It felt like January lasted for more than 5 weeks, and yet all of a sudden March is now just around the corner. Meanwhile, the wheels keep turning in the wonderful world of Digital Marketing.
We had an interesting, cautionary tale from Canada this week, where AirCanada fell foul of a rogue AI chatbot (and then by just not doing The Right Thing).
Meanwhile, LinkedIn is doubling down on trying to drive B2C advertisers to the platform, and I came across some interesting thinking around how Google deals with the amount of nonsense content out there.
What happens when your AI goes off script?
Very interesting case in Canada this week, with AirCanada falling foul of a ruling whereby they had to honour a refund to a bereaved passenger – the catch being, the refund policy didn’t technically exist. Instead, it was generated by their AI chatbot.
Quite apart from the point that we now live in a world where machines appear to have a stronger moral code than us humans (maybe the impending AI-pocalypse is a good thing?), it makes for an interesting case study in adequately training your AI tools to the same, if not a greater extent, than you would do the employees that they are replacing.
Also, if your AI tool stuffs up and it will only cost you relative peanuts to make it right, maybe do that instead. The negative PR and resource/time AirCanada is now devoting to addressing this surely outweighs the cost of just doing the right thing. But this isn’t a PR newsletter…
How does Google rank mediocre content?
There’s no shortage of marketing opinions out there (hi), but sometimes you get some really interesting takes that make you stop and take a second. This week I came across one on Search Engine Land seeking to address the question of What should Google rank in Search when all the content sucks?
There’s a few things at play here. Firstly, the hypothesis is largely spot on – time and time again I have googled a simple question, and waded through keyword-stuffed filler intros before I’ve actually got to the crux of the matter. Second, as the article notes, this isn’t an AI thing – humans have been perfectly capable of writing generic, unintuitive content for years now.
And thirdly, so what does this mean for you? Well, Google has hinted that changes are coming. This, combined with the fact that we all know landing pages need to get to the point and drive impact, poses a great opportunity to actually read the content you’re pulling together, and ask yourself that same question – so what? What is this adding to the person finding it? You’ll find that answering that question can still render content SERP-friendly (depending on how you structure it) AND you can future-proof yourself for the changes that are apparently coming.
LinkedIn highlighting B2C opportunities
The use of LinkedIn as a consumer-facing channel has been discussed for a while, and is something I’ve worked on for a few clients. The general thinking from my experience is that it can serve as a powerful tool to reach specific audiences, but it can also be a very expensive way to build awareness. It’s generally best to try it and see what, if anything, works for you.
However, LinkedIn is clearly making a push to try and bring new advertisers to the platform. They recently ran a Meet the Member survey, the results of which indicate that LinkedIn users are likely to be wealthier, own larger assets, and like spending on high-end products. Which… well, if you didn’t know that (or at least assume that), then you might want to go and brush up on your social networks.
But in any case, I do think it’s worth trying if you’re looking to shift units of a high value B2C product – but for anything with a minimal CPA, I’d steer clear.
Further Reading
The New York Times explores the growing impact of Instagram as a ‘news’ source, despite the fact the platform itself is trying to de-emphasise political content ($£)
Meta is continuing to double down on Threads, with experiments to cross post from Facebook.
Reddit has signed a deal, reportedly worth $60m per year, with an unnamed AI Company. The deal means that the AI firm can incorporate Reddit user responses into its model, enabling the system to provide more ‘human’ answers. One to watch.
Finally, a recent opinion piece on Slate explores whether recent developments (such as Joe Biden joining) are signalling the imminent decline of TikTok.
That’s it! If you enjoyed this article, then please feel free to tell your friends, say hello or if you’re feeling particularly generous, to buy me a coffee.