Marketing teams went from roughly $1,200 a month on AI tools in early 2025 to about $3,400 by early 2026, while only 13% say they fully trust AI output without a human checking it. The spend tripled and the trust stayed flat according to Anastasia Barba’s buyer’s guide for PMMs.
Her filter is to name the specific task the tool removes, and if you can’t say it in one sentence, you aren’t ready to buy. Ignore the model name, the demo polish and that competitor who has already signed.
I’d suggest adding one more… Make the vendor draw the mechanism, not the outputs or the case study, the path from input to answer. Anyone with a real one is happy to walk you through it.
Worth Your Time
- Three sites own three quarters of Google’s forum results (Ahrefs): Ahrefs checked 145 million US search results and found the Discussions and Forums block in 11.7% of them. Reddit appears in 83.9% of those blocks worldwide, Facebook in 38.3%, Quora in 32%, and if your category has an active subreddit, Google already picked who speaks for you there.
- 53% of marketers still point paid traffic at the homepage (DemandGen Report / Unbounce): Marketers confident in their landing pages were 4.5 times likelier to significantly beat ROI targets, 31% against 7%. Wild to me that half the field is skipping this optimization standard.
- Kroll says your category is worth more than your discipline (SaaStr / Kroll): Public software above 20% growth trades at 7.2x revenue against 2.4x for shrinking companies, while EBITDA margins stay flat across every bracket. The uncomfortable read: at an identical Rule of 40 score of 38, cybersecurity trades at 5.2x and marketing software at 2.1x.
- Owner.com’s customer research was three months old and completely wrong (SaaStr): Discovery interviews said restaurant owners feared AI; today 83% of new customers start inside an AI product. Put a decay date on your research the way you would on a pricing study, because one quarter is now long enough for a finding to flip.
- AI can organize your research, it can’t be your research (Nielsen Norman Group): Rachel Krause’s contrast carries the argument: a model invents a user who wishes the app would ask before swapping items, while real research finds the one whose oat milk became whole milk and got charged before anyone told them.
- Retention is a product marketing problem (Product Marketing Alliance): Thumbtack’s Jennifer Dewey reframes low-frequency retention as finding the high-frequency need sitting next to the rare transaction, the way home-value tracking props up a purchase people make twice a decade.
Wondering where your growth engine is leaking? The Growth Gap Diagnostic is where I’d start. Or just hit reply and tell me what you’re stuck on.