Four marketing technology announcements on September 18, 2026 addressed a single question most CMOs cannot answer with confidence: how much of your reporting rests on data you can describe. Infillion said it will acquire location intelligence platform Foursquare, Taboola made an offer for London finance ad network Dianomi, Fingerprint launched a public directory of AI bots and crawlers, and Handvantage published a free procurement handbook for agentic AI. Each release touches an input behind a marketing number - store visits, publisher relationships, session classification, and vendor data access - and none published a customer outcome.
Infillion's deal adds Foursquare's visitation data to its existing Catalina purchase intelligence, creating what the company describes as a privacy-safe way to measure whether advertising drove store visits and real-world purchases. Foursquare brings more than 100 million points of interest across 200 countries, 16 billion human-verified check-ins, and aggregated coverage of 250 million United States devices. Rob Emrich, founder and executive chairman of Infillion, told Axios the acquisition raises Infillion's revenue and headcount by roughly 20% to 25%, with most of Foursquare's approximately 150 employees joining Infillion's 650 full-time staff. Foursquare will keep operating as an independent brand and continue selling data to customers who buy media on competing platforms.
Store-visit measurement correlates device presence near a location with an ad exposure. That establishes a relationship between two logs. Incrementality - whether the visit would have happened without advertising - requires a holdout group that sees nothing. Infillion's release published no lift figure, no holdout design, and no price. A buyer evaluating the combined dataset at renewal should ask for the test design, the size of the control group, and the confidence interval before treating a visit report as proof that media caused the visit.
Taboola moves into finance advertising with Dianomi offer
Taboola announced its offer for Dianomi under Rule 2.7 of the UK City Code on Takeovers and Mergers, with completion expected before the end of 2026. Dianomi works with more than 600 advertisers including Charles Schwab, Invesco, and Bank of America, placing their ads with publishers such as Reuters, CNN Business, The Times, and The Wall Street Journal. The network will run through Realize, Taboola's performance advertising platform, which the company reports reaches more than 600 million daily active users. BusinessCloud reported the deal is worth up to 27 million pounds, with Dianomi posting first-half 2026 revenue of 13.4 million pounds, a 2% year-over-year increase. Taboola's own release disclosed no financial terms.
Adam Singolda, chief executive of Taboola, said the deal connects large financial brands with their target audiences. For marketers running campaigns through Dianomi, the transition raises practical questions about publisher lists, rate cards, and reporting continuity that should be confirmed before the deal closes.
Fingerprint's Bot Directory classifies AI traffic
Fingerprint launched Bot Directory as a continuously updated, searchable public repository of AI tools, bots, and crawlers, listing each entry by identity, provider, and purpose. Operators can test their own agents for Web Bot Auth compliance and submit them for inclusion. Valentin Vasilyev, co-founder and chief technology officer of Fingerprint, framed the launch around telling helpful automation apart from malicious activity as agents act on behalf of people. Fingerprint reports identifying more than 1 billion unique devices a month across more than 6,000 customers, including Plaid, Booking.com, Dropbox, Trustpilot, and Checkout.com. The release states that bots account for the majority of web traffic and that nearly half of technology leaders face AI-driven fraud attacks, without citing a source for either figure.
Bot classification determines two marketing outcomes at once. A crawler your team blocks never reads the product page you optimized for Answer Engine Optimization, and an automated session your analytics counts inflates the denominator under every conversion rate you report. One allow-and-deny list governs both. In most companies the security team owns that list and the marketing team owns the number it distorts. Agentic Commerce Protocols raise the cost of that split, because a buying agent arriving at checkout on a shopper's behalf looks like automation to a fraud rule written three years ago.
Handvantage publishes agentic AI procurement guide
Handvantage, a Surrey, British Columbia consultancy for AI security and governance, published the Agentic AI Procurement Handbook, a 43-page vendor-neutral guide available with no email form. The handbook organizes the questions procurement teams inherit from security architects: where company data goes, who approves what an AI agent does, and what record exists after the agent acts. Marketing signs more AI software contracts than most functions, so those three questions belong in the marketing contract review before they reach an AI Governance Board.
Work an example. A retailer runs 2 million sessions a month and reports a 2.4% conversion rate to its board. If 30% of those sessions come from automated traffic sitting in the denominator, the human conversion rate is closer to 3.4%, and a year of A/B tests that moved the reported rate by 0.1 points measured a mixed population. Correcting the denominator costs a bot-classification field in the analytics warehouse and one engineering sprint. Those figures are illustrative, and none of the September 18 releases published them.
Why this matters for marketing and creative professionals
Pull the share of sessions your analytics tags as automated, then recompute the conversion rates you have been reporting. A rate built on a mixed population makes small test wins unreadable. Assign one owner for the bot allow-and-deny list across marketing and security this week - Fingerprint's directory gives one public reference, and the same list decides which AI systems read your product pages and which sessions land in your conversion denominator. For teams building AI Marketing Leadership Courses into their planning, the Handvantage handbook's three questions - where data goes, who approves an agent action, and what record survives it - belong in the master services agreement template before the next vendor signs.
Your membership also unlocks:
