Screen Shot 2017-09-14 at 7.54.38 AM

“Do Not Touch Any Of These Wires”

I think that a lot of the complications and noise have been added intentionally to create a “wire” of money that the advertiser or publisher dare not pull.

Takeaway: Avoid adding complication to address complexity. Avoid adding noise. And win.

Posted in Leadership, Philosophy, planning | Tagged , , , , , | Leave a comment

Guitar Effects: Two Legends and a Sale

My core beliefs include positivity and assuming goodwill. If you only knew me from my writing here, you might not think that. It’s hard to be positive sometimes when I seem to spend the whole day mired in “digital” marketing problems.

So, before the Labor Day weekend in the U.S., let’s go positive with a brilliant piece of marketing by the great (and I chose that word specifically) Lee Anderton (“the Captain”), from Anderton’s Music Company in the U.K.

First, a few things about guitarists, the target market for the products in this video:

  1. Most guitarists are on a constant tone chase, wanting to sound like their favorite player(s) or getting a unique sound of their own.
  2. The music industry obliges that quest by producing tons of amplifiers, guitar pedals, and guitars that purport to help.
  3. Most guitarists (myself included) refuse to believe the truth that “tone is in the fingers.”
  4. Most guitarists would be better off spending 99% of their gear budget on lessons, and 98% of the time they spend watching YouTube videos on practice, myself included.
  5. Guitar players have the peculiar habit of watching many hours of videos and reading guitar forum opinions before spending just $99-$200 on a guitar effect pedal. This last part is the most important. It’s the behavior that Anderton’s has zeroed in on.

What makes Anderton’s use of YouTube great is that the Captain acknowledges the above and, understanding the needs of his target audience and his need as a retailer to sell product, has created a brilliant YouTube channel. He has great presenters, funny banter, zero sales pitch pressure, great education, and an easy way to see and hear the gear you’re contemplating.

I don’t know anything about their business, other than I bet they shift a lot of goods, have great customer care ratings, and I’d be the first on line to patronize them if they ever opened a shop here in the U.S.

With the above in mind watch a few minutes of this video of the incredible Peter Honoré and Tore, the product manager for TC Electronic. You get everything from how easy the products are to use, to the clever “Mash” technology for the pedals, to a great idea of what you could do with these TC products.

The video addresses all the needs and desires of a guitarist with the addition of one thing most “content marketing” types don’t understand: this is darn good entertainment. These guys are engaging, funny, and truly knowledgeable experts in their field. Even if I don’t buy from Anderton’s this time, the saliency of their brand is embedded in my mind every time I watch a video on the channel.

So how did this video work for me? Midway through the demo of the TC Flashback 2 product, I was over at Amazon (sorry Lee!, but good for Tore) and picked one up. I still haven’t totally figured it out and, of course, I can’t play anywhere near like Pete. Check it out:

Good marketing? Yep.

Takeaway: Understand the behavior of your target audience. Develop strategies that align with that behavior. Deliver compelling messages in appropriate media. And win.

Have a great Labor Day weekend and a great fall.

P.S. Lee, if you ever read this, a question. Why did you bring Pete on board? He sucks me in with his great playing and then makes me want to go to the basement and chop my guitars into firewood!

Posted in Branding, Communications, digital marketing, Guitar, Marketing, Media, Sales, Social Media Marketing | Tagged , , , , , , | Leave a comment

Mix modelling muddles marketers

See the final four recommendations!

Marketing Science

I think econometric modelling is over used in marketing. And routinely produces misleading results. Let me explain…

View original post 1,339 more words

Posted in Uncategorized | Leave a comment

Thought Experiment: Am I Just an Example of Survivorship Bias?

I’ve been a fairly successful continuity marketer, CMO, whatever. As such I’ve got a lot of what I think are good ideas about how to do marketing-type things.

But what if I’m just an example of an outlier that’s made it through 27 years by chance?

Hmm. Let’s see how well I sleep tonight!


Aside | Posted on by | Tagged , | Leave a comment

Need an EVP of Brand Safety? You’ve Got a Problem!

It took a second read of the “Lets Talk About The Brand Safety Tax” column in AdExchanger last Tuesday to catch this:

It’s a small price to pay, said John Montgomery, global EVP of brand safety at GroupM, running somewhere between two and 10 cents per thousand impressions for brand safety protection when it’s bundled in with viewability and anti-fraud measures from the same provider.

Notice that title? Why in the world does an agency need a global EVP of brand safety?

Do the account people or media planners and buyers understand what’s on-brand and what’s off-brand? Do the clients understand or clearly articulate what types of media are appropriate for their organizations?

An EVP of brand safety, plus the staff in that organization can’t be cheap. And the clients are paying for it. Once upon a time, you could trust your agency, like a physician, to “first, do no harm.” Now, you’ve got to have an in-house babysitter to catch the harm after the fact.

If the existence of EVP-level “brand safety” people at agencies isn’t a good example of what’s wrong with digital advertising, I don’t know what is.

Takeaway: Chasing “audiences” at the cheapest CPMs, with no segmentation strategy, is the root cause of brand safety problems. Think like a marketer. If it’s too good to be true, it is. Be skeptical, always. And win.

Posted in Branding, digital marketing, Media, Strategy | Tagged , , , | Leave a comment

Source Level Triage

Screen Shot 2017-08-22 at 8.34.48 AMLet’s assume you’ve built your marketing whiffletree and have a banding model at the source level that you can analyze every Monday.

Of course, you’ll sort your sources from most effective to least and do more of the things that work while cutting the dogs.

Correctly, we find ourselves spending a lot of time at the banding cutoff, i.e. looking at sources just above or below our allowable. That’s where you can scale your efforts and optimize marketing ROI.

I also suggest you think about your media sources using the triage method to help focus your efforts. Like classical triage theory, your sources belong in one of three “tents” where the source:

  1. Will never get any better. The performance you have is what you have. You probably can’t scale it much and changing creative and offer will be of limited efficacy. If it’s above the cutoff, great. If it’s below, cut it and move on.
  2. Will be hard for you to screw up. These sources basically print money and will work almost regardless of what creative you throw at them, how much you run, etc. (There are usually not many in this bucket, of course.) Your most junior employees can learn the ropes on these sources. Take advantage of the source’s resilience and use it to train your next generation of talent.
  3. Could go either way. Meaning if the creative, volume, or mix with other channels is right, they work (or could be improved to be made to work). If wrong, the source performs at a rate below your allowable. This is where you earn your pay.

And the source could be in any of the tents, regardless of the sortation on the banding model.

Spend most of your time in tent three. Look at the things that need your limited resources, whether they be from your channel managers, your agency, or your analytical team. Those sources could live or die, but they need your attention. And don’t have triage done by interns. It should be done by the marketing experts with the most experience.

Messy work? Sure. But if you’re going to figure out what’s going to survive, you’ve got to get into the blood and guts of your media performance, and make some hard decisions.

Takeaway: During source analysis, focus on the sources with variable outcomes which are dependent on your intervention. Spend your time in “tent three.” And win.


Posted in Marketing, Media, Philosophy, Tactics | Tagged , , , | Leave a comment

Who’s Going to Debug Digital Advertising?

In Alan Kay’s talk “How To Invent The Future II” he quotes Tony Hoare:

“Debugging is  harder than programming, so don’t use all your cleverness to write the program.”

The entire digital ecosystem, so incredibly complicated that nobody can tell why a particular ad was served to a particular identity (or even if the identity is a human), was programmed by thousands of companies in pursuit of quick profit.

Now that we know it’s a mess, who’s gonna fix it? Not more tech, that’s for sure. And there’s no money in the slog of debugging.

Takeaway: Don’t rely on buggy technology. Think like a marketer. Segment and prioritize before you go in search of “reach” or “audiences.” And win.

Posted in Data, digital marketing, Marketing, Media | Tagged , , , , , | Leave a comment

Live by the Algorithm, Die by Same

Think your fancy, all-seeing algorithm can’t be gamed?

So do the fools at Uber. Unfortunately, it looks like Uber drivers can just switch off the app at once and trigger surge pricing. No complicated competing algorithm, no data scientists needed.

Complex systems can be–when created by people–relatively fragile. And the bigger the complex system, the more likely it can be disrupted by small and perhaps unexpected things. Case in point, Uber and taxi drivers with cell phones.

Do you really think your fancy multi-touch attribution models can beat armies of profit-motivated criminals with (among other things) IOT botnets?

Takeaway: Start with proper marketing, then add technology to support your strategy. Check everything. De-complicate*. And win.

*Changed from “de-complexify” on 9/20/17. I’ve noticed several cases when I referenced complexity instead of complication and have tried to make those corrections as I find them.

Posted in digital marketing, fraud, Marketing, Philosophy | Tagged , , | Leave a comment