Learn · How-To

How to Verify Rank Tracker Data Against Search Console

On 10 October 2026 Nozzle, a rank-tracking vendor, reported that since mid-September Google has been returning altered results pages to tracking tools, and Ahrefs and Semrush have since described the same pattern. Google has not commented. You do not need to settle the cause to protect your reporting: you need a repeatable way to decide whether a position change in a tracker is real before it reaches a client or a budget decision. This guide uses Search Console, which records Google's own impressions and clicks for real searches, as the reference.

1. 1

   Freeze the evidence before you change anything

   Export the tracker positions for the keywords that moved, with dates, and note the device, country and tracker's collection time if it shows one. Save a copy of the report as it stood. Once you start re-checking or the provider cleans its data, the original reading may be overwritten, and you will want it for the written question to the provider in step 7.

2. 2

   Compare against the keyword's own history, not yesterday

   Nozzle judges a results page as altered when fewer than half of the sites that normally appear for that keyword show up on it. Do the same by eye: open the tracker's history for the keyword and ask whether the domains in today's top ten are the ones that usually appear. Comparing only to the previous day can mislead, because the previous reading may itself be affected.

3. 3

   Check page two and look for the wrong kind of domain

   Nozzle reports that page two was altered about one and a half times as often as page one, so a drop that appears only below position ten deserves the most suspicion. Also look at who replaced the usual sites. Nozzle found social platforms and forums, including Facebook, Instagram and Quora, appearing more often on altered pages; Facebook was in the top ten on 29.5% of altered pages against 21.9% of the same keywords' other pages since 13 September. A top ten that suddenly fills with them is a prompt to check, not proof.

4. 4

   Wait for the next reading before you call it a drop

   Search Engine Journal, relaying Nozzle's data, reports that 80% of altered spells lasted a single reading, 15% lasted two and 5% lasted three or more. A keyword that falls and returns within a day or two fits that pattern, while a genuine ranking change usually persists. Set a rule for your own reporting, for example "no movement is reported until it has held for three consecutive readings". That is our suggestion, not Nozzle's, and for a weekly tracker three readings means three weeks.

5. 5

   Pull the same query and page from Search Console

   In Search Console, open Performance, filter by the query or the page, and set the same dates as the tracker movement. Compare impressions, clicks and average position with the period before. Nozzle's own guidance is that if Search Console still shows impressions and clicks for a site that appears to have vanished, the site probably did not change. Low-volume queries may not appear because Search Console withholds rare queries, so check the page filter as well.

6. 6

   Classify each keyword and decide what you will say

   Sort every keyword that moved into three groups. Confirmed: the tracker and Search Console agree and the change has held. Suspect: the tracker moved but Search Console did not, or the move lasted one reading. Unknown: too little volume to confirm. Report confirmed movements as movements, hold suspect ones for another cycle, and say plainly that unknown ones cannot be verified yet.

7. 7

   Put your questions to the provider in writing

   Ask whether they detect and reject altered results and how, what changed in their collection after 13 September 2026, whether historical data points were recalculated or removed, and how a position is flagged as unverified in the export. Ahrefs said on 22 October that it added filters and is cleaning past data, and Semrush said on 27 October it excludes clearly invalid results. Ask each provider for the specifics for your keywords and keep the written answer on file.

8. 8

   Annotate the report so the reader knows how to read it

   Add a short note to every rank report that covers the period. Suggested wording: "Positions from third-party rank trackers are indicative for September and October 2026 while providers filter altered Google results. Impressions, clicks and average position are taken from Google Search Console, which we treat as the reference. Movements marked confirmed have held across at least three readings and are matched in Search Console."