Two colleagues send you results for “quiet coffee grinder.” One ran the search last Tuesday and the other ran it today. Before describing what appeared or disappeared, establish which saved items are actually the same. You can do this with the copies you already have. Open your search history
Write down what each search covered
Open both results and label them A and B. Record the network, query, filters, sort, capture date and the page or batch you have. If one search sorted by comments and the other by recency, include that difference in the first line of your notes. It changes what a missing thread can mean.
Count links before interpreting the posts
Keep the raw item count. Make a second list with each exact post URL once per result. Put rows that have no URL in a separate group. Matching a display name or a similar headline can merge two different posts. If a redirect or tracking parameter makes two URLs look equivalent, open both before recording that as a separately checked match.
| Saved rows | What the rows support | |
|---|---|---|
| Result A | A contains five rows. Two point to the same post P. The others point to Q, point to R, and have no URL. | Three distinct links can be compared: P, Q and R. Keep the unlinked row in the five-row total. |
| Result B | B contains three rows linking to Q, R and S. | Q and R are exact matches across the two results. S is visible only in B. |
| What changed in these copies | Two shared links, one only in A and one only in B. | You can say that P was absent from the captured B rows. To find out whether the original post was deleted, you would have to check P itself. |
Compare what the samples actually say
Open one shared post in each saved result. Quote the sentence you used so a teammate can inspect both copies. For a theme that appears only in B, cite a B item and check A for the same idea under different wording. You might find a new question about grinder vibration in B while the overall balance of conversation remains unknown.
Save a short handoff with the two result references, shared-link count, unmatchable rows and a cited finding for each side. If the inputs came from different networks or have almost no comparable links or text, write that down and stop. A difference in two ranked samples cannot establish a change in demand.
For a closer reading of one result, use Analyse a saved result. For the two-result side-by-side format, see the Compare two saved results preview.