All Intelligent Life

22 June 2023

The time has come for the Cog Blog to address the biggest problem / opportunity / elephant in the room our industry is facing; at least until the next cab on the rank comes along. The issue of artificial intelligence.

I would be the first to admit that when it comes to the technology this is a topic I know little about. I suppose that there’s the potential for AI to rule the world, to take our jobs and to cast us all into outer darkness, but there’s also the potential for me to run a marathon.

Neither seem very likely to me.

I’m no expert in how the technology works. In that, I like to think I’m aligned with 90% of the ad industry, 80% of the media industry, and 99% of online commentators.

I also know very little (and even that’s an exaggeration) about how broadcast signals are transmitted and received, how streaming actually works and how self-drive hybrid cars operate. I use all of these things quite satisfactorily.

I can see the great potential in AI; but then I witness most days how the reality fails to live up to the promise. I still get letters addressed to ‘The Occupier’; I still spend hours filling out forms in duplicate; I still get frustrated explaining the same thing over and over again to different departments of the same supplier.

Surely in a world ruled by big data and our AI-driven ability to harness it the end-execution should be a little better?

Even in our own little world, and only this week the ANA reported on the waste generated by programmatic buys, focussing in on the creeping insidiousness of made-for-advertising sites, or MFAs.

The ANA found that 15% of participants’ budget and 21% of impressions are accounted for by MFAs.

And that the average advertiser campaign ran on 44,000 sites. I don’t believe there are 44,000 sites that add unique reach, or any valuable frequency against any target market.

And remember – these are ANA advertisers, so the more sophisticated end of the market.

Maybe the ANA will revise its rather confusing and variable position on ad fraud whilst they’re at it. To remind you:

‘Juniper Research (estimates that) $68bn will be lost to ad fraud this year. ($68bn may be at the low end – the ANA in the US have quoted $120bn although they promptly ‘disappeared’ that figure when it started to attract attention)’.

Nick Manning’s The Media Leader column makes the point that the ANA work mirrors the principal findings of ISBA’s 2020 report into much the same thing.

Progress is it seems slow, possibly because the agency experts guiding their clients through this morass whilst avoiding the worst pitfalls are too busy thinking up the latest way to make money from the mess that is online media.

My point is if AI is all-embracing and smart why is the reality often so rubbish? How come we are still wasting millions / billions of our clients’ money by buying ads that reach an audience of bots?

Is AI a great threat to jobs? Will we enter a world within which all ads are written by Chatgpt or the next generation equivalent? Will all media plans and buys be automated, with budgets spent untouched by human hand?

It suits some people, in particular those with systems to sell to talk up the benefits that come from automation. And it suits our trade press to amplify the apocalyptic message – the end of the industry is news after all.

I see AI technology as a fuel to tools, a brilliant way for us to do our jobs better.

Take the creating of an ad campaign. Someone has a genius idea, in all likelihood sparked by a brilliant planner analysing all kinds of data and considering all sorts of insights.

I’ve worked with many great planners. You know how good they are when they say something to which your reaction is ‘yes, of course. That’s obvious.’ Whilst at the same time acknowledging that you, as a mere mortal would never in a million years have thought of that.

The marriage between data and creativity is helped hugely by an ability to analyse loads of stuff, very quickly.

That’s where the genius spark starts. It’s not in itself the genius spark.

I was encouraged to read this today in the WSJ, in which Omnicom discuss their new AI tool (and thanks to Kristin Engard for the post):

“Omnicom frames the move as a way to get smarter answers in less time, with more workers being able to use deeper data analysis — A virtual assistant to help ad agency employees with tasks across the advertising process such as compiling audience insights and building media plans.”

These things are tools. They help us do a better job, in multiple ways. But they’re not the job, at least not yet.

I wish more of us would use them to do a better job because right now we’re screwing things up pretty badly.