Sponsored
INSIDER SESSION: HOW TO CREATE A HOLISTIC BRAND EXPERIENCE
"Brand" is a big word, and everyone seems to be talking about it. But what does it mean in practice? A shinier logo? More display ads?
Parse notes
Sponsored
"Brand" is a big word, and everyone seems to be talking about it. But what does it mean in practice? A shinier logo? More display ads?
There's a spark of productivity panic in tech, pressuring engineers and executives to work longer hours and produce more code. Once seen as a way to “vibe code” casually, these tools now drive an expectation that everyone experiment with AI while maintaining existing workloads. Senior leaders using agents themselves raise output expectations. Studies show nonmanagers gain far less time savings, fueling “AI fatigue” and task expansion. While some teams report higher productivity, much of the output risks becoming disposable busyware (minor projects or features that consume time but add little real value).
WPP is launching a major restructure to become a simpler, lower-cost, AI-focused business. The London-based group will reorganize into 4 main divisions and regional markets. It is merging key ad agencies, including Ogilvy, VML, and AKQA, under a shared back office while keeping them separate. It is also reinvesting savings into AI and high-growth areas.
Small changes can massively shift results, with identical videos sometimes getting 10X more views. Format matters too, as greenscreen videos perform far better on Instagram while talking-head styles can do better on TikTok. Algorithms are unpredictable, so the best approach is to repurpose content across platforms and then test multiple hooks, captions, text, and formats. Iterate on underperforming videos instead of relying on a single version.
Build a foundation before running campaigns. Start with the audience, not the feature, and use segmentation by fit, behavior, lifecycle stage, role, and jobs-to-be-done to shape messaging. Gather insights from product data, support, CS, and sales. Use a shared calendar and intake workflow to manage requests, and apply priority tiers, QA checks, and post-send reviews to keep cadence and multi-channel messaging consistent. Treat ad-hoc sends as learning opportunities. Once the system works, use AI to speed repetitive tasks.
Leading with emotion instead of a heavy search form doubled conversions for a travel booking platform. User research showed visitors called the original mobile homepage boring and unengaging because the form appeared before any inspiring content. The winning redesign reduced the first step to one field and highlighted strong destination imagery and video. This change increased bookings by 103% during the test.
Sponsored
B2B buying is messy. Yet most ABM programs are still orchestrated at the account level—where buyer behavior gets averaged out, early signals get missed, and “personalization” becomes a best guess.
Sarah Adam, Head of Influencer Marketing at Wix, and Guy Dotan will discuss their experience scaling influencer campaigns and how to handle both organic and paid strategy. The webinar will take place on March 5 at 11 AM ET.
HubSpot workflows are normally limited to a single object, forcing teams to use APIs, custom scripts, or manual exports to connect Contacts, Deals, and Companies. Claude Code removes this barrier, letting you query across multiple objects in natural language without code. After installing the Claude for HubSpot app, you can open a workflow and ask Claude to update it using any objects. Review the draft it generates, then implement the changes directly in HubSpot. This enables faster and more flexible multi-object automation.
Google Discover reaches hundreds of millions daily, but its ranking system is mostly hidden. Content flows through a nine-stage pipeline. Missing og:image or images under 1,200px block or limit visibility, and dismissed items remain permanently suppressed. Collection-level filters can block an entire publisher before ranking. Discover uses 13 cluster types, multiple personalization layers, and about 150 concurrent A/B experiments with 51 feature flags, so similar users can see very different feeds.
Meta may have structural advantages over pure play LLM rivals because of its closed-loop ecosystem. Unlike Google, which risks disrupting search ads, Meta can experiment with AI answers inside engagement-driven platforms while preserving monetization. Social discovery behaviors, from restaurant recommendations to product research, are already happening inside Instagram and WhatsApp, creating an entry point for AI-powered search layered onto existing conversations. Local businesses could see renewed importance in optimized Meta profiles as discovery shifts from anonymous reviews to socially informed recommendations.
The Head of Marketing at Proton used Claude Code to write the bot, Anthropic API to process requests, HubSpot tokens to make updates, Railway to run it 24/7, and Supabase to log every action.
A breakdown of Storylane's marketing spend across 10 channels.
Google rotates results from 5-8 subreddits per query, so focusing activity on the most relevant communities drives stronger results.
GTM is the overall plan, a product launch is a marketing-focused subset, and product marketing connects messaging, insights, and enablement across both.