Week 11 Lab: Feed reverse-engineering

A lab where you deliberately engage with a narrow slice of content on a fresh or logged-out feed and log how quickly the recommendations shift

13 May 2027

Before the lab#

Pick a recommendation feed you can use in a logged-out or fresh-account state, and a narrow, specific slice of content to chase — not “cooking” but “one very specific cuisine,” not “sport” but “one team’s results.” A narrow target makes the shift easier to detect.

In the lab#

Interact only with content matching your slice — watch it fully, like it, whatever the platform’s strongest engagement signal is — and log the feed’s composition every few interactions. Most people find the shift happens faster than they expected, and that a few interactions in a genuinely different direction are enough to pull it back.

Discussion compares slices across the room: some topics shift the feed almost immediately, others resist for much longer, and guessing why is more useful than confirming it, since none of you have access to the actual ranking system.

Afterwards#

Nothing here is separately assessed, but this experiment is a fast way to generate first-hand evidence for anyone whose final project touches recommendation, and the pace of the shift you measured is a specific, citable number a general claim about “the algorithm” isn’t.