Somebody has to make the stock.

Signals
  1. 01

    Plan for AI to arrive as new budget lines, not just cheaper software

    Companies already using AI are spending far more on it, and budgets that treat AI as a software saving will understate what the work really costs.

    Menlo Ventures
  2. 02

    Find out which AI agents visit your website and whether they can finish a task

    Some of the visitors trying to buy from you are now software, and a form that a person can finish but an agent cannot is a lead you never see.

    Microsoft Clarity
  3. 03

    Check your search reports before you blame a falling click rate

    A drop in search numbers this autumn may be partly a measurement problem, and acting on it as a demand problem would waste budget.

    Search Engine Land
  4. 04

    Decide whether Reddit's search results belong in your paid plan

    Searches on Reddit tend to happen mid decision, so ads there may reach buyers closer to a choice than most social placements do.

    Reddit
  5. 05

    Ask what your retail media partners now know about purchases made elsewhere

    Retail media is moving beyond each retailer's own customers, and the brands that understand where the purchase data comes from will buy it more wisely.

    GlobeNewswire
  6. 06

    Treat AI answers as the default face of Google Search

    Google has confirmed that AI is part of Search itself, so visibility in AI answers is now part of basic search performance rather than an experiment.

    The Guardian
  7. 07

    Walk your own website as a buyer would, because most B2B sites lose people before they ask

    Most of a B2B purchase is decided before the first sales conversation, and a website that hides the next step loses buyers no report will ever show.

    PR Newswire
Trend watch

The work is getting cheaper, and the checking is getting dear.

Across this week's news, the cost of producing marketing output keeps falling while the cost of knowing whether it is right keeps rising: AI spending growing inside the companies already using it, agents visiting websites, search data distorted by automated traffic, new ad slots and new purchase data to evaluate, and AI built into Search itself. The teams that invest in people and systems able to check the work, by testing what agents can do, verifying numbers against real leads and sales, and training the next generation of judgment, will get the value the cheaper output promises.

From Magnet today

Somebody Has to Make the Stock

Marketing teams are cutting the junior seat because software now writes the first draft, but the junior seat was never really paying for drafts. It was where judgment got made, and the agents every team is adding now run on judgment.

Read the full article →