This is going to sound semi-obvious, but we should be thinking a lot harder about direct relationships with customers and supporters.
Who knows how AI will shake out day-to-day over the next five years. Still, it seems clear to me that a Pew report on how people interact with Google’s “AI Overviews” in search results is one of those, “You had to study this to know that?” kind of reports.
The gist is what it sounds like: when people do a search, Google’s AI summarizer just spits out “an answer” to people. And people very rarely check the sources.
Google has repeatedly said this is false. That people click through more on search summary links. But this feels thin. They’ve never produced evidence, they hide stats on how often a site appears in AI summaries in their reporting tools, and it doesn’t pass the most obvious test: how often do you click a link?
A lot of searches on the web are for information gathering, and likely not ride-or-die. “How long do goldfish live?” Is one example. So long as the summary result is in the realm of plausible (“10-15 years”, or “about 12 years” or “up to 30 years with proper care”,) it’s all mostly accurate and good enough.
By most measures, Google traffic for a lot of sites will drop 70%. Large news sites and forums already see this.
Mark Twain’s quip that history doesn’t repeat itself, “but it sure does rhyme” comes to mind. Google’s pulling a Facebook and using all the stuff people have shoveled into the web for years and are re-selling it back to people. The difference is Facebook is Facebook’s own platform. No confusion there if you don’t like it. Just don’t use Facebook. But Google’s just assuming “the entire Internet” is their platform. Kinda hard to avoid that.
I’m of a mind several things are going to happen and maybe should happen:
- More sites are going to use paywalls or login walls. This works if you have a large following already. Traffic will still drop a lot, but overall revenue will be flat or up. This will be annoying for a lot of people because we’re so used to “getting things for free”, but ads pay the bills and no views = no ad impressions.
- We’re going to turn back to a pre-web interaction model where more businesses use more direct interaction. Think personal phone calls, mailed letters or postcards, etc.
- Email, which I have long called “the cockroach of the Internet” still shines. People own their own lists, you can mostly import/export those lists around, and the direct interaction is still very high at around 30-50% (compared to 1-5% on Facebook).
- More sites are going to turn to smaller, internal communities. Think blog comment threads, Slack or Discord groups, self-hosted forums, etc.
Digitally, more sites and organizations are likely to turn to:
- More live QA sessions, like Zoom hangouts or YouTube livestreams.
- “Exclusive stuff”, like bonus videos, podcasts, guides, kits, gifts, etc. for people who subscribe/pay/donate, etc.
Social media has a role in all this, too, but “history sure does rhyme.” It’s malpractice to say, “Welp, time to invest all our attention in growing our LinkedIn followers.” You don’t own those platforms, and the algorithms are not our friends.
This is why Facebook’s audience is collapsing, more people spend more time in DM chats/threads, Instagram has peaked, X has cratered, and social platforms are splintering.
In some ways this is good and kinda exciting. Not having a single dominant player is usually better for everyone long-term. But it’s unsustainable for most people to maintain high-quality accounts across five or six networks.
So we’re going to have to think about what arenas are worth fighting in. This notion of “We just post to such-n-such because ‘everyone’s there’” is over.
In my mind, I’m starting to think, “How would you grow a business or supporters if it was 2005 and newspapers didn’t exist?” That’s about what 2025 feels like right now.
