Reddit tests scale, LinkedIn automates: the July 2 brief

Test budget got a cleaner price tag today. Reddit made experiment design more accessible, LinkedIn pushed more creative variation into Campaign Manager, Google tested AI text inside ads, and HubSpot moved deeper into intent capture. Platforms want more of the decisioning layer. Your margin depends on which decisions you still audit.

Reddit puts a floor under ad testing

Reddit made Split Testing generally available on July 1, giving advertisers self-serve A/B experiments inside Ads Manager with audience splits, controlled spend and winner calls at 65% confidence, according to Reddit for Business. The catch matters: Reddit says the feature has a $1,000-a-day minimum spend.

That makes this a CAC tool for advertisers already spending enough to learn cleanly. Reddit says beta data found 4 out of 5 split tests identified a winning ROAS variant, with templates for Reddit Max, automated targeting, CBO and creative.

The money angle is waste avoidance. Without a controlled split, a campaign manager sees two date ranges, two creative sets, and a dozen outside variables pretending to be a decision. With one variable changed, the answer is narrower.

Smaller accounts should not force the spend just to use the feature. Put the same discipline into cheaper channels first. Our rule from Google Ads cleanup still applies: isolate the variable before trusting the winner.

LinkedIn shifts B2B creative work into the machine

LinkedIn introduced new creative tools for ads on July 1, including Brand Kit, Draft with AI, ad personalization, AI ad variants and Flexible Ad Creation, according to LinkedIn Marketing Blog. Flexible Ad Creation mixes supplied images, videos and copy, then shifts delivery toward better-performing creatives as signals emerge.

The money lever is creative throughput. B2B teams often under-test because every headline, intro and variant takes human time. LinkedIn is trying to compress that cost. It also says flexible ad creation generated roughly 7% more creative options.

The failure mode is false confidence. More variants do not mean more learning if every variant says the same thing with a different adjective. The operator still owns the offer, the audience promise and the landing page.

For B2B advertisers, separate message testing from format testing. Group variants by angle: pain, outcome, proof, urgency. If one angle wins, feed that learning back into your landing page, not just the next ad.

Google tests AI summaries inside sponsored results

Google is testing AI-generated summaries below Google Ads descriptions, Search Engine Roundtable reported on July 1 after screenshots appeared on X from Darcy Burk. The unit included a disclaimer that Google AI responses are generated independently and can make mistakes, according to Search Engine Roundtable.

This one hits conversion rate. Search ads are expensive because the click usually arrives with intent already formed. If Google inserts its own generated comparison or caution inside the sponsored result, the ad is no longer just your copy plus extensions. It becomes your copy, Google's summary, and a trust caveat in the same unit.

That can help or hurt by category. For commoditized services, an AI note that reframes price or risk can pull clicks away before the landing page works. For high-trust categories, it can pre-qualify users who would have bounced anyway. The dashboard can misread the change as creative fatigue when the ad surface changed.

Watch branded and non-branded search separately, and annotate any CTR or conversion-rate change around June 30 and July 1. This connects to AI Overviews and lead flow: Google is inserting more interpretation between the query and the site.

HubSpot buys closer to the intent signal

HubSpot's Warmly acquisition, covered by MarTech on July 1, points to the next CRM battleground: identifying who is showing buying intent, not just which company visited a page. Warmly is known for identifying individual website visitors and using AI agents to help sales teams engage them.

The money lever is pipeline speed. Anonymous intent is useful, but slow. A company visit creates a research task. A person-level signal can create a routed follow-up, enriched record and next-best action. That compresses the gap between "someone is looking" and "sales has a reason to reach out."

The risk is over-routing weak signals. Not every visit is pipeline. If AI agents turn every anonymous hit into activity, the CRM looks busy while reps chase noise. Better: high-fit account, high-intent page, recent visit gets routed. Everything else stays in nurture.

This is why CRM keeps moving away from contact storage and toward follow-up mechanics. We see the same budget logic in CRM selection for agencies: the tool matters less than whether lead, booking and follow-up data live close enough to act quickly.

What to watch

Cloudflare's new default AI-crawler policy could raise the value of owned audience and direct traffic if more publishers restrict unpaid scraping.

Google AdSense Ad Intents adding Gemini-generated content is worth watching for publishers because monetized answers can change the revenue split on informational pages.

TikTok's Agentic Hub is still early, but if it turns creator and commerce workflows into agent-managed campaigns, the next paid-social fight is who controls the brief.