Why did my Google Ads CPC spike? Find the change in cost, clicks and traffic mix

CPC rises when the cost per reported click rises for the selected population

Your CPC spike means spending increased relative to clicks in the selected period. It can reflect higher within-campaign prices, a shift toward more expensive traffic or both. In a connected Slack channel, ask: “Compare cost, clicks and CPC for the last seven completed days with the previous seven, then break the change down by campaign.”

That live comparison is the starting point, not the final diagnosis. The account average cannot identify the responsible campaign or establish what changed in its auctions. A useful investigation confirms the inputs, locates the contribution and inspects relevant settings and traffic context before recommending a bid, budget or targeting change.

Google defines average CPC as cost divided by clicks and distinguishes it from the maximum bid in its average-CPC explanation. Use the measured average rather than treating it as the bid the owner set. A changed bid and a changed observed CPC are related questions, but they are not the same measurement.

Start with the actual amount and scale of the change. A movement from 1.00 to 1.20 is a 20% increase, while a movement from 0.10 to 0.12 has the same percentage with a different monetary consequence. Spending, click volume and campaign contribution help the team determine whether the spike deserves immediate investigation or a proportionate routine review.

The answer should end with a concrete next check supported by the data. “The increase is concentrated in the newly expanded campaign; inspect its changed traffic selection” is more useful than a list of every possible reason CPC can rise. A broad educational explanation can support that finding, but it should not replace the account evidence the user requested.

Calculate the total from spending and clicks, not average row averages

Account CPC is a weighted outcome. Combine spending and clicks across the intended population, then divide. Do not average the displayed CPC values of campaigns, days or devices without their contribution. A small expensive row should not carry the same numerical weight as a large inexpensive row merely because both occupy one line in a table.

Consider two fictional campaigns. Brand spends 100 for 200 clicks, giving a CPC of 0.50. Services spends 900 for 300 clicks, giving a CPC of 3.00. The account spends 1,000 for 500 clicks, so its CPC is 2.00. The unweighted average of 0.50 and 3.00 is 1.75 and does not describe what the account actually paid per click.

The current ppctor reporting semantics retain cost and click inputs for CPC aggregation. Money reported in micros is scaled into the account currency for readable output. This keeps weighted totals and presentation separate: a correct internal amount should not appear as millions of currency units because a raw provider field was displayed without its unit conversion.

Check the denominator when click volume is low or zero. A period with spending and no reported clicks does not have an ordinary finite CPC from that calculation. Do not replace the undefined ratio with zero and declare it cheap traffic. Show the relevant inputs and explain that the chosen denominator does not support the requested average.

A percentage change also requires a valid baseline. If the earlier period has no meaningful CPC, a conventional relative increase is unavailable. The assistant can still report the later spending, clicks and observed CPC, but it should not manufacture an infinite percentage or a dramatic rounded figure to make the table appear complete.

Reconcile the comparison at the same scope. A campaign table filtered to enabled campaigns can differ from a whole-account history that includes activity from campaigns now paused. Use the intended status population consistently or describe the difference. The reported spike should not be a side effect of comparing two different sets of campaigns.

Comparable periods prevent a timing artifact from becoming a diagnosis

Use completed dates in the relevant account time zone. A partial current day can have a different click and spending pattern from a complete earlier day. If the request says the last seven days, establish whether it means completed days or explicitly includes today. The application should not silently reinterpret that phrase as the previous calendar week.

Keep comparison windows equal in length unless the owner asks for another basis. A seven-day period and a ten-day period have different volumes and potentially different weekday composition. CPC itself is a ratio, but the population and business context still matter. Equal completed windows make the first comparison easier to interpret.

Inspect daily detail when the aggregate conceals a sharp event. One unusual day could account for much of the change, or CPC could have risen gradually across the whole period. Those patterns call for different next checks. A daily series should come from actual date-segmented observations rather than dividing the aggregate evenly across the days.

Compare the same weekday where a business has strong weekday patterns. A services campaign may buy different traffic on weekends and working days. This is a practical comparison choice, not proof that the weekday caused the change. State the selected basis so a later follow-up uses the same population rather than drifting into another baseline.

Seasonal promotions and opening hours can also change the traffic available to the campaign. Record relevant business context from the owner alongside measured data. The assistant can use that fact as a hypothesis to inspect, but it should not present an assumed promotion or closure as something the provider API verified.

A clear report states the two actual date ranges, currency and selected account or campaign. If multiple sources are used, retain their time-zone and metric definitions. Cross-source totals can differ for reasons unrelated to CPC arithmetic. The primary question here is what Google Ads spending and clicks show for the intended population.

Find the campaigns that account for the movement

Break the account comparison into campaign spending, clicks and CPC for both periods. Look at contribution, not just the largest percentage increase. A tiny campaign can double its CPC while barely changing the account total. A large campaign with a moderate movement can explain most of the business consequence.

Retain campaigns with activity in either period. A newly active campaign has no ordinary earlier baseline, but its later spending and clicks can change the account mix substantially. Omitting it because the comparison lacks an earlier row hides exactly the evidence the team needs to understand the account-wide spike.

A campaign that stopped receiving clicks can also change the average. If an inexpensive brand campaign is paused, the remaining account traffic can become more expensive without any other campaign's own CPC rising. The explanation should identify that contribution change rather than claiming that every auction became more competitive.

Use names and actual scope in the summary. “The account average rose mainly because Product Discovery received a larger share of clicks” tells the owner where to investigate. A vague statement that “campaigns changed” does not. If campaign names are duplicated, resolve the distinction internally and present enough readable context to avoid ambiguity.

After locating the major contributor, inspect that campaign's relevant breakdown. A whole-account device table may be less informative than a device comparison inside the campaign that moved. Narrowing the read can improve both clarity and efficiency, because the assistant requests evidence for the established question instead of collecting every possible segment.

Maintain the original total as a reconciliation check. The campaign breakdown should explain the selected account comparison within its provider coverage. If the requested resource or filter changes the population, disclose that boundary rather than forcing totals to match through local allocation. A missing contribution is a question to resolve, not a number to invent.

A traffic-mix shift can raise average CPC without a within-group price increase

Suppose two fictional campaigns keep the same CPC in both periods: Brand stays at 0.50 and Specialist Services stays at 3.00. In the first period most clicks come from Brand; in the second, more come from Specialist Services. The account average rises because the composition changes. Nothing in that example establishes that either campaign paid more for comparable clicks.

This distinction changes the response. If the shift is intentional and the specialist campaign produces valuable results, restoring the old cheap mix may be undesirable. If the shift came from an unintended pause or targeting expansion, a configuration review could be appropriate. The average alone cannot select between those cases.

Within-campaign changes and mix changes can occur together. A useful analysis compares each campaign's own spending-to-click ratio and its share of account clicks. The summary can then identify a measured increase inside the major campaign plus a changed contribution. It need not claim one simple cause when the data shows two distinct movements.

The same principle applies at ad-group, keyword and device levels. A campaign may shift toward more expensive but more valuable searches while each subgroup's own CPC remains stable. Do not compare the aggregate as though it were a controlled experiment with an unchanged traffic composition.

A mix review should remain tied to the business goal. Cheap clicks are not automatically useful clicks. A campaign could buy fewer low-intent interactions and more commercially relevant ones, increasing average CPC while improving cost per meaningful outcome. Conversely, a shift to expensive weak traffic can increase waste. The outcome evidence determines which interpretation fits.

If the owner asks for a quantified decomposition, use a clearly defined method and actual inputs rather than a rough narrative percentage. The ordinary contribution table is often sufficient for the initial decision. The assistant should not label a locally estimated causal share as a provider metric or pretend the API returns the reason for each account-average movement.

Inspect keyword and search context in the affected Search scope

For Search, compare the relevant ad groups and keywords within the campaign that changed. Identify which targets contributed spending and clicks in each period. A keyword's observed CPC is still a performance result, not a direct explanation of auction conditions. Its match type, intended topic and current status help establish the configuration context.

Search terms can reveal a change in the actual reported queries, but a keyword and a search term are different objects. Do not apply a keyword mutation to a term as though it were an existing target. If a negative is appropriate, resolve the intended scope and match behavior rather than simply copying a phrase into an account-wide exclusion.

A broader query mix can change average price and business intent. Inspect the reported terms that materially contribute to spending and outcomes, with the provider's actual reporting coverage. A visible example list is not necessarily the complete query universe. The assistant should avoid calculating a full-account percentage from only the few terms printed in a Slack preview.

Look for changes in product or service focus. A campaign might receive more searches for a premium service than for an entry-level one, producing a higher CPC that is commercially acceptable. Another campaign might drift into a less relevant topic. The relevant distinction is supported intent and outcome, not whether the query contains a word that looks expensive.

Do not assume a match-type change caused the spike merely because it appears in history. Establish its timing and the subsequent reported traffic change. The edit is evidence of configuration, while the performance comparison supplies observed results. Together they can support a focused investigation without claiming a fully isolated causal experiment.

The next action could be to retain the traffic, refine one unintended query pattern or inspect the destination for a relevant high-cost target. Each has a different scope and justification. The negative-keyword workflow explains how to make a deliberate exclusion without turning a CPC review into an indiscriminate keyword cleanup.

Device, location and network can explain where the change occurred

Choose a segment because it addresses the established question. If the campaign's click composition changed by device, compare device spending, clicks and CPC across the same periods. If targeting or market coverage changed, inspect the relevant location data. Avoid pulling every segment by default when the campaign contribution already points to a narrower investigation.

A device aggregate can rise because a more expensive device receives a larger share of clicks. It can also rise inside one device category. Preserve both the ratio and contribution. This is the same weighted reasoning used for campaigns, applied to a different dimension rather than a new formula.

Location reporting requires particular care. Physical presence and locations of interest are different contexts. The current account audit emphasizes physical location and can present United States state detail, while a campaign location reader has a different aggregation context. Do not compare these as identical populations or label a mixed geographic result as physical presence only.

A location targeting change is also different from a measured location shift. Adding a country, removing a positive target or changing Presence behavior are distinct settings operations. A campaign's increased CPC in one market does not automatically prescribe which of those controls should change. First inspect the desired reach and actual configured selection.

Network and campaign-family context can affect what the click measure describes. A broad account comparison can combine different advertising uses. Keep the requested resource's supported segments and metrics together, and avoid inventing a combined breakdown from separate marginal tables. Available device and country totals do not establish every device-by-country intersection.

When a supported segment returns no observations, retain that meaning. When the read fails, say it could not be established. Neither state should become a confident zero contribution to the spike. A report is useful because it preserves the coverage and actual inputs, not because every conceivable category has been given a numeric cell.

Auction evidence adds context without exposing competitors' private decisions

Google's Ad Rank guidance describes bids, quality, thresholds, competition, search context and expected asset impact as relevant factors. That confirms that observed price is not determined by one setting alone. It does not let the assistant infer a competitor's exact bid change from the account's CPC series.

When available for the selected context, auction-related metrics can show aspects of visibility or competitive overlap. Use the actual supported resource and period. Do not treat a competitor domain appearing in one report as proof that it caused the spike, or present an unsupported exact spending estimate as a Google Ads fact.

Impression-share and position-related measures answer different questions from CPC. They can help establish whether the campaign's visibility changed alongside price, but they are not click-cost measurements. Retain their definitions and denominators. A visibility improvement with a CPC increase may be informative without demonstrating whether the extra exposure was worth the cost.

Compare auction context inside the same relevant campaign selection rather than attaching a broad competitor list to every account-level increase. The question is what evidence helps explain the established movement. Pulling unrelated competitor data can make the response longer without making the causal conclusion stronger.

If the available evidence only confirms a price movement and changed visibility, state that conclusion precisely. The next check can be the campaign's recorded bidding and targeting changes. A useful investigation does not need an invented rival action to feel complete. It can identify the decision the owner should review based on measured account behavior.

Keep competitor intelligence and account auction evidence separate. Public creative can inform messaging research, while account auction metrics describe the selected advertising context. Neither should be used to claim access to another advertiser's private budgets, conversions or bidding strategy. Accurate boundaries preserve trust in the parts of the report that are actually verified.

Quality Score helps investigate Search experience but is not the auction formula

Google describes Quality Score as a keyword-level diagnostic, with expected clickthrough rate, ad relevance and landing-page experience components. It explicitly says the score is not an auction input or a KPI to aggregate. Use it to locate useful questions about the Search experience rather than calculating a promised CPC discount.

A weak component can justify inspecting the actual ad and destination. Does the message reflect the intended query? Does the page give the visitor the promised information and useful next step? Those practical checks are more meaningful than raising a numeric score for its own sake or averaging scores across unrelated campaigns.

The current ppctor audit can surface ad-group keyword quality context. That informational evidence should not be described as a complete real-time auction analysis. A historical diagnostic and a daily CPC series have different meanings. Use them together carefully without claiming that one reported score directly caused a precise amount of price movement.

A missing score is not a zero score. The provider's availability condition matters. Do not insert a zero into a campaign average or report a dramatic quality collapse when the field is absent. If the owner requests a quality review, present the returned values and components with their actual availability.

Creative and destination improvements can be valuable even when the score is not the deciding evidence. Their business purpose is to serve the relevant intent and produce useful outcomes. The CPC review should explain that purpose and define a follow-up measure rather than promising an immediate price reduction after one headline change.

If a live creative change is requested, prepare the complete supported desired state and confirm it. A recommendation to improve relevance is not a ready API operation. The assistant needs the target, text, URL and applicable asset behavior before execution. This keeps a sensible diagnostic suggestion from becoming an accidental broader edit.

Use recorded changes as evidence of timing and configuration

Inspect changes around the start of the movement, especially in the identified campaign. Bidding, targeting, status, budget and creative changes can provide relevant context. The record shows that something changed; it does not isolate its effect from all other conditions. Combine the actual timestamp and resource with the measured series before selecting the next action.

The current audit change-history read is bounded to recent events and a capped number of rows. A summary from that selection is not an unrestricted lifetime history. If the relevant change falls outside coverage or the cap is reached, retain the boundary. Do not describe “no returned change” as proof that no one altered the account.

Automation labels also require careful interpretation. The current summary groups specific recognized automated client types separately from other returned sources. A remaining API-related event is not necessarily a human clicking in the interface. Describe the recorded source accurately rather than blaming a particular person from an aggregate category label.

A newly enabled expensive campaign can explain a mix shift, while a changed target may help explain a within-campaign movement. A budget increase might alter available volume without proving a direct per-click price rule. The assistant should name the actual relevant edit and what the performance comparison shows afterward.

A restoration request needs a known desired state. “Undo whatever caused the spike” is not a safe or precise operation when several changes exist. Read the relevant recorded fields and current configuration, then establish which change the owner actually wants restored. Not every resource operation has a simple reversible equivalent.

The change-history guide explains the current audit's scope and interpretation. Use history to narrow a concrete investigation, not to generate a fictional automatic rollback. The owner's decision should be based on the selected setting and expected business purpose, with live confirmation for the actual supported edit.

A higher CPC is not automatically a worse business result

Pair CPC with the selected outcome. For a retailer, review conversion value, purchases where requested and value relative to spending. For a lead advertiser, review the relevant primary lead actions and cost per result. The purpose is to determine what the more expensive traffic produced, not simply whether the account could buy cheaper clicks.

If CPC rises while conversion rate improves enough to reduce cost per meaningful result, a bid reduction may undermine useful growth. If CPC rises and outcome efficiency deteriorates, the same price movement deserves a different response. Preserve the actual conversion selection and volume so the owner can see which interpretation fits.

Conversion value may include values assigned to primary nonpurchase actions, depending on the selected business measure. Do not call every value total verified ecommerce sales. If the owner wants purchase-only evidence, select the appropriate conversion group. A CPC investigation should not introduce a new measurement ambiguity while trying to explain a spending change.

Account goals and business capacity matter. A service company with limited availability may prefer fewer suitable enquiries, while a retailer may seek profitable volume. Record the owner's stated objective rather than inventing a margin or lead-quality threshold. The report can support the decision using available inputs without pretending to see facts outside the connected data.

Recent outcomes may be incomplete, so a immediate apparent decline should be reviewed with timing in mind. Spending and clicks can establish the CPC increase before later conversions fully settle. The assistant can distinguish the confirmed cost movement from the still-developing outcome comparison and choose a proportionate next review.

The final recommendation should connect the measured change to the business question. Retain useful expensive traffic, inspect an unintended expansion or prepare one supported correction when the evidence justifies it. A blanket instruction to lower bids treats a symptom as the goal and can discard the very traffic that contributes valuable results.

Three fictional spikes lead to three different decisions

The inexpensive campaign stopped contributing

A fictional account has stable CPC inside both Brand and Specialist Services, but Brand is paused halfway through the later period. Account CPC rises because the remaining clicks come mainly from Specialist Services. The recorded status change and contribution table support a mix explanation. The correct next question is whether pausing Brand was intentional, not why every auction supposedly became more expensive.

If the pause was accidental and the owner wants restoration, the assistant prepares that specific campaign-status proposal. If the pause was deliberate, the new account average is part of the chosen strategy. Both decisions can be reasonable. The data does not prescribe reactivation merely because it would make the average CPC look lower.

The expanded campaign bought different traffic

A fictional Search campaign receives more clicks from a newly emphasized service topic. Its CPC rises and its reported lead outcome weakens. The assistant identifies the contributing ad group and relevant reported query change, then inspects the associated configuration. It does not claim a specific competitor raised bids just because the campaign became more expensive.

The owner may choose a scoped refinement or further review depending on intent and volume. If a negative keyword is requested, the proposal names its match behavior and actual parent scope. The action follows the verified unwanted traffic pattern rather than excluding every high-CPC keyword in the account.

The higher price produced more useful outcomes

A fictional retailer sees CPC increase from 1.00 to 1.25 while its selected conversion-value efficiency improves. The contribution is concentrated in a product campaign with a changed commercial offer. The appropriate initial decision is to retain the traffic and review whether the outcome remains stable, rather than reverse the strategy solely to recover the earlier click price.

These invented cases illustrate measured reasoning, not predicted results. CPC is one input in a commercial decision. The same visible spike can reflect an unintended pause, an unhelpful expansion or a useful shift. A reliable investigation identifies which evidence exists and chooses the next action that follows from it.

Choose one proportionate next action from the evidence

End the initial review with a specific decision. It can be to retain the configuration, inspect one contributing segment, request a missing business fact or prepare a supported edit. The decision should name the target and the measured reason. A long list of unrelated recommendations makes the owner do the diagnostic work the assistant was asked to perform.

Do not apply a bid change without checking the current strategy and field eligibility. A manual bid, a target CPA and a target ROAS are different controls. The same casual phrase “lower the bid” cannot safely determine the operation for every campaign. Discover the actual supported setting and desired value before preparing the request.

A budget change is also distinct from a price change. The standard focused campaign-budget action currently supports nonshared daily budgets, not a universal shared-budget editor. If the owner requests a broader configuration, inspect the applicable path rather than advertising a capability from a generic provider schema alone.

Keep targeting actions specific. Removing a positive location is not the same as adding an exclusion, and adding an audience signal is not a hard restriction. A CPC recommendation should describe the actual reach behavior intended. The confirmation text must not conceal materially different settings under the broad word “optimization.”

If the evidence remains narrow, state the confirmed result and the exact additional check. For example, the comparison can establish a within-campaign increase without enough auction evidence to identify a causal competitor. That is a reason to inspect relevant configuration or business context, not to invent an answer. The next action should resolve the actual uncertainty.

Record the intended follow-up measure before execution. A change meant to remove unwanted traffic should be reviewed for that traffic and meaningful outcomes, not only a lower CPC. This keeps the intervention connected to its purpose and makes a later evaluation more reliable than a retrospective search for whichever metric improved.

The Slack investigation should reuse the established data plan

The conversational reporting framework uses field discovery, a measured-data request and optional visual rendering. It retains explicit metric, population, period and scope for the report. A follow-up should preserve those selections unless the user changes them. This allows an account CPC comparison to become a campaign or device review without losing the original question.

Ask for the inputs as well as the ratio: “Show spending, clicks and CPC for both periods by campaign.” The assistant can return a readable table, identify the major contribution and answer a follow-up from the same available results. It should not reread the whole account simply to repeat a total it already returned.

Use a narrower new read when the next question genuinely needs new grain or settings. A campaign-level aggregate cannot supply daily detail, and a performance table cannot prove the current bid configuration. Efficient reuse means retaining the appropriate evidence, not pretending every follow-up can be answered from the first response.

Creative requests remain conversational. “Suggest clearer headlines for the expensive service ad group” can be drafted from context without a new provider report. “Which headline currently exists?” requires current account data. The assistant should route by the requested evidence rather than treating every mention of advertising as a measured API task.

For a live change, prepare the concrete target and desired state with confirmation. If the conversation already identified the campaign, retain that context. If a deeper target is required, resolve it and show the actual parent. Clear hierarchy prevents a measured CPC diagnosis from being applied to another similarly named ad group.

The account-question guide explains precise conversational scope, and the reporting guide explains measured output. The aim is a focused sequence of evidence and decisions, with unnecessary reads removed and every new request tied to the question it resolves.

Choose a chart that communicates the actual change

A CPC line chart can show when the movement occurred. A campaign table can explain its contribution. A bar chart can compare category values across an appropriate selection. Choose the visual for the user's question rather than producing several charts by default. The underlying measured series must exist at the requested grain.

Two period series can be aligned for a comparison when the plan preserves the actual period meaning. Calendar alignment and relative day alignment answer different questions. The title and labels should make the chosen basis explicit. Do not present overlapping week positions as though they were the same calendar dates.

When the user requests several compatible series, ppctor supports a maximum of three in one readable native chart. Metrics with very different units need care on a shared scale. A CPC series beside spending can be mathematically renderable yet visually unhelpful. The assistant should honor the requested format while making the metric units understandable.

For longer time series, use an appropriate coarser grain when daily detail was not explicitly required. If the owner requires daily data beyond the native chart's point capacity, preserve the complete period in explained consecutive charts. Do not silently stop at the first portion or claim that the second chart represents a separate business question.

Show readable dates and sensible decimal precision. A monetary CPC rarely benefits from six decimal places in a client conversation. Retain enough precision internally for calculations while displaying clear currency values. A current-year date can omit its year when the period title establishes context; older periods need sufficient distinction.

The chart should reinforce the conclusion rather than replace it. A concise accompanying sentence can identify the measured spike and major contribution. If the requested chart cannot be built from the available data, explain the precise missing selection and use the supported next request instead of inventing arrays or declaring a false provider limitation.

Confirm the chosen edit and evaluate its actual purpose afterward

A supported conversational proposal names the campaign or deeper asset, its actual parent context and the requested operation. It lists logical actions clearly and waits for owner confirmation. The current flow states that no changes have been made yet and that the proposal expires in 15 minutes. The CPC analysis itself does not perform the live edit.

On confirmation, relevant settings are checked, the applicable mutation is executed and the resulting state is read back. The visible progress message and final outcome should make the sequence clear. Validation and verification serve different purposes: the first checks the proposed request, and the second establishes what the account now contains.

If the execution was sent but cannot be fully verified, inspect the current state before requesting another edit. Do not blindly replay the operation. A correct response can preserve the confirmed analytical context while treating the mutation outcome as a separate uncertainty that needs a targeted read.

A verified budget, status or targeting edit does not guarantee lower CPC. It establishes the saved configuration. Later reporting measures the commercial effect under the new conditions. Use the chosen follow-up window and selected outcome, and account for other relevant changes rather than claiming the edit caused every later movement.

Review contribution and efficiency together after the intervention. A lower CPC with fewer valuable outcomes may fail the business purpose. A stable or higher CPC with better useful volume may be acceptable. The owner should receive the result that answers the original decision, not a cherry-picked favorable metric.

The confirmation workflow keeps the approved action reviewable. Combining it with weighted reporting and explicit business outcomes turns a CPC spike into a disciplined investigation: establish the change, locate its contribution, choose the appropriate response and verify both configuration and later measured results at their proper stages.

Frequently asked questions

Can CPC rise without any campaign's own CPC rising?

Yes. The account can receive a larger share of clicks from an already more expensive campaign. Compare each campaign's ratio and contribution across the same periods. The account average is spending divided by clicks, so a changed traffic mix can move it even when subgroup prices remain stable.

Should I immediately lower bids?

First establish the contributing scope, current bidding strategy and business outcome. A higher click price can accompany useful traffic, and the applicable bid control differs by strategy. Prepare a concrete supported setting change only when the evidence and owner's intent identify the desired operation.

Does a lower Quality Score prove the cause?

No. It is a Search diagnostic, not an isolated real-time auction explanation. Inspect its available components and the actual ad experience where relevant, then combine that evidence with the measured CPC series and campaign context. Do not calculate a promised price reduction from the score alone.

Can the bot identify a competitor's exact bid increase?

The CPC series does not reveal another advertiser's private bidding decision. Supported auction context can help inspect visibility and overlap, but it should not be expanded into an invented exact competitor action. Use the actual available evidence to choose the next account investigation.

What is the minimum useful first report?

Request spending, clicks and CPC for two comparable completed periods, then a campaign contribution breakdown. That establishes the weighted movement and where to inspect next. Add narrower segments or current settings only when they answer the established question rather than reading every surface by default.