Improvements
26 engines that find the work, so you can just do it
Cliqda syncs your accounts and runs every engine against them continuously. What comes back is not a dashboard to interpret — it is a ranked queue of specific changes, each with the money at stake and the evidence attached.
- Engines running against every account
26
Engines running against every account
- Distinct improvement types produced
31
Distinct improvement types produced
- Platforms with live one-click push
2
Platforms with live one-click push
- Before-and-after change measurement
Impact
Before-and-after change measurement
Ranked by money
Start with the opportunities that matter most
Every improvement carries the spend at stake, not a colour-coded severity guess. A $2,000-a-month leak outranks a tidy-up, because it should.
- Estimated saving or opportunity attached to each item
- The underlying rows are on the card — you are not asked to trust a number you cannot check
- Findings whose sample is too small to be significant are suppressed, not padded into the list
Block “free” across 18 search terms
$2,148 spent in 30 days across 18 queries containing this pattern. One conversion. Adding a phrase negative to the shared list stops it everywhere.
Est. saving
$2,148/mo
| Search term | Cost | Conv. |
|---|---|---|
| free camping gear | $612 | 0 |
| free hiking boots | $443 | 0 |
| free tent giveaway | $388 | 1 |
| + 15 more terms | ||
Product illustration · Sample data, not customer results
Then it grades itself
Applied is not the same as worked
Change Impact takes every change you applied and measures the before-and-after on the metric it was meant to move. When something made performance worse, Cliqda says so and offers the rollback.
- Measured outcomes on the target metric
- Platform change events ingested, so out-of-Cliqda edits sit in the same timeline
- Guardrails flag regressions instead of leaving them in place
Every applied change, graded against the metric it was meant to move.
Paused 34 non-converting keywords
CPA before → after
$61.20 $48.90
Device bid −18% on mobile
CPA before → after
$54.80 $50.10
Broad match test, 4 ad groups
CPA before → after
$49.30 $57.40
The broad match test raised CPA. Cliqda flagged it for rollback rather than quietly leaving it in place.
Product illustration · Sample data, not customer results
Safety
Nothing reaches a live account by accident
The approval gate is the default, not a setting you have to find.
You decide, every time
Improvements are proposals. Nothing is written to the ad platform until a person pushes it — unless you have deliberately enabled an automation and set it to apply directly.
Automations can propose instead of apply
A rule can be set to drop its changes into an approval queue rather than execute them, so scheduled work still gets a human read.
Everything is logged
Who applied what, when, and what happened to the numbers afterwards. Every applied change flows into Change Impact automatically.
The full list
All 26 engines
Each looks for one well-defined problem, and stays quiet when it does not find it.
Keywords and queries
Search terms
Promotes converting queries to keywords in the right ad group.
Negative keywords
Blocks queries that spend without converting.
N-gram patterns
Turns wasteful query patterns into negatives.
Industry negatives
Applies vertical-specific negative sets to new accounts.
Duplicate keywords
Finds keywords competing with themselves across ad groups.
Match type
Recommends match-type changes where the data supports them.
Cannibalisation
Surfaces keywords bidding against each other for the same query.
Quality score
Flags keywords whose score slipped, ranked by spend behind them.
Ads and creative
Ad comparison
Finds ad groups with a significant winner and a loser to pause.
Extensions
Detects missing or underperforming extensions and assets.
Bid adjustments
Below first-page bid
Identifies keywords bidding under the first-page threshold.
Keyword bids
Adjusts keyword bids against CPA and ROAS targets.
Device bids
Sets device modifiers where performance genuinely differs.
Location bids
Adjusts geographic modifiers on statistically meaningful gaps.
Demographic bids
Age, gender and parental status modifiers from observed performance.
Schedule bids
Ad schedule modifiers built from hour and day-of-week performance.
Budget and pacing
Budget
Recommends budget changes from pacing and opportunity.
Account health
Disapproved ads
Escalates disapprovals with the serving impact attached.
Broken sitelinks
Catches sitelinks pointing at dead or redirecting URLs.
Landing pages
Flags landing pages failing technical or performance checks.
LP message match
Compares what the ad promises against what the page says.
Conversion health
Distinguishes tracking outages from real performance drops.
GA4 engagement
Uses GA4 engagement signal to find traffic that arrives and bounces.
Search partners
Evaluates whether search partner traffic is earning its spend.
Performance Max
PMax asset health
Rates Performance Max assets and proposes replacements.
PMax placements
Surfaces placements worth excluding from PMax campaigns.
Find out what your accounts have been hiding
Tell us about your accounts. We’ll help you explore the findings, the workflow and the results with a 14-day trial.
14-day trial · No credit card · Approval required by default