Your feed is a black box. This report turns your experience into evidence.
Reconstruct one short session with precise facts. Don’t guess. If you can’t verify
something, choose “Not sure.” You don’t have to be perfect. If a section doesn’t apply,
skip it. If you complete the full form, email us proof to receive monetary compensation.

Deep Feed Trace Report (Reverse-Engineering Grade)

Answer what you can. If you don’t know, choose “Not sure.” Do not guess.

0) Consent

1. I understand this report is used to study platform behavior, not to fix my account.
2. I consent to my report being anonymized and used in aggregated research outputs.
3. Optional: I consent to share screenshots/screen recordings with identifying info blurred.

1) Identify the exact environment

4. Platform

5. Surface you were using (pick one primary)

6. Device
7. Access method
8. Network conditions
9. Location context (choose one)
10. Time
MM slash DD slash YYYY
Local time started
:
Approximate session length
11. Had you just updated the app within 48 hours
Had you just logged in/out or switched accounts within 30 minutes?

2) Account state (signals that affect ranking)

13. Account age
14. Account type
15. Approx follower count range
16. Are you verified?
17. Any enforcement in last 90 days?
18. In the 24 hours before this session, did you do any of these? (select all)
19. In the last 7 days, did you post?
20. In the last 7 days, did you get a sudden spike in views/followers?

3) What triggered this report?

21. What are you reporting? (choose the best match)

23. When did you first notice it?

4) Feed trace: what you actually saw (the core)

You are going to reconstruct a slice of the feed. If you can’t do all 20, do at least 10
Feed Slice A (first 10 items shown)
For each of the first 10 pieces of content you saw, answer:
For Item #1:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #2:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #3:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #4:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #5:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #6:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #7:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #8:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #9:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #10:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
Feed Slice B (next 10 items)
For Item #11:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #12:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #13:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #14:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #15:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #16:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #17:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #18:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #19:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?
For Item #20:
24. Content type
25. Source type
26. Topic category (best guess)

27. Language

28. Age of post (if visible)
29. Your action on it
30. If video: watch time
31. Was it an ad or sponsor?

5) Ranking symptoms: what changed in composition?

Compared to your normal experience, did you see more or less of the following?
For each, choose: ☐ Much more ☐ Slightly more ☐ Same ☐ Slightly less ☐ Much less ☐ Not sure
32. Posts from accounts you follow
33. Posts from accounts you do NOT follow
34. Posts from small accounts (low followers)
35. Posts from big accounts (high followers)
36. Posts from your geographic area
37. Posts in your language
38. Posts matching your usual interests
39. Repeated creators (same creator multiple times)
40. Repeated topics (same theme over and over)
41. Political/news content
42. Sexual/“borderline” content
43. Violent content
44. Ads
45. Did the feed feel “stuck” (looping the same stuff)?
46. Did you see content you had already watched earlier?

6) Missing content diagnostics (the “absence” problem)

47. What exactly felt missing? (select all)
48. If a specific account was missing, how do you normally encounter them?
49. Did you search for the missing account/topic?
If yes:
51. Did the account/topic appear in search?
52. If it appeared, where was it ranked?
Did autosuggest show the query as you typed?
Did related searches look “steered” away from your query?

7) Your micro-behavior during the session (signals)

In THIS session, did you do any of the following? (select all)

8) Possible algorithm “reset” events

In the last 72 hours, did any of these happen? (select all)

9) Content production (if you posted)

If your report involves YOUR post performance, answer this section.
81. Did you post content within 24 hours of noticing the change?
82. Post format
83. Post length (if video)
85. Did you include any of these? (select all)
86. Did you edit it inside the app or upload from camera roll?
87. Were there any “sensitive” elements?
88. First-hour metrics (best guess)
Views:
Likes:
Comments:
Shares:
89. Did the post get pushed, then suddenly stop?
90. Did you receive any notification about the post (limited, ineligible, copyright, etc.)?

10) Controlled test (optional but extremely high value)

If you can, do a tiny experiment and report results.
91. Did you run a test?
If yes, what test type?

For the test:
101) Result:

11) Hypothesis prompts (to structure inference)

103. If you had to name the dominant shift, which is it?
104. Which is most consistent with what you observed?
105. Confidence level in your report
Why:

12) Evidence upload (optional, but best)

106. Upload any of the following (all optional):

Drop files here or
Max. file size: 1 GB.

    13) Final: one clean factual summary

    108. In 2–3 sentences, describe:
    Scroll to Top