Kill bugs fast

Can AI devise advertising strategies? I ran an experiment.

A few months ago, I gave Claude one of my favourite case studies to tackle. Jon Steel's Porsche turnaround story from the 1990s. Porsche sales had tanked 59% by 1993. There were several reasons for the decline. But one was this: People still loved the cars, but they hated what they represented. Wall Street greed. Gordon Gekko types. 80s excess. They didn't want to be seen as the kind of person that drives a Porsche.

Among other things, Porsche came up with a very successful advertising campaign that helped reverse this. I wondered how Claude would tackle it.

Claude suggested five approaches:

1. Reconnect with engineering heritage. Focus on German craftsmanship and driving experience rather than status.
2. Profile the "true" Porsche enthusiast, create clear daylight between status-seekers and passion-driven owners.
3. Emphasize racing heritage, shift the narrative from Wall Street to racetrack.
4. Take a self-aware approach, acknowledge the stereotype but cleverly subvert it.
5. Position as timeless versus trendy, frame the 80s association as merely a passing trend.

All seemed like directions worth exploring. I was pretty impressed. And to be fair to our robot friend, this was only after a few moments. If I were doing this again, I'd give it more context and wrestle with it more, and try a few techniques I've done since with it, such as river-jumping.

Afterwards, I told Claude about the actual campaign that worked, from Jon Steel. The turnaround campaign was "Kill bugs fast" created by Goodby, Silverstein & Partners. I asked Claude if it would have gotten to this with more time and prompting.

Claude admitted it couldn't have made that jump: "Honestly, I might not have arrived at 'Kill bugs fast' even with more time. The approach is brilliant precisely because it's unexpected and doesn't follow conventional strategic reasoning. It represents a creative leap that bypasses the entire problem rather than solving it directly."

Again, to be fair, Claude's direction of "Profile the true Porsche enthusiast" is not far off this. Digging into a true enthusiast could have led to something about capturing the sheer joy and feeling you get from driving a Porsche if working with a creative team. Maybe.

That said, "Kill bugs fast" came from someone who knew that feeling. Somebody that understood that Porsche wasn't just about status or engineering heritage. Our robot friends can analyze what customers do. They can spot patterns in behaviour. They can even generate strategies based on perception gaps.

They just can't experience what customers experience. Every good marketer has these moments. When the breakthrough insight comes not from analysis but from experience. From using the product. From feeling the frustration when it doesn't work, from watching customers in the shops get confused. It's more than sentiment analysis or journey mapping. It's often about human understanding.

Maybe it's the difference between knowing about something and knowing something. My take on this is that AI might get us to something half-decent. But still needs a human behind the wheel.

For now anyway.