Did my traffic really drop?
The Quack Report has fewer clicks from Google. Check the reporting periods before recommending a rewrite.
September 1–30, 2026
444 Google clicks
October 1–23, 2026
182 Google clicks
Google Search Console, The Quack Report; Sep 1 to Oct 23, 2026.Check the reporting periods before explaining the change.
Check 1: what would you tell the site owner?
The Quack Report covers Oregon Ducks football. Would you assign a rewrite based on this report?
| Period | Clicks |
|---|---|
| September 1 to 30, 2026 | 444 |
| October 1 to 23, 2026 | 182 |
Source: The Quack Report’s archived Search Console export, through 2026-10-23.
See why each answer is right or wrong
Not quiteNothing. A drop that size is worth investigating straight away.It may deserve a look, but these totals cover different numbers of days, so they can't tell you yet how much traffic changed. Check the dates first.
CorrectThey cover different numbers of days.September is a full month; October has 23 days in this data. Matching the dates still gives 375 clicks for September 1 to 23 against 182 for October 1 to 23, so October is lower even on the same dates. Keep going.
Not quiteClicks are the wrong measure; look at impressions instead.Impressions are worth reading, but they have the same problem: the months are different lengths. Fix the comparison first, then choose the measure.
Compare the same dates
September 1 to 23375
October 1 to 23182
Watch the dates change
Notice which total changes when the periods match.
10-second silent walkthrough, drawn from the archived export. This is not a recording of the Search Console interface.
Read the video explanation
Compare the same dates.
Match September to October’s 23 days.
The cause still needs investigation.
What would you recommend?
Choose which periods to compare. Then explain your choice in the full exercise below.
Fictional practice data. K means thousand. Not results from my sites or clients.
Difficulty
Beginner: Choose a fair comparison
A fictional bicycle website recorded fewer Google search clicks in May so far. A click means someone selected a search result; it is not the same measure as an analytics session.
Before deciding whether to rewrite anything, choose two periods you can fairly compare.
Evidence for your decision
| Period | Days covered | Google clicks |
|---|---|---|
| April | 30 | 1K |
| May so far | 15 | 600 |
Analyze the reports
Download the reports and attach them to your AI tool. Paste the analysis prompt, then check its answer against the evidence.
Read or select the prompt
Then check its work. Check the reporting periods, one important source row and any calculation. Does its recommendation actually follow from that evidence?
Help checking the analysis
- Download the complete case and give it to an AI tool with the analysis prompt.
- Ask AI to organize the evidence. Check the dates, one important row and any calculation it uses.
- Choose or reject its recommendation. Explain the page action you would take and what you would check next.
If you’re stuck: Which option uses recorded results covering 15 days in both months?
Explain your decision
Use the evidence to explain your choice and next check. Your writing is not submitted or graded.
Name the start and end dates for both periods.
Explain what makes the periods more comparable.
Separate a change in clicks from its cause.
Compare with a worked answer
Partly right: Compare all of April with May so far, and label May as incomplete.
That is an honest progress update, but the windows still differ. It does not answer whether clicks fell over a comparable period.
Does not fit this case: Double the May count and compare that estimate with the April total.
That compares an estimate with a recorded result. It assumes the remaining days behave like the first 15.
Fits this case: Compare the first 15 days of April with the first 15 days of May.
Both totals would then describe observed clicks over 15 days. This improves the comparison without establishing why the counts differ.
How I would approach it
I would request April 1 to 15 and compare it with May 1 to 15. I would keep the full-April total in a separate progress note.
What we still don’t know
Equal-length windows can still differ in weekdays, holidays or events. Matching the days is a first check, not a full explanation.
Check your own explanation
- Name April 1 to 15 and May 1 to 15.
- Explain why 600 versus 1K does not establish a decline.
- Do not claim a cause or predict the rest of May.
Keep your work
Choose an answer and write your reasoning. Compare with the worked answer.
Notes are not saved automatically. Download your notes before leaving.
Intermediate: Separate the loss from the explanation
A fictional bicycle site has two complete 28-day reports. Commuting and repairs are separate page groups. The total combines them; it is not another group.
Find which page group lost clicks. Use impressions and average position to decide what to check next.
Check the source evidence
| Page group | Period | Impressions | Clicks | Average position |
|---|---|---|---|---|
| Commuting guides | Earlier 28 days | 20K | 2K | 5 |
| Commuting guides | Later 28 days | 10K | 1K | 5 |
| Repair guides | Earlier 28 days | 10K | 1K | 5 |
| Repair guides | Later 28 days | 10K | 1K | 5 |
| Total, not an additional group | Earlier 28 days | 30K | 3K | Not combined here |
| Total, not an additional group | Later 28 days | 20K | 2K | Not combined here |
Analyze the reports
Download the reports and attach them to your AI tool. Paste the analysis prompt, then check its answer against the evidence.
Read or select the prompt
Then check its work. Check the reporting periods, one important source row and any calculation. Does its recommendation actually follow from that evidence?
Help checking the analysis
- Download the complete case and give it to an AI tool with the analysis prompt.
- Ask AI to organize the evidence. Check the dates, one important row and any calculation it uses.
- Choose or reject its recommendation. Explain the page action you would take and what you would check next.
If you’re stuck: Compare both measures inside each group. Does the click loss happen without an impression change?
Explain your decision
Use the evidence to explain your choice and next check. Your writing is not submitted or graded.
Use the two groups and their counts to explain the total change.
Compare impressions and average position without claiming a cause.
Name the breakdown and what you would compare.
Compare with a worked answer
Does not fit this case: The commuting group lost clicks at a similar average position, so title changes should be the first fix.
Similar average position does not isolate a title problem. Impressions also halved, and an average can hide changes in the searches that appeared.
Fits this case: The observed loss is in commuting; inspect its page and query changes before selecting an edit.
Commuting accounts for the 1K-click decrease. Its recorded impressions and clicks both halved, while repairs stayed steady. The query mix is the next useful comparison.
Does not fit this case: The stable repair group suggests waiting for a site-wide decline before investigating any commuting pages.
Repair guides are steady, but that does not justify delaying a specific commuting investigation. A change need not affect the whole site to merit review.
How I would approach it
I would report that commuting explains the total click decrease, then compare its pages and queries across the two periods. Recorded exposure fell too; the table does not tell me whether demand, visibility or the mix of searches changed.
What we still don’t know
Average position and group totals can hide different query-level changes. These rows do not establish an algorithm effect or a title defect.
Another approach worth considering
If a separate page check finds a broken or inaccurate commuting page, repair that observed problem while keeping the traffic diagnosis separate.
What would change my mind
A query breakdown showing stable exposure for relevant searches but a changed result presentation would make a title or snippet investigation more relevant.
Check your own explanation
- Show that commuting accounts for the 1K-click decrease.
- Notice impressions and clicks both halved in that group.
- Use repair guides as context without treating them as proof of a cause.
- Ask for commuting page/query rows before recommending a title rewrite.
Keep your work
Choose an answer and write your reasoning. Compare with the worked answer.
Notes are not saved automatically. Download your notes before leaving.
Advanced practice: Prioritize a repair using separate data sources
A fictional cycling publisher has four developer hours today. The agreed goal is to protect the email-sign-up journey before a newsletter campaign tomorrow. Tracking and consent settings are unchanged. A separate editor has already corrected a bridge-closure warning.
Choose which problem to fix first with four developer hours. Use the search and analytics reports, without treating clicks and sessions as the same thing.
Check the source evidence
| Landing page | Period | GSC clicks | Google landing sessions | Email sign-ups |
|---|---|---|---|---|
| Commuting | April 1–28, 2026 | 2K | 1.8K | 100 |
| Commuting | May 1–28, 2026 | 1K | 900 | 100 |
| Repairs | April 1–28, 2026 | 1K | 800 | 50 |
| Repairs | May 1–28, 2026 | 1K | 800 | 50 |
| Bike reviews | April 1–28, 2026 | 500 | 1K | 100 |
| Bike reviews | May 1–28, 2026 | 500 | 1K | 20 |
Download the fictional source comparison (CSV)Six dated page/period rows. Calculate the sign-up rate within analytics; do not divide analytics events by GSC clicks.
| Check or proposal | Observed evidence | Estimated effort |
|---|---|---|
| Bike-review sign-up button on phone | The button points to an old sign-up route that returns 404 in three repeated checks. The desktop button reaches the current form. | Repair and phone/desktop verification: 2 hours |
| Commuting traffic investigation | The matched-period query breakdown has not been inspected. No live access problem found in the sampled pages. | Investigation: 4 hours |
| New bike-review comparison content | An editor suggested it; no current evidence shows it would repair the sign-up journey. | Prototype: 2 hours |
Analyze the reports
Download the reports and attach them to your AI tool. Paste the analysis prompt, then check its answer against the evidence.
Read or select the prompt
Then check its work. Check the source periods, a decisive row and any calculation. Challenge its strongest assumption and ask what would reverse the recommendation.
Help checking the analysis
- Download the complete case and give it to an AI tool with the analysis prompt.
- Ask AI to organize the evidence. Check the dates, one important row and any calculation it uses.
- Choose or reject its recommendation. Explain the page action you would take and what you would check next.
If you’re stuck: Use the analytics session denominator for the sign-up calculation. Then compare the result with the stated campaign goal, observed fault and effort limit.
Explain your decision
Use the evidence to explain your choice and next check. Your writing is not submitted or graded.
Show the analytics counts and calculation for both periods.
Connect the observed problem to the campaign goal and available hours.
Explain which other work can wait under this goal.
Include the page on phone and desktop and whether sign-ups are recorded.
Name a concrete check result, and avoid claiming the fault explains every lost sign-up.
Compare with a worked answer
Fits this case: Repair the bike-review phone button, verify the journey and use the remaining time to check measurement.
There is a reproducible sign-up obstacle on the page where analytics sessions with a sign-up fell from 100 to 20 while landing sessions stayed at 1K. The repair fits the stated goal and budget. The observations do not prove it caused the entire event decline.
Partly right: Use the developer’s four hours to investigate commuting queries because that page lost the most search clicks.
That is a real investigation opportunity, but it uses the full budget without addressing the observed sign-up failure. It would rank higher under a traffic-recovery brief or after the button was repaired.
Does not fit this case: Calculate sign-up conversion from sign-up sessions divided by GSC clicks, then prototype content for the weakest page.
Sign-up sessions must be divided by analytics landing sessions, not search clicks. Use counts from the same analytics report and period. The result is not a session conversion rate. The packet supplies a compatible denominator and a reproduced functional fault.
How I would approach it
I would calculate the bike-review share of analytics landing sessions with a sign-up as 100/1K = 10% earlier and 20/1K = 2% later. Then I would issue a two-hour fix brief for the phone button’s destination, with phone and desktop journey checks. I would keep the commuting investigation in the backlog and separately verify post-fix event collection.
What we still don’t know
An event-rate decrease and a reproduced 404 do not prove that the faulty route explains every lost sign-up. Device mix, visitor intent and other changes could matter.
Another approach worth considering
If the broken route had already been repaired and verified, the remaining time could go to the commuting query breakdown. That is a changed condition, not an equally supported first action under the current brief.
What would change my mind
A test showing the current phone button reaches the form correctly, or showing sign-up tracking changed, would alter the repair or measurement priority.
Check your own explanation
- Use 100/1K and 20/1K for the bike-review analytics rates: 10% and 2%.
- Do not use GSC clicks as the analytics denominator.
- Prioritize the reproduced phone-route fault under the campaign brief and four-hour limit.
- Specify the destination correction, phone and desktop regression checks, and event verification.
- State what the evidence cannot establish and when commuting would become the next priority.
Keep your work
Choose an answer and write your reasoning. Compare with the worked answer.
Notes are not saved automatically. Download your notes before leaving.
Something wrong, or a different approach? Include your evidence and lesson link. Leave out private information.
Optional: explore a worked real report
Use this as a reference when checking your analysis. You do not need to complete a second set of notes.
Checks 2 and 3: a total hides the days inside it
Look for unusually large days and check the dates before explaining the change. A daily pattern can suggest where to investigate; it cannot tell you the cause by itself.
Inspect or download the daily data
| Date | Clicks | Context |
|---|---|---|
| 2026-09-01 | 10 | No game listed |
| 2026-09-02 | 8 | No game listed |
| 2026-09-03 | 12 | No game listed |
| 2026-09-04 | 12 | No game listed |
| 2026-09-05 | 187 | Boise State |
| 2026-09-06 | 31 | No game listed |
| 2026-09-07 | 17 | No game listed |
| 2026-09-08 | 6 | No game listed |
| 2026-09-09 | 5 | No game listed |
| 2026-09-10 | 5 | No game listed |
| 2026-09-11 | 4 | No game listed |
| 2026-09-12 | 15 | Oklahoma State |
| 2026-09-13 | 11 | No game listed |
| 2026-09-14 | 5 | No game listed |
| 2026-09-15 | 5 | No game listed |
| 2026-09-16 | 2 | No game listed |
| 2026-09-17 | 4 | No game listed |
| 2026-09-18 | 10 | Portland State |
| 2026-09-19 | 4 | No game listed |
| 2026-09-20 | 3 | No game listed |
| 2026-09-21 | 8 | No game listed |
| 2026-09-22 | 4 | No game listed |
| 2026-09-23 | 7 | No game listed |
| 2026-09-24 | 5 | No game listed |
| 2026-09-25 | 2 | No game listed |
| 2026-09-26 | 28 | USC |
| 2026-09-27 | 15 | No game listed |
| 2026-09-28 | 11 | No game listed |
| 2026-09-29 | 5 | No game listed |
| 2026-09-30 | 3 | No game listed |
| 2026-10-01 | 8 | No game listed |
| 2026-10-02 | 9 | No game listed |
| 2026-10-03 | 8 | bye week |
| 2026-10-04 | 7 | No game listed |
| 2026-10-05 | 9 | No game listed |
| 2026-10-06 | 4 | No game listed |
| 2026-10-07 | 7 | No game listed |
| 2026-10-08 | 7 | No game listed |
| 2026-10-09 | 5 | No game listed |
| 2026-10-10 | 36 | UCLA |
| 2026-10-11 | 12 | No game listed |
| 2026-10-12 | 8 | No game listed |
| 2026-10-13 | 4 | No game listed |
| 2026-10-14 | 9 | No game listed |
| 2026-10-15 | 14 | No game listed |
| 2026-10-16 | 4 | No game listed |
| 2026-10-17 | 7 | Nebraska |
| 2026-10-18 | 2 | No game listed |
| 2026-10-19 | 3 | No game listed |
| 2026-10-20 | 2 | No game listed |
| 2026-10-21 | 4 | No game listed |
| 2026-10-22 | 7 | No game listed |
| 2026-10-23 | 6 | No game listed |
See why each answer is right or wrong
Partly rightClicks per day, September against October.That corrects for the number of days, but one big day still pulls the average up. Use it as a first description, then look at which dates and pages make up the change.
CorrectGame days against game days, and ordinary days against ordinary days.The table below groups the same kinds of day and leaves out the opener and the day after it, so the largest day does not dominate. Per day, October's Mondays to Thursdays are about the same, its other Fridays to Sundays are about the same, and its game days are higher. These averages describe only these selected days; small groups and exclusions do not establish a normal level or explain the change. That narrows where to look, not why it happened. Next, check which pages lost clicks.
Not quiteWait for the full October before comparing anything.You can compare the dates you already have now, as long as you say October isn't complete and that its newest days, October 22 and 23, may still rise. A full month will sharpen the comparison but won't explain a change by itself.
| Days included | September | October |
|---|---|---|
| Game days | 17.7 a day over 3 days | 21.5 a day over 2 days |
| Monday to Thursday | 6.8 a day over 18 days | 6.6 a day over 13 days |
| Other Fridays to Sundays | 7.3 a day over 7 days | 6.6 a day over 8 days |
See the worked decision and day groups
How the days are grouped
Game days are the days the team played. Other Fridays to Sundays are weekend days without a game; Monday to Thursday are ordinary weekdays. The opener on September 5 and the following day are kept apart so the largest part of the report does not dominate this comparison. That choice does not establish why either day had those clicks. September 18, a Friday game, counts as a game day; October 3, the bye week, counts as an ordinary weekend day.
Investigate before assigning a rewrite
October’s recorded clicks are lower, but the two totals cover different periods. September also contains a large day: 187 clicks on 2026-09-05, or 42% of its total. Taking that day out drops September from 444 clicks to 257, which shows how much one day moves the total (a different cut from the matched-dates comparison in check 1). But October has big days too, so rather than deleting days, the fairer test is to compare game days with game days and ordinary days with ordinary days.
Compare relevant dates and inspect the pages and queries behind the change. A schedule can help choose comparisons. It does not establish whether a matchup, search demand, a ranking change or something else caused the difference.
Check 4: now look at one page
Now look at the roster page. Impressions count appearances in Google’s results; they are not clicks. Write your first explanation before opening the worked answer.
The owner also asks about one page. On the October 10 game day, the roster page got more impressions than any other page on the site; it lists the team's players and its depth chart, the order of starters at each position. Searches containing “depth chart” gave it 1.3K impressions on October 10 and 94 on October 17. What would you look at before recommending a rewrite?
| Date | Impressions | Clicks | Average position |
|---|---|---|---|
| 2026-10-10 | 1330 | 9 | 9.2 |
| 2026-10-17 | 94 | 1 | 8.7 |
Inspect the query rows behind these totals
“Qb” means quarterback. Average position is the page’s result position averaged over its appearances for these exported searches; 1 is the top result. Search Console leaves out rare queries, so a missing row does not establish zero searches.
| Date and query | Impressions | Clicks |
|---|---|---|
| 2026-10-10 oregon qb depth chart | 395 | 0 |
| 2026-10-10 oregon football depth chart | 262 | 2 |
| 2026-10-10 oregon depth chart | 254 | 0 |
| 2026-10-10 oregon ducks depth chart | 105 | 1 |
| 2026-10-10 oregon ducks depth chart 2026 | 66 | 1 |
| 2026-10-10 oregon football depth chart 2026 | 59 | 2 |
| 2026-10-10 oregon depth chart 2026 | 53 | 0 |
| 2026-10-10 oregon qb depth chart 2026 | 49 | 1 |
| 2026-10-10 oregon ducks football depth chart 2026 | 25 | 2 |
| 2026-10-10 oregon ducks qb depth chart 2026 | 16 | 0 |
| 2026-10-10 oregon football qb depth chart | 12 | 0 |
| 2026-10-10 oregon football roster 2026 depth chart | 8 | 0 |
| 2026-10-10 ducks qb depth chart | 7 | 0 |
| 2026-10-10 orgeon depth chart | 4 | 0 |
| 2026-10-10 oregon ducks football roster 2026 depth chart | 3 | 0 |
| 2026-10-10 oregon 2026 depth chart | 2 | 0 |
| 2026-10-10 orgeon football depth chart | 2 | 0 |
| 2026-10-10 2026 oregon football depth chart | 1 | 0 |
| 2026-10-10 oregan football depth chart | 1 | 0 |
| 2026-10-10 oregon depth chart football | 1 | 0 |
| 2026-10-10 oregon ducks 2026 depth chart | 1 | 0 |
| 2026-10-10 oregon ducks 2026 football depth chart | 1 | 0 |
| 2026-10-10 oregon ducks football qb depth chart | 1 | 0 |
| 2026-10-10 oregon fb depth chart | 1 | 0 |
| 2026-10-10 oregon quarterback depth chart | 1 | 0 |
| 2026-10-17 oregon qb depth chart 2026 | 42 | 0 |
| 2026-10-17 oregon ducks qb depth chart 2026 | 30 | 1 |
| 2026-10-17 oregon ducks football depth chart 2026 | 5 | 0 |
| 2026-10-17 oregon football depth chart 2026 | 4 | 0 |
| 2026-10-17 oregon depth chart 2026 | 3 | 0 |
| 2026-10-17 oregon football depth chart | 2 | 0 |
| 2026-10-17 oregon ducks depth chart | 1 | 0 |
| 2026-10-17 oregon ducks depth chart 2025 | 1 | 0 |
| 2026-10-17 oregon ducks depth chart 2026 | 1 | 0 |
| 2026-10-17 oregon ducks football 2026 depth chart | 1 | 0 |
| 2026-10-17 oregon ducks football depth chart | 1 | 0 |
| 2026-10-17 oregon ducks football roster 2026 depth chart | 1 | 0 |
| 2026-10-17 oregon football qb depth chart | 1 | 0 |
| 2026-10-17 oregon football roster 2026 depth chart | 1 | 0 |
See why each answer is right or wrong
Not quiteThe roster page slipped in the results, so it needs rewriting.Fewer impressions don't show a ranking loss or a problem with the page. Compare the same searches and their positions, and read the page, before choosing a fix.
Not quiteThe roster page's average position stayed about the same, so its rankings held.Average position is the average spot the page held across the searches it appeared for. A steady average can hide a change in which searches the page appears for: a page can drop out of one search while its average across the others stays the same. Look at the individual searches before ruling out a ranking change.
Partly rightImpressions fell, and the cause isn't clear yet.That's an honest description. Make it useful by naming the next check: compare the same searches on both dates, then read the page.
CorrectCompare the searches on both dates and read the page before choosing a change.The search table shows what makes up each total. On October 10, one search, “oregon qb depth chart”, gave the page 395 of those impressions. On October 17 that search isn't in the export at all. That's the lead: what changed for that search, and does the page still answer it? A search missing from the export doesn't mean nobody searched it.
Text version of the worked explanation
These searches ask for a depth chart.
The same question appears in each search.
Read the page before choosing an edit.
Download the selected query rows (CSV)
The same October 10 and 17 rows shown above. These are selected exported queries; a missing row does not mean zero searches.
Your real-report answer review
- Right
- 0
- Partly right
- 0
- Wrong
- 0
- Unanswered
- 0
Use the explanations beside each question to review your choices, then write your recommendation.
Only your latest choice for each real-report question counts. Your written recommendation is not graded.
Review each answer
Compare with a worked recommendation
Recorded clicks are lower in the October window, including when the same calendar dates are compared. September contains a large spike, but the totals and daily chart do not establish why the change happened. I would compare the affected pages and the same queries across relevant dates, then inspect the roster page’s answer before deciding whether it needs rewriting.
This is a worked recommendation for this data. Compare its evidence and its next check with yours; if you'd check something else first, say what it would settle.
To get the same views for your own site in Search Console, open Performance, then Search results. Set the two date ranges, then use the Pages and Queries tabs, and add a Page filter to see one page’s searches. For more on reading these reports, start with Search Console as a visibility dashboard.
Ready to build a whole page? Explore the Moat Map course (Free during early access).