Ad Platforms: Should You Or Shouldn’t You Take Their Recommendations? via @sejournal, @jonkagan

Recently, I did the math, and realized I’ve been in the biddable media business (search, social, programmatic, etc.) for over 20 years now.

(Shoutout to Didit.com for taking a chance on a Hofstra University senior with no experience and giving me a paid internship.)

In those 20 years, I have looked back at all of the changes within the industry, including but not limited to:

But, what really has changed in my mind is how we view support and insights from the platforms.

Google and Yahoo had true human support, and we took their recommendations as gospel.

There were few to no automated recommendations telling us to spend more, and when the platform reps made a suggestion, it was taken as a degree of good faith.

Before I go off on a rant, I should provide a disclaimer that I have numerous friends at all the major platforms in the U.S.

I’ve applied for jobs at all of them and have also been offered jobs from some of them. So, this article is not directed at a particular person or platform but the industry, in general, from the vendor side.

Backstory

While I can’t pinpoint a precise date, I’d say around 2012 is when we really saw consistent evidence of shifts in platforms moving from “We want what is best for you and your ads” to “We have recommendations to move your business forward if you can just spend more, or adopt these things that will cause you to spend more.”

I was with a large holding company agency, running a bit of pharma and a lot of financial services advertising with a small team.

In my not-so-humble and completely biased perspective, we had one of the best search marketing teams around. We knew the industry standard and best practices, and we knew what would help or hurt the business – and so did our reps at “big search and big social.”

Suddenly, their recommendations were not best practices and would clearly drive higher spend and lower efficiency.

Even more noticeably, when something went awry, the reps we turned to to help troubleshoot would say something to us that gave me a cold chill down my spine: “You’ll need to file a support ticket for that. These are now handled by a different department.”

This was the beginning of us really having to scrutinize what was recommended to us.

Human Vs. Machine

Yes, I am well aware that robots are on their way to take our jobs.

But, I keep getting reminded that if I let the machines have at it, several brands would lose a lot of money for absolutely no logical reason.

Google budget reco for a lot more money but not a lot more production (Image from author, March 2025)

Meta making a recommendation for a function we don’t even have (Image from author, March 2025)

Bing recommending a move with a questionable network (Image from author, March 2025)

These are samples of the same suggestions I get every day. The platforms want to drive our business forward, but ignore the following facts:

Yes, I recognize that these recommendations are both automated and optional.

But does that mean scenarios like these should be given a discount and not held to the same industry standards as others? There is only one correct answer: No.

We observe that the automated recommendations are faster, more real-time, and more correlated to actual data. But, the human-based suggestions have a human objective and the little bird saying whether these are actually a good idea.

We also note – and this is more specific to advertisers with a dedicated rep (and I don’t mean those you get for a fiscal quarter and call you at inappropriate times of the day) – that recommendations coming from them (while still salesy), are taking a look at the bigger picture, and making more specific recommendations to enhance your business.

Google wanting to expand into a network known to have questionable placements (Image from author, March 2025)

This is a machine/AI recommendation to opt into a network I intentionally opted out of due to extreme amounts of invalid activity coming from it. (I have more trust in Jenn Shah of RHOSLC safely holding onto my SSN than this recommendation.)

This ratio of Google Search vs. Partner Network does not make sense (Image from author, March 2025)

Rule of thumb: If the click volume in your search partner network is grossly exceeding your Google search volume, then there is likely something wrong.

Play stupid games, win stupid prizes … in credits, months later (Image from author, March 2025)

Look at all those credits tied back to that April campaign (this outcome reminds me of Aaron Rodgers leading my NY Jets in 2024).

Meanwhile, my account rep gives me emails like this:

What a proper recommendation should look like (Image from author, March 2025)

Net-net: Quantity (machine/AI) is not better than quality (human), but I will admit, it is faster.

Don’t Disregard Automated Recommendations Completely

While I may sound like a spokesperson for the anti-AI/anti-machine lobbyists, these systems can help save your butt here and there.

But, review these recommendations with a grain of salt while using them for a gut check.

Showing when something was forgotten (Image from author, March 2025)

Catching conflicts I would’ve missed (Image from author, March 2025)

Not Controlling What Happens Is The Devil

Yes, rep support is helpful, and sometimes, machine/AI can be helpful, too.

But, if you are not reviewing what happens in terms of placements, creative, or optimization, you may find yourself up a creek. This is particularly important on the creative side, especially if there are multiple layers of creative approval.

If you do not disable auto-apply, or you let Advantage+ run without oversight, then creative may not meet legal department guidelines.

Nothing gives you a heart attack on a Saturday more than when your client calls to let you know that they saw their ad in a live video on a social platform about some nefarious activity, with part of the creative cut off inappropriately.

If you run Performance Max, you definitely want to keep a close eye on this.

What does this mean? If you have a degree of control over the machine rather than letting it run with it, take control. An extra hour of work now will save you 12 hours of fixing problems later.

Having Control Of What Happens And Not Handling Properly Makes You The Devil

Yes, you read that right. If you want to test out the machine functions and AI suggestions, and you have the control to decide how and when it executes the work, and you still don’t use it properly, well, then that is your fault.

This isn’t just isolated to creative, but it definitely is much more awkward with it:

I really don’t know how Meta thinks this will sell Corned Beef Hash (Image from author, March 2025)

Bing thinks that this is a Jamaican Beef Patty … it is not (Image from author, March 2025)

But, this is the tip of the iceberg; this is the easy stuff to prevent.

When you allow it to add in music, create videos, adjust text, etc., you have the ability to prevent that. You’re even notified if you want to use it, and if you don’t take action, it’ll implement it. That is on you.

What Is This Rant Really About?

To do a very long-winded cut to the chase, this is about separating support and recommendations made by the platforms and the reps from what are industry and standard best practices.

No, your Google Adwords (it’ll always be Adwords), Bing Ads (I refuse to call it Microsoft Advertising), and Meta reps are not out to ruin you financially, with the sole goal of getting you to spend more, by any means necessary.

But, they have quotas and adoption requirements they have to meet. Companies that are not charities are money-making machines.

So, sales-focused recommendations, auto-apply functions, and AI-generated suggestions are designed to ultimately get you to spend more money.

It is your job as the digital marketer to recognize which function, recommendation, and capability your brand should utilize.

Do not be bulldozed – push back. If the platform or the rep makes a recommendation, ask them to show the work and explain how it will help you meet your goal. After all, the platform’s goals may not be your goals.

More Resources:

Featured Image: Roman Samborskyi/Shutterstock