How to diagnose a drop in Google Ads conversions
Conversions are clicks multiplied by conversion rate, so a drop is either fewer clicks or a worse rate — and the first thing to establish is which, because they share almost no causes. Before either, rule out tracking: conversions falling to zero or near-zero while clicks continue is a measurement break, not a performance one, and it is the single most common cause of a dramatic-looking drop.
To diagnose a conversion drop, check in this order:
- Tracking — is the conversion action still firing at its usual volume
- Clicks versus conversion rate — which side actually moved
- Impression share — did you simply lose visibility
- Query mix — has match type drift changed who is clicking
- CPC and budget — did cost push you out of the auction
- Landing page and offer — did the destination change
- Recent account changes, including anything that reset learning
Match the pattern to the cause
| Clicks | Conversion rate | Likely cause |
|---|---|---|
| Flat | Collapsed to near zero | Tracking break — check the conversion action first |
| Down | Flat | Visibility or budget — check impression share and CPC |
| Flat | Down moderately | Traffic quality or landing page |
| Up | Down | Match type drift bringing in worse queries |
| Down | Down | Often a bidding change, or learning reset after an edit |
| Flat | Flat, conversions still down | Attribution or conversion-action configuration changed |
The diagnostic path
Step 1
Rule out a tracking break first
Compare the conversion action's daily volume against its own recent baseline. A genuine performance decline is gradual and partial. Conversions dropping to zero, or close to it, while clicks and impressions continue normally is the signature of a broken tag, a changed conversion action, or a site deployment that removed the trigger. Diagnosing anything else first wastes the day.
Also check whether the conversion action's counting or attribution settings changed. Switching counting from every to one, or changing the attribution model, lowers reported conversions immediately without anything changing in performance.
Step 2
Split clicks from conversion rate
Conversions equal clicks times conversion rate. Pull both for the affected period and the one before it, and establish which moved. Fewer clicks at a stable rate is a traffic problem; stable clicks at a worse rate is a quality or destination problem. They lead in opposite directions and conflating them is why conversion-drop investigations stall.
Step 3
If clicks fell, find out why you lost the traffic
Check impression share and its two loss reasons. Losing to budget means you ran out of money; losing to rank means you stopped winning auctions. Check CPC at the same time — if cost per click rose on a fixed budget, you bought fewer clicks for the same spend without anything else changing.
- Budget-lost impression share up — you capped out. Check whether spend actually hit the daily limit.
- Rank-lost impression share up — competition or quality. Check CPC and CTR together.
- Neither moved but clicks fell — check whether search volume for your terms fell, which is seasonal rather than fixable.
Step 4
If the rate fell, find out who is now clicking
A falling conversion rate at stable click volume usually means the mix of queries changed. Pull the search terms for both periods and compare the tiers: new losing or invalid queries appearing is match type drift, and it brings in clicks that were never going to convert.
Rising CPC alongside new junk terms is the specific signature of match type leakage. The fix is negatives and tighter match types rather than bid changes — raising bids here buys more of the wrong traffic.
Step 5
Check the destination
If query mix is stable and the rate still fell, the problem is probably past the click. Landing page changes, site speed regressions, checkout errors, stock outages and offer changes all show up as a conversion rate decline with no visible cause inside the ad account.
Step 6
Check what changed in the account
Pull the change history across the window. Bid strategy changes, budget changes, structural edits and new campaigns all reset learning, and a campaign relearning delivers worse for a period. The question to answer is whether the drop started before or after the change — which is why having the inflection date from step two matters.
Do not stack changes while diagnosing. If three things were edited the same week, you cannot attribute the drop to any of them, and the same applies to the fixes you are about to make.
Step 7
Confirm there is enough volume to be reading anything
Before concluding, check the numbers are large enough to carry meaning. Below roughly 30 conversions in 30 days on Search, week-to-week movement is mostly variance, and the honest answer may be that nothing diagnosable happened. The same applies at segment level, where we want at least 50 clicks and 5 conversions before drawing a conclusion.
A worked example
Illustrative example. The numbers below are made up to show the method. They are not a client account and not a benchmark.
Conversions fell 34% week over week on a Search account. The decomposition takes about ten minutes.
| Metric | Previous week | This week | Change |
|---|---|---|---|
| Impressions | 412,000 | 398,000 | −3% |
| Clicks | 9,850 | 10,410 | +6% |
| Average CPC | $2.21 | $2.08 | −6% |
| Conversion rate | 3.1% | 1.9% | −39% |
| Conversions | 305 | 198 | −35% |
Clicks went up, so this is not a traffic problem. That eliminates impression share, budget and most bidding explanations immediately.
CPC fell while clicks rose — cheaper clicks arriving in greater volume. That combination is suspicious rather than good: it usually means the account started matching broader, cheaper, less qualified queries.
The conversion rate fell by almost 40%, which is far too sharp for a landing page drift and exactly what query dilution looks like.
The confirming check is the search term comparison. If new losing and invalid terms appeared this week and took meaningful spend, this is match type drift and the fix is negatives — not bids, not budget, and certainly not the landing page.
What not to conclude too quickly
Conversion drops attract fast wrong answers:
- Do not assume performance before checking tracking. Near-zero conversions with normal clicks is a measurement break almost every time.
- Cheaper clicks are not automatically good. Falling CPC with rising clicks and a falling rate is usually dilution.
- A conversion count drop is not a conversion rate drop. Establish which one you have.
- Do not blame the landing page before checking query mix. The traffic may have changed rather than the destination.
- Attribution and counting changes lower reported conversions with no performance change at all.
- Below 30 conversions in 30 days, week-to-week movement is largely noise.
- Check whether learning was reset by a recent edit before diagnosing anything deeper.
How to automate this diagnosis
The decomposition itself is quick once the data is assembled. The slow part is the assembly: two periods, clicks and rate separately, impression share by loss reason, a search term comparison across both windows, conversion action health, and the change history.
GoMarble runs that assembly against the live account and returns the decomposition with the cause located, rather than a chart showing that conversions fell.
The shape of it
- Connect the Google Ads account
- Scheduled monitoring watches conversion volume against its own baseline
- A drop triggers the split between click volume and conversion rate
- Query mix, impression share and change history are checked automatically
- The likely cause and the recommended fix come attached
Catching a tracking break the day it happens rather than at the end of the month is most of the value here.
Conversion drop checklist
- ☐ Check the conversion action is firing at its usual daily volume
- ☐ Check counting and attribution settings did not change
- ☐ Split the drop into clicks versus conversion rate
- ☐ If clicks fell, check impression share by loss reason
- ☐ Check whether CPC rose on a fixed budget
- ☐ If the rate fell, compare search term tiers across both periods
- ☐ Look for new losing and invalid queries appearing
- ☐ Check landing page, site speed, checkout and stock
- ☐ Pull the change history across the window
- ☐ Establish whether the drop pre-dates or follows the change
- ☐ Confirm the volume is large enough to diagnose at all
Questions people actually ask
Why did my Google Ads conversions suddenly drop to zero?
My clicks went up but conversions went down. What happened?
How do I tell a tracking problem from a real performance drop?
Can a bid strategy change cause a conversion drop?
How long should I wait before reacting to a conversion drop?
Where this method comes from
This page composes the decomposition from the Google Ads analysis methodology GoMarble's agents run — the query tier classification, the impression-share loss split, the match-type leakage signature from the CPC diagnosis, the smart bidding volume thresholds and the segment significance floors. There is no single named conversion-drop workflow in that methodology, so the sequence here is assembled from those primitives rather than lifted whole. Thresholds are ours rather than published Google guidance.
Related
Find the cause, not just the chart.
Connect Google Ads and GoMarble splits the drop into clicks and rate, then locates the cause.