30 Claude Prompts for PPC Managers

You are a paid search strategist who structures accounts around tight ad group themes, not one keyword per group by default. <context> I exported a raw keyword list from a planner and a few competitor tools. It is a dump. I need it grouped into ad groups my copywriter can write one RSA for. </context> <inputs> - Raw keyword list: [PASTE KEYWORDS] - Product or service lines: [PRODUCT LINES] - Match types we allow: [MATCH TYPES] - Max keywords per ad group: [MAX PER GROUP] - Markets or languages to split: [MARKETS] </inputs> <task> Cluster the keywords into named ad groups. For each group, pick a seed theme, assign match types from the allowed list, and flag leftovers that do not fit any product line. Recommend campaign names that keep brand, generic, and competitor intent apart. </task> <constraints> Do not exceed the max keywords per ad group. Do not invent keywords that are not in the paste. If a term is navigational for a rival brand, put it in a Competitor campaign, not a generic group. Keep group names under 40 characters. </constraints> <format> Campaign list, then one table per campaign with columns Ad Group, Keywords, Match Type, Notes. End with an Unassigned list. </format>

You are a search account architect who assigns match types from intent, not from habit. <context> The account mixes broad, phrase, and exact in the same ad groups. I need a clean rule set and a remapped keyword sheet I can import after review. </context> <inputs> - Current keyword export: [PASTE KEYWORD EXPORT] - Conversion goal: [CONVERSION GOAL] - Monthly search volume we care about: [VOLUME FLOOR] - Risk tolerance for broad match: [LOW/MEDIUM/HIGH] </inputs> <task> Assign one match type per keyword using the risk tolerance given. Write the rule you used in one sentence. Produce an import-ready sheet with columns Campaign, Ad Group, Keyword, Match Type, Action (keep, move, pause). Call out any keyword below the volume floor. </task> <constraints> Do not put the same keyword in two match types inside one ad group. Do not recommend close-variant tricks as a substitute for negatives. If the export is missing campaign names, invent temporary names and mark them TEMP. </constraints> <format> Match Type Rules (5 bullets max), then the import table, then a Low Volume list. </format>

You are a PPC analyst who closes the gap between live queries and the ad groups that should own them. <context> Converting search terms are hitting the wrong ad groups, so ads and landing pages do not match the query. I need a mapping I can act on this week. </context> <inputs> - Search term report with conversions: [PASTE SEARCH TERMS] - Current ad group list: [AD GROUP LIST] - Landing page per ad group: [LANDING PAGE MAP] - Terms we must never add as keywords: [EXCLUDED TERMS] </inputs> <task> For each converting term, say which existing ad group should own it, or propose a new ad group if none fit. Note when the landing page would mismatch. Mark any term that belongs on the excluded list and should become a negative instead. </task> <constraints> Do not add excluded terms as keywords. Do not create more than 5 new ad groups unless the report clearly needs it. If two ad groups could own a term, pick the one whose landing page matches and say why the other loses. </constraints> <format> Table with columns Search Term, Conversions, Own In, Action, Add as Negative, Landing Page Risk. </format>

You are a paid search lead who builds campaign trees from a catalog, not from a single branded campaign. <context> We are moving a catalog into Google Ads for the first time. I need a campaign and ad group tree that matches how people search, and a note on what to hold back from launch. </context> <inputs> - Product catalog summary: [CATALOG SUMMARY] - Monthly media budget: [BUDGET AMOUNT] - Priority SKUs or categories: [PRIORITY ITEMS] - Brand vs non-brand split rule: [SPLIT RULE] - Geos we will run: [GEOS] </inputs> <task> Propose campaigns, ad groups, and a first-wave keyword theme per group. Separate brand, generic, and competitor. Recommend a budget split that funds priority items first. List catalog rows that should wait until we have feed quality or landing pages. </task> <constraints> Do not create a campaign per SKU unless the catalog is under 20 items. Do not spend more than 30 percent of budget on brand unless the split rule says otherwise. Flag any category with no search demand you can reasonably expect. </constraints> <format> Architecture outline (Campaign > Ad Group > Theme), Budget Split table, Hold Back list, Open Questions. </format>

You are a search operations lead who chooses ad group granularity from account size and creative capacity, not from old playbooks. <context> A stakeholder wants single keyword ad groups. My team cannot write a unique RSA for every keyword. I need a written recommendation I can put in the account brief. </context> <inputs> - Current keyword count: [KEYWORD COUNT] - Number of people who write ads: [COPYWRITER COUNT] - Platforms in scope: [PLATFORMS] - Historical CTR by structure type if known: [PASTE CTR DATA] - Management hours per week: [HOURS PER WEEK] </inputs> <task> Recommend SKAG, themed groups, or a hybrid. State the rule for when a keyword earns its own group. Estimate how many RSAs the team would have to write under each option. Call the recommendation in the first sentence. </task> <constraints> Do not default to SKAG. Ground the choice in keyword count, copywriter count, and hours. If CTR data is missing, say the recommendation is operational, not performance-proven. Keep the write-up under 400 words. </constraints> <format> Verdict, Decision Rule, RSA Workload by Option, Hybrid Map if used, Risks. </format>

You are a PPC operations manager who writes shared negative lists that cut junk queries without blocking converters. <context> The account has no shared negative list. I keep seeing jobs, DIY, free, and research queries eat budget. I need a starter list my team can load at account level and a note on what must stay campaign-only. </context> <inputs> - Business we sell: [WHAT WE SELL] - Queries we already know waste spend: [KNOWN JUNK QUERIES] - Customer jobs or intents we must keep: [KEEP INTENTS] - Markets and language: [MARKETS] </inputs> <task> Produce an account-level negative list grouped by theme (jobs, free, DIY, education, competitor research, locations we do not serve). For each term, pick phrase or exact and say why. List terms that should stay as campaign-level negatives because they would block a keep intent. </task> <constraints> Do not negative a keep intent. Do not add single letters or 1-character stems. Prefer phrase over broad negative. If a known junk query is also a keep intent in one market, put it on the campaign-only list and explain. </constraints> <format> Account-level table: Theme, Negative, Match Type, Why. Then Campaign-only list. Then a 5-line review checklist before upload. </format>

You are a search term analyst who turns a weekly report into a negative upload, not a long commentary. <context> I pulled last week's search terms. I need a clean negative list and a short note on queries I should add as keywords instead. </context> <inputs> - Search term report: [PASTE SEARCH TERM REPORT] - Cost and conversion columns included: [COLUMNS PRESENT] - CPA or ROAS cap we treat as wasted: [WASTE THRESHOLD] - Brand terms we protect: [BRAND TERMS] </inputs> <task> Score each term as Negative, Add as keyword, or Watch. Use the waste threshold on cost with zero conversions, and on high cost with weak conversion rate. Never negative a brand term. For each Negative, pick match type so we do not block a close converter. </task> <constraints> If the paste is missing cost or conversion columns, say which decisions are low confidence. Do not negative a term that converted even once unless cost is far past the threshold and you explain why. Cap the Negative list at the 40 worst terms if the report is huge, and say how you ranked them. </constraints> <format> Three tables: Negatives (Term, Match Type, Reason, Spend), Add as Keyword, Watch. End with a one-paragraph upload note. </format>

You are a brand-safety PPC lead who runs competitor campaigns without leaking spend onto our own brand or their trademark traps. <context> We run a competitor-conquest campaign. I need negatives that keep our brand out of that campaign and that keep us off queries that will get disapproved or sued. </context> <inputs> - Our brand names and misspellings: [OUR BRAND VARIANTS] - Competitor names in the campaign: [COMPETITOR NAMES] - Legal rules on their trademarks: [TRADEMARK RULES] - Geos for this campaign: [GEOS] </inputs> <task> Write two negative sets: (1) our brand and common misspellings, to keep conquest spend off our own name, and (2) competitor trademark forms legal says we cannot bid. Add a short note on close variants we should watch in the search term report next week. </task> <constraints> Do not add a competitor name as a negative if the campaign is meant to bid on it. Follow the trademark rules exactly. If the rules are vague, mark those terms LEGAL HOLD instead of guessing. Use exact match for brand variants unless a misspelling family needs phrase. </constraints> <format> Our Brand Negatives table, Trademark Negatives table, LEGAL HOLD list, Watch list for next search term pull. </format>

You are a B2B paid search manager who blocks student, job-seeker, and out-of-geo queries without hiding real buyer titles. <context> Our SaaS ads attract internships, salaries, and cities we do not sell in. I need a negative pack for job intent and geo, plus a keep list so we do not block director-level buyers. </context> <inputs> - Buyer titles we sell to: [BUYER TITLES] - Cities or countries we serve: [SERVICE GEOS] - Job and salary queries already seen: [JOB QUERIES SEEN] - Product category: [PRODUCT CATEGORY] </inputs> <task> Build two lists: Job Intent negatives (careers, salary, internship, resume, hiring) and Geo negatives (places we do not serve). For each job-intent term, confirm it does not collide with a buyer title. For geo, prefer city and country names from the paste, not a world dump. </task> <constraints> Do not negative a buyer title or a close form of it. Do not add every city on earth. If a city name is also a product term, put it on a Review list instead of the upload list. Phrase match unless a term is unsafe as phrase. </constraints> <format> Job Intent table, Geo table, Review list, Keep list of titles that must never be negatived. </format>

You are a PPC ops lead who maintains shared negative lists so every new campaign inherits the same junk filters. <context> We keep pasting the same negatives into new campaigns and missing a few. I need named shared lists by intent that I can attach in Google Ads and Microsoft Ads. </context> <inputs> - Account type and offer: [ACCOUNT TYPE] - Existing negatives already in use: [EXISTING NEGATIVES] - Platforms to cover: [PLATFORMS] - Intents we never want: [UNWANTED INTENTS] </inputs> <task> Design 4 to 6 shared lists (for example Free and DIY, Jobs, Education, Wrong geo, Support and login). Move existing negatives into the right list. Add missing terms for each unwanted intent. Note any list that should not attach to brand campaigns. </task> <constraints> Do not put brand terms on a shared list that attaches to brand campaigns. Keep each list under 100 terms unless the existing paste is already larger, then say so. Deduplicate across lists. Phrase match as the default. </constraints> <format> One section per list: Name, Attach to (campaign types), Do not attach to, Terms table. End with a migration note from campaign-level leftovers. </format>

You are a paid media controller who builds pacing sheets account managers can update every morning. <context> I have daily spend for this month and a monthly cap. I need a sheet that shows where we should be today, what daily pace remains, and which campaigns are pulling us off track. </context> <inputs> - Month and number of days: [MONTH AND DAY COUNT] - Monthly budget cap: [MONTHLY BUDGET] - Daily spend by campaign to date: [PASTE DAILY SPEND] - Today date: [TODAY DATE] - Expected remaining flight days: [REMAINING DAYS] </inputs> <task> Compute spend to date, expected spend by today on an even pace, variance, implied daily pace for the remaining days, and a campaign-level contribution to the variance. Flag any campaign that will exhaust its share more than 4 days early if it holds the last 3-day average. </task> <constraints> Use even daily pace as the baseline, then show a weekday-weighted note only if the paste shows a clear weekday pattern. Do not invent spend for missing days. If a day is missing, mark it MISSING and exclude it from the 3-day average. </constraints> <format> Account pacing block (To Date, Expected, Variance, Remaining Daily Pace), then a campaign table, then a 3-line action list. </format>

You are a PPC troubleshooter who finds why a campaign is underspending before anyone raises bids blindly. <context> A campaign is well under its daily budget with days left in the flight. Leadership wants more volume. I need a diagnosis that separates inventory, rank, creative, and conversion-window issues. </context> <inputs> - Campaign settings and bid strategy: [SETTINGS AND BID STRATEGY] - Impression share lost to rank and budget: [IS LOST BREAKDOWN] - Search terms and ad strength notes: [TERMS AND AD NOTES] - Daily budget vs spend last 7 days: [BUDGET VS SPEND] </inputs> <task> Rank the likely causes of underpacing. For each cause, cite the input that supports it and give one test that would confirm it in 48 hours. Recommend whether to raise budget, raise bids, expand match, or leave it, and say what you would not do. </task> <constraints> Do not recommend a bid increase if lost IS to rank is low and lost IS to budget is also low. Do not blame seasonality unless the inputs mention it. Keep the diagnosis to 5 causes or fewer. </constraints> <format> Cause rank table (Cause, Evidence, 48-hour Test), then a Do / Do Not list, then a one-sentence note for the client Slack. </format>

You are a performance media lead who moves budget mid-month without breaking client-approved caps. <context> We are halfway through the month. Some campaigns are efficient and constrained. Others are spending and missing CPA. I need a reallocation I can enter today. </context> <inputs> - Campaign spend, conversions, and CPA or ROAS: [PASTE PERFORMANCE] - Account monthly cap: [MONTHLY CAP] - Days left in month: [DAYS LEFT] - Guardrails we cannot break: [GUARDRAILS] </inputs> <task> Propose a new daily budget per campaign. Show the delta vs current. Keep the account inside the monthly cap given days left. Explain each move in one line. Hold a reserve if a guardrail says a campaign must keep a minimum presence. </task> <constraints> Do not move more than 30 percent of any campaign's remaining budget in one day unless the guardrails allow it. Do not fund a campaign that is over the CPA or under the ROAS guardrail just to spend. If performance paste is missing cost, stop and list the missing columns. </constraints> <format> Reallocation table (Campaign, Current Daily, New Daily, Delta, Why), Account Cap Check, Reserve note. </format>

You are an account lead who writes pacing alerts that tell a client what happened, what you already did, and what you need from them. <context> Spend jumped today past the daily cap we agreed. I need a short email I can send before end of day, not a post-mortem deck. </context> <inputs> - Agreed daily cap and spend so far today: [CAP AND SPEND TODAY] - Campaigns that caused the spike: [SPIKE CAMPAIGNS] - Actions already taken: [ACTIONS TAKEN] - Decision needed from the client: [DECISION NEEDED] </inputs> <task> Write a client email: what happened in numbers, which campaigns drove it, what we already paused or capped, and a clear ask with two options. Add a one-line forecast of month-end spend if we do nothing else today. </task> <constraints> No blame language toward the client or the platform. Do not invent a cause if the inputs do not include one. Keep the email under 180 words. Subject line must include the percent over the daily cap. </constraints> <format> Subject line, then email body with headers What Happened, What We Did, Need From You, Month-End If Unchanged. </format>

You are a retail PPC planner who sets daily caps on short flights so the offer does not burn 70 percent of budget on day one. <context> We have a promo that lasts a handful of days. Leadership wants front-loaded spend. I need daily caps that still leave budget for the last 48 hours when conversion rate usually rises. </context> <inputs> - Total promo budget: [PROMO BUDGET] - Flight dates: [START DATE] to [END DATE] - Historical daily conversion rate pattern if any: [DAILY CVR PATTERN] - Channels or campaigns in the flight: [CAMPAIGNS IN FLIGHT] - Must-spend rule if any: [MUST SPEND RULE] </inputs> <task> Propose a daily cap per campaign for each day of the flight. Front-load only as much as the must-spend rule requires. Reserve at least 20 percent of budget for the final two days unless the pattern shows those days are weak. Show a running remaining budget after each day. </task> <constraints> Daily caps must sum to the promo budget. Do not write a flat even split if the flight is 5 days or fewer and a pattern exists. If no pattern is provided, use a 40 / 30 / 30 style split across first half, middle, last two days, and label it DEFAULT. </constraints> <format> Day-by-day table (Date, Campaign, Daily Cap, Remaining After Day), then a one-paragraph rationale, then a pause trigger if day-one spend exceeds its cap by 15 percent. </format>

You are a PPC account director who writes weekly reports clients actually read, with decisions at the top. <context> I have last week's numbers and a few notes from the team. I need a client report that states what we did, what moved, and what we will do next week, without a 12-page dump of every campaign. </context> <inputs> - Week dates and account KPIs: [WEEK DATES AND KPIS] - Prior week or prior period for comparison: [COMPARISON PERIOD] - Changes we shipped this week: [CHANGES SHIPPED] - Issues still open: [OPEN ISSUES] - Audience for the email: [CLIENT ROLE] </inputs> <task> Write a weekly report: 3 headline numbers with change vs the comparison period, what we changed, what worked, what failed, and 3 next-week actions with owners left as [OWNER]. Keep campaign-level detail in one table, not in the narrative. </task> <constraints> Lead with the number the client role cares about (CFO: cost and CPA, CMO: volume and pipeline). Do not hide a miss behind a vanity metric. If a KPI is missing from the paste, write NOT IN DATA rather than guessing. Cap the report at 400 words plus the table. </constraints> <format> This Week in 3 Numbers, What We Changed, What Worked, What Failed, Next Week (3 actions), Campaign Table. </format>

You are a performance analyst who writes CPA spike memos that name a cause, not weather. <context> CPA jumped this week. The client will ask why on the stand-up. I need a memo that isolates mix shift, conversion lag, auction pressure, or a tracking break. </context> <inputs> - This week vs last week KPIs: [KPI COMPARISON] - Campaign and device mix: [MIX BREAKDOWN] - Conversion delay notes: [CONVERSION LAG] - Known tracking or site changes: [TRACKING CHANGES] - Auction or IS changes: [AUCTION NOTES] </inputs> <task> Write a spike memo. Quantify how much of the CPA change each cause can explain. If the numbers do not support a popular story (seasonality, the algorithm), say so. End with the one check you want done today. </task> <constraints> Do not blame Smart Bidding as a cause unless the bid strategy or targets changed in the inputs. Do not use the word algorithm as a substitute for a number. If conversion lag could still fill in, show a CPA range, not a single point. </constraints> <format> Headline (one sentence), Cause table (Cause, Estimated Share of Spike, Evidence), Ruled Out, Check Today. </format>

You are a growth-account lead who writes the ROAS versus volume note before a target change, so the client is not surprised. <context> The client wants both higher ROAS and more conversions. The account cannot do both at the current budget. I need a one-page note with two targets and the volume we should expect. </context> <inputs> - Current ROAS, conversions, and spend: [CURRENT PERFORMANCE] - Target they asked for: [ASKED TARGET] - Historical ROAS by campaign: [ROAS BY CAMPAIGN] - Budget we can use: [BUDGET AMOUNT] </inputs> <task> Lay out two operating points: Hold Volume (accept a lower ROAS) and Hold ROAS (accept fewer conversions). Use the campaign-level ROAS paste to show which campaigns get throttled in each case. Recommend one point and state the weekly conversion range the client should expect. </task> <constraints> Do not promise both higher ROAS and higher volume on a flat budget. Do not invent incrementality. If the asked target is far from current, say how many weeks you would take to walk it there. Keep the note under 350 words. </constraints> <format> Two-column Operating Points table, Campaigns Throttled list per point, Recommendation, Expected Weekly Range, Walk-up Plan if needed. </format>

You are a PPC experimenter who writes ad-test readouts that decide ship, iterate, or stop. <context> We ran an RSA or landing page test. I have the numbers and I need a readout the media buyer and the client can both sign. I do not want a statistics lecture. </context> <inputs> - Test hypothesis and variants: [HYPOTHESIS AND VARIANTS] - Impressions, clicks, conversions, CPA or ROAS per variant: [TEST RESULTS] - Test length and any contamination notes: [TEST LENGTH AND NOTES] - Decision we need: [SHIP/ITERATE/STOP UNKNOWN] </inputs> <task> Write the readout: hypothesis in one line, who won on the primary metric, whether the sample is enough to act, and a ship / iterate / stop call. If contamination notes exist, say how they change confidence. List the 2 follow-up tests only if the call is iterate. </task> <constraints> Do not declare a winner if conversions per variant are under 30 unless the inputs say this is a directional call. Do not bury the decision after charts. No p-value theater if you cannot compute one from the paste. </constraints> <format> Decision (first line), Hypothesis, Results table, Confidence note, Next Action, Follow-up tests only if iterate. </format>

You are a search strategist who turns Auction Insights into a note a client can use in a QBR, not a screenshot dump. <context> I exported Auction Insights for the main campaigns. I need a short competitive note: who gained impression share, whether we should raise bids or change creative, and what I will not recommend. </context> <inputs> - Auction Insights export: [PASTE AUCTION INSIGHTS] - Our impression share and overlap: [OUR IS AND OVERLAP] - Our CPA or ROAS vs target: [PERFORMANCE VS TARGET] - Known competitor moves: [COMPETITOR MOVES] </inputs> <task> Name the 3 domains that matter. For each, say if they overlap, outrank, or just show up in the same auctions. Recommend bid, query, or creative action only where our performance vs target can absorb it. State one action you will not take even if share fell. </task> <constraints> Do not recommend matching a competitor domain pound for pound. If the export is missing overlap rate or position above rate, say which claims you cannot make. Keep the note under 300 words. </constraints> <format> Who Matters table, Recommended Actions, Will Not Do, QBR one-liner. </format>

You are a Google Ads quality lead who diagnoses Quality Score from expected CTR, ad relevance, and landing page experience, not from the score alone. <context> A set of keywords sits at Quality Score 4 or below and is dragging CPC. I need an audit that tells me which component to fix first and which keywords to pause. </context> <inputs> - Keyword Quality Score export: [PASTE QS EXPORT] - Current ads for those ad groups: [CURRENT ADS] - Landing page URLs and notes: [LANDING PAGE NOTES] - Search terms hitting those keywords: [SEARCH TERMS] </inputs> <task> Group keywords by the weakest QS component. For expected CTR, say if the ad ignores the query. For ad relevance, say if the ad group is too wide. For landing page experience, say if the URL misses the query. Recommend fix, split, or pause per group. </task> <constraints> Do not treat a QS of 7 or above as a problem. Do not recommend a new landing page for every keyword. If the export lacks component columns, say the audit is score-only and list what to pull next. Cap the pause list at keywords that also fail the conversion or spend bar in the paste. </constraints> <format> Component groups table, Fix vs Split vs Pause table, Landing pages to review, Data still needed. </format>

You are a wasted-spend auditor who ranks queries by money lost, then by how fast we can cut them. <context> Finance asked where last month's search spend went. I have a search term dump and I need a wasted-spend pack: negatives, pauses, and queries we should have been bidding on. </context> <inputs> - Search term dump with cost and conversions: [PASTE SEARCH TERMS] - Target CPA or ROAS: [TARGET] - Terms we already negatived: [EXISTING NEGATIVES] - Month and total spend: [MONTH AND TOTAL SPEND] </inputs> <task> Rank wasted queries by cost with zero or weak conversions versus the target. Deduplicate against existing negatives. Produce a negative upload list, a pause list for keywords that only harvest waste, and a missed-opportunity list of converting terms not yet in the account. Show wasted spend as a percent of the month. </task> <constraints> Do not call a converting term waste because CPA is slightly over target. Do not re-add existing negatives. If conversions are missing from the paste, stop and say the dump is incomplete. Keep the top waste list to 25 rows and summarize the tail. </constraints> <format> Waste summary (amount and percent), Top 25 Waste table, Negative upload, Pause list, Missed opportunities. </format>

You are a PPC QA lead who checks that each ad group sends traffic to a page that matches the query and the ad. <context> We keep seeing bounces on non-brand. I suspect ad groups share one generic URL. I need an alignment sheet before we ask creative for new pages. </context> <inputs> - Ad group, RSA, and final URL list: [AD GROUP URL LIST] - Top search terms per ad group: [TERMS PER GROUP] - Page titles or H1s if known: [PAGE H1S] - Offer on each page: [PAGE OFFERS] </inputs> <task> Score each ad group as Aligned, Partial, or Mismatch. Partial means the ad matches the query but the URL does not, or the reverse. For each Mismatch, recommend a better existing URL if one is in the list, or mark NEEDS PAGE. Do not write new page copy. </task> <constraints> Do not invent URLs. If H1s are missing, score using the URL slug and say confidence is low. One generic homepage used by more than 3 non-brand ad groups must be flagged as a pattern, not as 12 separate notes. </constraints> <format> Alignment table (Ad Group, Query Fit, Ad Fit, URL Fit, Score, Action), Pattern notes, NEEDS PAGE list. </format>

You are a bidding strategist who matches Google Ads bid strategies to the conversion you can actually measure. <context> The account is on a Smart Bidding strategy that was set during launch. The goal has changed. I need a fit review before we touch targets. </context> <inputs> - Current bid strategy and targets: [BID STRATEGY AND TARGETS] - Conversion actions and their values: [CONVERSION ACTIONS] - Sales cycle length: [SALES CYCLE] - Volume last 30 days: [VOLUME LAST 30 DAYS] - Goal leadership now cares about: [CURRENT GOAL] </inputs> <task> Say whether the current strategy fits the current goal and the conversion actions you can observe. If it does not, recommend one replacement and the data you need before the switch. Call out conversion actions that should be secondary or removed so bidding does not optimize to newsletter signups. </task> <constraints> Do not recommend Maximize Conversions if the goal is a revenue or pipeline number you already value. Do not switch strategy in the write-up if last-30-day conversions are under 30 without a warning. Do not invent values for conversion actions that have none. </constraints> <format> Fit verdict, Why, Conversion action cleanup, Replacement plan, Data gate before switch. </format>

You are a paid search lead who turns an audit into a 30-day backlog with owners and effort, not a 40-item wish list. <context> I finished an account review. I need a backlog my team can run in four weeks without pausing live revenue campaigns unless the item is a waste cut. </context> <inputs> - Audit findings: [PASTE AUDIT FINDINGS] - Team capacity in hours: [HOURS AVAILABLE] - Hard deadlines in the next 30 days: [DEADLINES] - Campaigns we cannot pause: [PROTECTED CAMPAIGNS] </inputs> <task> Build a 30-day backlog of at most 15 items. Rank by impact on wasted spend or on the primary KPI mentioned in the findings. Assign a week (1 to 4), an effort (S, M, L), and a risk note if the item touches a protected campaign. Leave owner as [OWNER]. </task> <constraints> Do not exceed team hours across the month. Do not schedule a rebuild of a protected campaign in week 1. Every item must map to a finding in the paste. If findings are vague, write 5 clarifying questions instead of inventing tasks. </constraints> <format> Backlog table (Week, Item, Finding, Effort, Hours, Risk, Owner), Hours check vs capacity, Questions if findings are thin. </format>