• by Joe Mandese , Yesterday

I hate reporter notebook round-ups as much as the next guy, but so many relevant things have piled up, I'm using the format in this post to catch up. Let's start with the vibe at two important events I covered last week: MediaPost's Planning & Buying Insider Summit and the Association of National Advertisers' Measurement and Analytics Conference.

And in a surprising way, discussions at both of them blended together in a way for me that I'd like to call "multimodal planning."

I didn't know the term while covering the ANA conference in Chicago, but over the course of the brand marketer discussions I heard many of them articulate the same state-of-the-art method for measuring and adjusting their advertising, media and marketing plans and activations. As I listened to them, I was thinking of the term "tripartite," because it broke down into three methodological buckets -- marketing mix modeling, incrementality testing, and attribution modeling -- implemented simultaneously to determine the best course of action.

Never mind that the brand analytics experts said the three methods often produce different results, and even more infuriating, are all dependent on the quality of the data that go into them, they concurred that is the current state-of-the-art.

That's a lot of heavy lifting, and signals just how complex and multivariate the business has grown from the quaint days of pre-modeling measurement, but that's where we're at.

Still, the term "tripartite planning" wasn't doing the trick for me. That's when I listened to a sponsor presentation at the MediaPost summit later in the week in which an executive from Dstillery discussed the platform's "multimodal" approach, and it clicked for me. That's the world we live in now -- multimodality -- and we're constantly inducing and deducing insights from multiple sources of data simultaneously in an effort to understand what's actually going on.

So let's call it multimodal planning, because frankly, I was never a big fan of terms like "omnichannel planning," etc., which imply you're actually looking at everything. That's impossible. and the trick with multimodal inputs, is to ensure you're looking at the most important ones, and that you're post-analyzing them in the most scientific and meaningful way.

Interestingly, one of the presenters at the ANA conference actually had metric for how much all those multimodal metrics have improved the science -- as well as ROI -- of marketing.

"It was over a hundred yeas ago that [Philadelphia retailer John] Wanamaker allegedly gave us this quote that I'm not going to repeat, because it gets overused, but the problem is the "I don't now which half [of my advertising] is actually wasted," Measured's Will Post reminded the ANA attendees of the century-old mantra about half of ad spending being wasted."

"A hundred years," he continued, noting, "and we've got that narrowed to only 42% that's not incremental today."

Sheesh, so I guess we'll all have to paraphrase Wanamaker going forward as saying, "Forty-two percent of my advertising is wasted. The trouble is, I don't know which 42%." Or some improved percentage better than that, thanks, presumably to advances in AI and/or data and marketing science.

Meanwhile, during MediaPost's "Planners of Tomorrow" panel in Nashville, Harmelin Media's Dan DeLozier prompted what could well be the best description of what the current and near future of planning is in an era of generative and agentic AI.

"Show of hands, who has heard of the term vibe coding," DeLozier asked the summit attendees, going on to explain, "The idea is people who are not coders -- have never coded anything in their life --use a large language model and just natively describe what it is they're looking for, and it will produce code and output the thing you're trying to do with no coding background.

"We took that concept and we applied it to planning to an extent.

"Our normal process is we reach out to an activation team and they come back with a proposal. That could take a week. The client doesn't want to hear that it will take a week. So how do we get from budget to plan faster?... We built a tool that sort of lets you vibe code your audience. You can describe in natural language the audience you are looking for and via API pipelines we can actually go in and it will surface obvious audiences that sort of fit that vibe code description. But it will also surface things that you maybe didn't think of and build out an audience segmentation for you across all the platforms you're considering."

DeLozier said that manual process could normally take a week, but by working with Harmelin's vibe planning tool it would take only "one hour or two hours."

You can watch the full panel discussion embedded above, and all of the MediaPost summit sessions on our events page on YouTube here or by clicking through the summit's agenda here, but you can watch the Planners of Tomorrow" discussion embedded above.

I'm sorry to say the ANA's conference videos are behind a members-only wall, but I'll follow up with a couple more sessions that deserve their own explicit drill down shortly.