Rank Tracker vs Search Console: Which Position Number Should an Arabic Site Trust?

Two position numbers for the same Saudi site disagreed by up to 5.9 places in 28 days, and neither was wrong. We pulled Semrush Position Tracking and Google Search Console for one Arabic medical site on 19 October 2026. Trust trends and clicks first, and treat any single day of either tool as a prompt to check, not a result.

Update, 19 October 2026: We added a manual check. On 19 October we searched 20 keywords from this Semrush campaign (the 20 highest-volume first-page keywords among the first 40 the report returned) in a signed-out browser with Google’s Saudi settings and Arabic interface, from a computer in Cairo. Semrush put Delta at position 1 for 16 of the 20 that day, with a mean of 2.1. Our count of organic result titles gave a mean of 3.6. The two matched exactly on 2 keywords, came within one place on 10, and our count was lower on 15. This gap does not show poisoning: a one-off check from Cairo is not a Saudi desktop crawl, and each tool counts positions its own way. It shows that one keyword on one day can read as 1 in a tool and 3 or 4 in a browser, which is why the checklist below puts Search Console clicks first.

Why are Arab marketers suddenly asking if rank trackers can be trusted?

Nozzle, a rank-tracking vendor, reported on 12 October 2026 that Google returned altered results pages to its collection systems. By its count, 8.0% of 27.5 million pages gathered since 13 September were affected, against 0.17% before that date. Search Engine Roundtable relayed the claim. Google has not commented, so the figure is one vendor’s finding, not a confirmed fact.

The study covered 49 countries between 20 August and 9 October 2026. The country figures we could read list Australia, India, Canada, the US, Thailand and a few European markets, and name no Arab market. That means nobody has published a poisoning rate for Saudi Arabia, the UAE or Egypt, and anyone quoting one is guessing. You can read the summary of the Nozzle study and Barry Schwartz’s report on tracker problems for the vendor status notes.

Schwartz added on 16 October that volatility tools showed a spike with no confirmed update, and that third-party tracking tools were struggling. The only confirmed event in the window is the September 2026 spam update, which ran from 24 September to 8 October. That overlap is why a tracker chart can look alarming in exactly the weeks Arab agencies prepare White Friday reports.

What did we compare, and how?

We compared two daily series for Delta Medical Labs, a Saudi Arabic medical site: Semrush Position Tracking (Google, Saudi Arabia, Arabic, desktop, 200 keywords) and Search Console average position for the blog property, filtered to Saudi Arabia and desktop. The window is 19 September to 16 October 2026, 28 days, pulled on 19 October.

The two series do not measure the same thing, and we say so up front. Semrush follows a fixed list of 200 keywords across the whole domain. Search Console averages every query that produced an impression for the /blog/ folder, weighted by impressions. Google also states that position is an approximation that can differ from what you see in your own search. We did not run an incognito spot-check from Riyadh or Cairo, so this test cannot say whether any single day was poisoned.

Both series are VOCTOS pulls from 19 October 2026, shown with their scope. Treat them as one site’s sample, not a ranking guarantee or a regional benchmark.

How far did the two daily numbers swing?

Semrush’s average position for the 200 tracked keywords ranged from 4.33 to 7.995 over the 28 days, a spread of 3.67 places, with a mean day-to-day move of 0.57. Search Console’s Saudi desktop series ranged from 6.0 to 11.9, a spread of 5.9, with a mean daily move of 1.11. Both are noisy.

The third row is the quietest. Once the Saudi desktop filter is removed, Search Console’s blended average sat between 6.1 and 6.4 for 26 straight days. Slicing by country and device is what makes the line jump, because each slice holds fewer impressions and a different mix of queries each day.

Do the two tools agree on which days were bad?

No. The correlation between the Semrush daily average and the Search Console Saudi desktop daily average over the 28 days is -0.01, effectively zero. They moved in the same direction on 15 of 27 day-to-day changes, which is 55%, barely better than a coin flip. Two honest tools can still tell different stories.

Take 7 and 8 October. Semrush’s average slid from 5.08 to 6.33 to 7.995, then recovered to 5.245 by 11 October. Search Console’s Saudi desktop series sat at 7.1, 7.2 and 8.9 on 6, 7 and 8 October. A reader who only opened Semrush would have written “we dropped three places after the spam update ended.” A reader who only opened Search Console would have written nothing at all.

We cannot tell you the Semrush dip was poisoned data. Nozzle describes poisoned pages as short spells, with about 80% lasting one page, which would fit a one-day dip. But ordinary query-mix noise fits it equally well. What the test shows is narrower and more useful: two correct tools can disagree by several places on the same day.

Why does Search Console move so much on Saudi desktop?

Search Console averages whichever queries happened to get impressions that day, so a thin slice changes its mix daily. Our Saudi desktop slice drew roughly 10,000 to 20,000 impressions a day. The four Fridays in the window were also the four lowest-click days (77 to 96 clicks), because the Saudi weekend is Friday and Saturday.

Those four Fridays averaged position 9.0 against 8.2 for the other 24 days. Four data points is too few to call it a pattern, so treat it as a reason to annotate Fridays, not a finding. The point for Arab teams is practical: a weekend, Ramadan evening or a regional holiday changes who searches and what they type, and a daily average shifts with them.

Arabic adds a second mix problem. The same intent is typed in Modern Standard Arabic, Gulf dialect and Latin letters, and a tracked keyword list covers only the spellings someone chose. Search Console counts all of them. That is another reason the two series will not line up.

Which number should an Arab site report to its client?

Report Search Console clicks and impressions for the whole property as the primary number, then use a tracker for direction on a fixed keyword list, and never quote a single day. Google’s own help page defines position as the topmost position of your link on that results page, averaged. A tracker is a sample, and the sample can be wrong.

If you are comparing AI visibility scores too, see our audit of AI visibility tools and our guide to the Search Console AI report for Arabic sites.

Want a second pair of eyes on a report before it goes to your board? Send us the tracker export and the Search Console date range, and ask for a plain check of which numbers support the story.

What is a five-step check for Arabic keyword reports?

Before you act on a tracker drop, run five checks in this order: compare Search Console for the same queries and dates, look at 14 days instead of one, spot-check five keywords by hand from the target country, annotate known events, and ask the vendor how it detects swapped pages. Stop at the first check that disagrees with the tracker.

1. Open Search Console for the same property, country and device, and read clicks and impressions, not only position. If clicks and impressions held steady, the tracker drop is likely noise. Nozzle recommends the same step.
2. Look at a 14-day line. Nozzle says most poisoned spells resolve within three days, so a drop that vanishes in a week was likely bad data.
3. Search five of the affected Arabic keywords yourself in a private window from the target country, with the Arabic and the Gulf-dialect spelling.
4. Annotate the confirmed update dates, Fridays and holidays on the chart before you send it.
5. Ask your tracker vendor in writing how it detects replaced results pages and what it does with them.

We added a measurement note to our own spam-update article on 18 October, because the Semrush deltas there carry the same limits. You can read that piece on what moved in Arabic search after the September 2026 spam update.

Our read

Most Arab agency reports lean on one tracker line because it fits in a slide and the client can follow it. We think that habit is the real risk, more than Google’s blocking. A report that says “average position fell 2.5 places on 8 October” without Search Console clicks beside it should be sent back, whoever made it.

We also think the poisoning claim deserves a calm reading. It is one vendor’s study, its method is published but not independently checked, and it names no Arab market. We will not tell clients their data is poisoned. We will tell them which numbers to read in which order. Our own sample, one site and 28 days, shows ordinary disagreement and cannot prove anything about Google’s behaviour.

FAQ

My rank tracker shows a big drop but my traffic is flat. Which one is lying?

Probably neither. Trackers sample a fixed list of keywords and Search Console counts real impressions, so they can disagree. In our test the two daily series had a correlation of -0.01. When clicks and impressions hold steady for the same queries and dates, treat the tracker drop as unconfirmed.

Is Google poisoning rank-tracker data in Saudi Arabia and the UAE?

Nobody has published an Arab-market figure. Nozzle reports 8.0% of pages poisoned across 49 countries, and the country numbers we read name no Arab market. Google has not confirmed the practice. Until a regional number exists, check each drop against Search Console instead of assuming either answer.

Should I stop using Semrush or another tracker for Arabic keywords?

No, but change what you ask it. A tracker is useful for direction on a fixed keyword list over weeks. It is a poor judge of a single day. Pair every tracker chart with Search Console clicks for the same dates, and say which tool each number came from.

Why does Search Console average position jump on Saudi desktop?

A country and device filter leaves fewer impressions, and the query mix changes daily. Our Saudi desktop slice swung from 6.0 to 11.9 in 28 days, while the unfiltered blog line stayed between 6.1 and 6.4 over 26 days. Smaller slices make noisier averages, so widen the window before concluding anything.

What should I tell my client when the tracker chart looks bad?

Tell them what you checked and in what order. Say the tracker moved, say Search Console clicks did or did not, name the confirmed update dates, and give a date when you will re-check. A short note with that sequence beats a defensive explanation of the tool.

What is the better question to ask about your rankings?

The useful question is not whether your tracker is poisoned, but whether any decision would change if the number were wrong. If a budget, a client call or a content rewrite depends on one day of one tool, add Search Console and a 14-day window first.

If your Arabic keyword reports already pair tracker direction with Search Console clicks and annotations, keep going and re-check after the November data lands. If they do not, fix the report before you fix the strategy, and book a review of how we would structure your Arabic reporting. For a working example of Saudi Arabic search results measured end to end, see the Delta Medical Labs case study.