The Platform Squeeze

The WSU network team in 1993 did not stop using computers when Doom saturated their bandwidth. They changed how they allocated it. They changed which processes ran during peak hours. The solution was not to abandon the infrastructure. It was to adapt how they used it.

Every story in this week’s Pulse follows the same pattern. A platform changed. The existing setup started breaking. Someone had to decide whether to panic or adapt.

Every major digital marketing platform right now is in a power struggle between two forces: automation and advertiser control. The platforms are winning.

Google wants you in Performance Max. Meta wants you in Advantage+. Both systems use AI to decide who sees your ads, when they see them, and which version of your creative gets shown. You set the budget and the goal. The algorithm handles the rest. This can work. Meta’s Advantage+ Shopping Campaigns have shown real results for e-commerce brands. Some advertisers report 20 to 30% improvements in return on ad spend compared to manual campaign setups.

But here is the problem. When the algorithm makes every decision, you lose the ability to learn from your own campaigns. You stop understanding why something worked. You just know that it did, until it does not, and then you have no idea why. That is not adaptation. It is the same panic response my boss had in 1993, except nobody is lobbying anyone or changing how the work gets done.

The marketers navigating this well are doing what the WSU team did. Using the new system while staying in control of the outcome. Feeding the algorithm strong creative inputs, testing enough variations to understand what their audience actually responds to, and auditing results on a weekly schedule. They treat the AI like a capable junior employee. Fast, useful, not in charge.

The TikTok story is not really about TikTok. It is about what happens when you build your marketing operation on infrastructure you do not own. Thousands of brands woke up with zero access to their primary channel. The brands with email lists, strong SEO, or audiences on other platforms were uncomfortable for 24 hours. The brands with nothing else were in crisis mode with no clear path out.

The Semrush data on AI Overviews is significant and getting more significant. For years the formula was simple: find a question, write the best answer, rank on page one, get traffic. When Google answers the question before a user ever clicks, the traffic does not come. The brands that will win in search over the next two years are the ones shifting toward content where the reader needs to make a choice, not just get an answer, and building enough brand authority that users search for them by name.

The common thread across all of it: platforms change without asking permission. The network at WSU ran fine until Doom showed up. Then it did not. The fix was not waiting for Doom to go away. It was changing how the network worked.

Here is what to do this week:

1. Run automation with guard rails. Give Meta Advantage+ and Google Performance Max at least three to five creative variations to test, set clear conversion goals, and check results weekly. Let the machine optimize delivery. You keep the strategy.
2. Audit your top ten traffic pages right now. Identify which are informational (“how to,” “what is”) versus decision-stage (“best,” “vs,” “review,” “near me”). Prioritize building out decision-stage content. That is where search traffic is holding in the age of AI Overviews.
3. Pick two social platforms where your audience actually lives and go deep. Whatever you post, use it to drive people to something you own: a newsletter signup, a free resource, a community you host. The WSU team did not keep running Doom-vulnerable processes indefinitely. They changed the architecture.