The feed
Martech news and analysis. Every story says what changed, why it matters and what marketers should do next.
At its Agentic World summit on October 6, 2026 Treasure AI launched Personalization Studio for real-time website personalization without a developer, and an engagement-based price in place of per-email pricing.
Why it matters
Outcome-based pricing shifts risk to the vendor but depends on how clicks and engagements are defined and counted.
A MarTech guide of October 9, 2026 says a BAA covers only the vendor that signs it, that HIPAA regulates entities rather than data, and that tracking class actions and FTC health-data rules still apply to non-HIPAA brands.
Why it matters
Health-related brands that use ad pixels or retargeting can face liability even when a vendor holds a BAA.
On October 1, 2026 Advocate General Laila Medina advised that a diocese which struck through a man's baptism record but left it legible had not erased it, and must show compelling grounds to keep it (Case C-12/25).
Why it matters
If the Court follows, redaction that leaves data legible will not satisfy deletion requests, which matters for CRM and consent records.
Italy's Garante fined IQVIA Solutions Italy EUR 7 million (decision of September 23, announced October 2, 2026), finding a database of about one million patients from roughly 800 doctors was pseudonymised, not anonymous.
Why it matters
Marketers and data teams buying or sharing "anonymised" datasets cannot rely on the label alone.
Reddit announced on September 30, 2026 a timeline that ends public Data API access in March 2027: new requests stop October 31, RSS ends November 13 and unregistered apps begin losing access January 12, 2027.
Why it matters
Social listening, research and monitoring tools built on Reddit data will need a licensed route or a Devvit rewrite.
Lunio's October 6, 2026 report on more than five million education-sector clicks found competitor-brand campaigns at 9.3% invalid traffic against 4.7% for generic ones, and a sector rate rising from 4.08% to 5.42%.
Why it matters
Invalid clicks distort conversion rates and cost metrics that feed bidding and budget decisions.
Advertisers can now start Conversion Lift studies for Search and Performance Max without a Google representative, with a $5,000 minimum budget. Google also changed how a view counts for Demand Gen display ads.
Why it matters
Cheaper, self-serve experiments make it easier to calibrate platform-reported results with real holdouts.
A study published September 30, 2026 analysed 1,934 Kroger ads on Meta and DV360. It says creative following the recommendations converted about four times better and that reallocating spend could double conversions.
Why it matters
If creative drives a large share of results, scoring ads before launch could shift budgets, but the evidence so far is observational.
A September 2026 essay by measurement lead Karan Dhir revisits a 2023 Marketing Science study of 663 Facebook experiments: median lift for upper-funnel campaigns was 29% in randomised tests but 83% under double machine learning.
Why it matters
If observational estimates overstate lift, budgets can chase numbers that look causal but are not.
A preprint by Niklas Heusch, submitted August 21, 2026, finds a conventional MMM reported paid search ROAS of 10.61x against a true 4.20x on synthetic data, while a method built on four geo-experiments recovered 4.14x.
Why it matters
If MMMs inflate channels that are correlated with demand, budget shifts based on them can be wrong.
