Paid Advertising: Campaign Strategy and Optimization
Mandatory Content Standards
- Match output length to the skill, request, and deliverable type. Use concise answers for quick checks, structured detail for audits and plans, and full-length output only when the user asks for a complete deliverable.
- Write in a way that sounds like a knowledgeable human wrote it. No robotic or templated phrasing.
- Use short sentences. One idea per sentence. One focus per paragraph.
- Use active voice. Never passive constructions.
- Address the reader directly using "you" and "your."
- Use bullet points only when they genuinely improve readability.
- Replace all em dashes with commas, parentheses, semicólons, or a new sentence. No hidden Unicode characters.
- End every sentence with a period.
- No hashtags, emojis, or asterisks.
- No introductory or closing filler phrases such as "in conclusion," "in summary," or "in a world where."
- No warnings, notes, or disclaimers. Stick to requested output.
- No AI cliches: no "game-changer," "unlock," "leverage," "dive into," "delve," "cutting-edge," "transformative," "revolutionize."
- No excessive adjectives or adverbs. Let specifics do the work.
- No broad generalizations. Every claim tied to specific context.
- Use specific examples, data, and scenarios.
- Pose at least one thought-provoking question per skill.
- Mobile-friendly: short paragraphs, clear headers, scannable.
- Practical and actionable. Every section connects to a next step.
What This Skill Covers
This skill covers paid advertising across the major platforms: Google Ads, Meta, LinkedIn, and Twitter/X. It covers how to design campaign strategy from scratch, structure campaigns correctly, set bids and budgets, choose and build audiences, select creative rotation approaches, analyze performance, and continuously optimize toward your business goals.
Paid advertising is the fastest way to generate demand you can measure. It is also the fastest way to burn budget on traffic that never converts. The difference between those two outcomes is strategy, structure, and discipline in optimization.
Campaign Strategy Before You Spend
The most common paid advertising mistake is launching campaigns before answering the questions that determine whether any campaign structure can succeed.
Before you allocate budget, answer these:
What is your conversion event? Not "generate awareness." A specific, trackable action: a purchase, a completed form, a qualified demo booking, a trial signup. Everything downstream of this decision depends on it.
What is your target cost per conversion? Work backward from your unit economics. If your average contract value is $3,000 and your sales close rate from a demo is 25%, each qualified demo is worth up to $750 in expected revenue. What portion of that can you spend on customer acquisition? That determines your target cost per lead or cost per trial.
What conversion rate does your landing page currently achieve? If you send 1,000 people to a page that converts at 1%, you get 10 conversions. If your CPC is $3, that is $300 per conversion. If your page converts at 4%, that same spend produces 40 conversions at $75 each. Page conversion rate is often more important than targeting or bidding decisions.
What is your minimum viable budget? Every platform has a minimum threshold below which the algorithm does not have enough data to optimize. On Meta, campaigns spending less than $50 per day often do not exit the learning phase. On Google, smart bidding requires at least 30 conversions in 30 days to function well. If your budget cannot support that, adjust your expectations or structure.
These questions are the foundation. Teams that skip them optimize the wrong things.
Google Ads: Structure and Strategy
Campaign Types
Search campaigns show text ads to users actively searching for your keywords. This is demand capture: users who are already looking for something you offer.
Performance Max (PMax) campaigns use Google's full inventory: Search, YouTube, Display, Discover, Gmail, and Maps. PMax gives Google maximum control over placement and targeting in exchange for broader reach.
Display campaigns show image ads across Google's Display Network: millions of third-party websites. Display is primarily for awareness and retargeting.
Shopping campaigns show product listings with images and prices directly in search results. These are essential for e-commerce.
Video campaigns run on YouTube. Pre-roll, mid-roll, and in-feed video ads for awareness and consideration.
Campaign Structure: The SKAGs Approach vs. Broad Match with Smart Bidding
Two dominant philosophies exist for Google Search campaign structure.
The first is Single Keyword Ad Groups (SKAGs), where each ad group contains one keyword match type. This gives you tight control over which ads show for which searches and precise Quality Score management. It requires significant time to build and maintain.
The second is a broader structure using broad match keywords with Smart Bidding. Google has argued for this approach as its algorithm improves. The logic: broad match with Target CPA or Target ROAS bidding lets Google find conversions that exact match structures miss, because users search in ways you have not anticipated.
The right choice depends on your situation. SKAGs work best for accounts with clear, established keyword lists and a desire for control. Broad match plus Smart Bidding works best when you have enough conversion volume (30 or more per month) for the algorithm to optimize against and when you want to capture longer-tail queries at scale.
Keyword Match Types
Exact match: your ad shows only when someone searches your exact keyword or close variants. Most precise, typically highest intent, usually highest CPC.
Phrase match: your ad shows when someone searches a phrase containing your keyword. More flexible than exact, still relatively controlled.
Broad match: your ad shows for searches Google considers related to your keyword. Highest volume, lowest precision, requires strong negative keyword management to avoid wasted spend.
Negative keywords are as important as positive keywords. Add negatives proactively based on the search terms report. Any query that triggered your ad but is clearly irrelevant should become a negative keyword. Review the search terms report weekly.
Bidding Strategy
Manual CPC: you set bids yourself. Full control, but requires continuous management. Works when you do not have enough conversion data for Smart Bidding.
Enhanced CPC: Google adjusts your manual bids up or down based on likelihood of conversion. A light step toward automation.
Target CPA: Google targets your specified cost per acquisition. Requires 30 or more conversions in the last 30 days. Effective when your CPA target is realistic.
Target ROAS: Google optimizes to achieve your specified return on ad spend. Best for e-commerce with clear revenue values attached to conversions.
Maximize Conversions: Google spends your budget to maximize total conversions. No efficiency target. Use when you want to grow volume and have budget flexibility.
Start with Manual CPC or Maximize Conversions if you are below the 30-conversion threshold. Move to Target CPA or Target ROAS once you have the data to support it.
Meta Ads: Structure and Strategy
Campaign Objective Selection
Meta's campaign objectives determine how the algorithm optimizes and whom it targets. Choosing the wrong objective is one of the most common structural errors in Meta accounts.
Awareness objectives optimize for reach and impressions. Use only for brand campaigns where you genuinely want to maximize exposure, not conversions.
Traffic objectives optimize for link clicks. The algorithm sends you people who click ads. Not necessarily people who convert. Using Traffic objective for a conversion campaign is a common and expensive mistake.
Lead Generation objectives optimize for Facebook native lead forms. These convert well on mobile because users do not need to leave Facebook. Lead quality is often lower than website leads because the friction is lower.
Conversions objectives optimize for your specified conversion event (purchase, lead, trial signup) on your website. This is the correct objective for most direct response campaigns. It requires your pixel to be installed and conversion events to be firing correctly.
App Installs objectives optimize for app store installs. Use for app-focused campaigns.
Campaign Structure: CBO vs. ABO
Campaign Budget Optimization (CBO) manages budget at the campaign level and distributes spend across ad sets automatically. Meta's algorithm decides which ad sets and ads get the most spend based on performance.
Ad Set Budget Optimization (ABO) gives each ad set its own budget. You control how much each audience or placement gets.
CBO is generally recommended now because it lets Meta's algorithm optimize spend in real time. ABO still makes sense when you need guaranteed spend to a specific audience (for example, a retargeting audience that is too small to compete with prospecting in CBO).
Audience Targeting
Core Audiences: targeting based on demographics, interests, and behaviors. Facebook's interest targeting has become less precise as privacy changes reduced signal. It works best for broad demographic targeting combined with behavioral signals.
Custom Audiences: audiences built from your own data. Website visitors (pixel-based), customer email lists, video viewers, app users. These are your most valuable audiences because they have already interacted with your brand.
Lookalike Audiences: Meta finds users who resemble your Custom Audience. A 1% Lookalike of your customer list targets the 1% of Facebook users who most resemble your customers. Usually the highest-performing prospecting audience for established businesses.
Broad targeting: no audience restrictions beyond basic demographics. Counter-intuitively, broad targeting with strong creative often outperforms interest targeting because Meta's algorithm has maximum flexibility to find converters. Worth testing, especially for campaigns with strong conversion history.
Creative Rotation Strategy
Meta rewards fresh creative. Ads that have been in rotation for too long fatigue the audience. Frequency rises, CPM rises, and CTR falls.
Monitor ad frequency at the ad set level. When any ad exceeds a frequency of 4 to 6 within a 30-day window, introduce new creative. Do not wait until performance drops sharply. Prevention is cheaper than recovery.
Run at least three to five creative variants per ad set at all times. This gives the algorithm choices and slows the pace at which any single piece of creative fatigues.
Set up an ongoing creative review cadence. Every two weeks, review frequency and CTR by creative. Pull the lowest performers. Add new variants. This keeps the account healthy without constant manual intervention.
LinkedIn Ads: Structure and Strategy
When LinkedIn Makes Sense
LinkedIn CPCs typically range from $6 to $12 or higher. This makes LinkedIn viable only for offers with sufficient deal value to support those costs.
A general rule: LinkedIn makes economic sense if your target customer has an average deal value above $5,000, or if your monthly recurring revenue per customer exceeds $300 to $500. Below those thresholds, the CPA math rarely works.
LinkedIn is particularly effective for reaching specific professional roles that are difficult to target elsewhere: enterprise buyers, specific job functions, specific seniority levels, or employees at specific companies (Account Based Marketing).
Campaign Structure
LinkedIn campaigns are structured as: Campaign Group > Campaign > Ad.
Within a campaign, you set one objective, one audience, and one budget. Unlike Meta, LinkedIn does not have sophisticated automatic budget optimization across campaigns. Manage budgets at the campaign level.
LinkedIn's daily budget minimum is $10. The platform recommends a minimum of $50 to $100 per day per campaign to generate meaningful data. Below that threshold, campaign delivery is inconsistent.
Audience Targeting Options
Matched Audiences: upload your own contact lists or company lists. Target existing prospects or customer lookalikes.
Account Targeting: target employees at specific named companies. This is the core of Account Based Marketing on LinkedIn.
Retargeting: target users who visited your website, viewed your LinkedIn ads, or engaged with your LinkedIn content.
Interest and Skill Targeting: target by professional skills, group memberships, or interests. Less precise than account or list targeting.
Job Function and Seniority: target by department and level. This is often the most practical approach for broad demand generation to a defined professional profile.
Ad Formats
Single Image Ads: the standard feed ad. Effective for most campaigns.
Message Ads: sent directly to a prospect's LinkedIn inbox. Higher CPL but higher engagement rates from relevant audiences.
Conversation Ads: branching message ads that let users self-select their interests. Good for segmenting intent levels.
Document Ads: promote a downloadable document directly in the feed. Effective for content marketing and lead gen.
Retargeting: Structure and Priority
Retargeting reaches users who have already interacted with your brand. These users convert at significantly higher rates than cold audiences because they have prior exposure. Your retargeting audiences are your highest-value segments.
Structure retargeting campaigns by intent level:
High intent: users who reached your pricing page, checkout, or demo booking page but did not convert. These users are closest to a decision. Retarget them with specific offers, social proof, or objection-handling content.
Medium intent: users who visited key product or feature pages but did not reach a conversion-point page. Retarget them with education and value proposition content.
Low intent: users who visited your homepage or blog but did not go deeper. Retarget them with awareness-level content or case studies that build interest before making an offer.
Keep retargeting audiences separated from prospecting campaigns. This lets you set different bids, use different creative, and report results separately. Combining retargeting and prospecting into one audience is a common mistake that inflates your prospecting results with retargeting performance.
On Meta, create website custom audiences with appropriate time windows: 7 days for high-intent pages, 30 days for mid-funnel pages, 90 days for general site visitors.
ROAS and CPA: Setting and Managing Targets
Setting Your CPA Target
Your target CPA should reflect your unit economics, not a number someone found in an industry benchmark.
Work backward from these inputs:
Customer lifetime value (LTV): what is the average customer worth over their full relationship with your business?
Gross margin: what percentage of revenue is gross profit?
Payback period: how many months are you willing to wait to recover acquisition costs?
Sales close rate (for B2B with a sales team): what percentage of leads or trials become customers?
Example: SaaS business with $400 monthly LTV, 70% gross margin, and a 12-month payback period target. Allowable LTV = $400 x 12 months x 70% = $3,360 per customer. With a 20% trial-to-customer conversion rate, the allowable cost per trial is $3,360 x 20% = $672.
That number is your ceiling, not your target. Running campaigns at the ceiling leaves no profit. A reasonable starting target CPA might be 50% to 60% of that ceiling, with room to expand if LTV assumptions prove conservative.
Setting Your ROAS Target
For e-commerce, ROAS (return on ad spend) is the primary efficiency metric.
Minimum viable ROAS = 1 / gross margin percentage.
If your gross margin is 50%, you need a minimum ROAS of 2x just to break even on ad spend at a gross profit level. A target ROAS of 3x or 4x is necessary to cover overhead and generate actual profit.
Segment ROAS targets by customer type. New customer ROAS targets should be more conservative than repeat customer ROAS targets, because the incremental value of a new customer includes future LTV while a repeat customer's repeat purchase is already captured in LTV modeling.
Budget Allocation Across Platforms
When you are advertising across multiple platforms, budget allocation is a strategic decision, not just a math problem.
Allocate budget proportional to proven ROAS or CPA performance, not proportional to reach or audience size. If Google search converts at $40 CPA and LinkedIn converts at $200 CPA for the same customer profile, Google should receive a higher share of budget unless there is a strategic reason to invest in LinkedIn awareness that you expect will pay off downstream.
Reserve 10% to 20% of your total paid budget for testing. This budget goes to new channels, new audiences, or new creative concepts that are not yet proven but have strategic potential. Keep this test budget separate from performance budget. Evaluate test budget on learnings, not on CPA.
Adjust allocation monthly based on performance data. Platforms that improve get more budget. Platforms that remain inefficient get less. Do not maintain a fixed allocation plan across a quarter without reviewing performance data.
Campaign Optimization Cadence
Set a structured cadence for reviewing and adjusting campaigns. Ad hoc optimization leads to reactive decisions. A structured cadence leads to systematic improvement.
Weekly review tasks:
For Google: review the search terms report and add negatives. Check Quality Score changes by keyword. Review impression share for your top keywords. Adjust bids for ad groups where CPA has moved significantly above or below target over the past 14 days.
For Meta: review frequency by ad set and creative. Check CPM trends. Review CTR by ad. Pull low-frequency, low-CTR ads. Add new creative where frequency is approaching the threshold.
For LinkedIn: check delivery pacing. LinkedIn campaigns often underspend. Check relevance scores and engagement rates by ad. Pause low-engagement ads.
Monthly review tasks:
Audience performance: which audiences are converting at the lowest CPA? Increase budget to high performers. Consider building lookalike audiences from your highest-quality converters.
Creative analysis: which creative concepts have performed best over the month? Document the pattern. Brief new creative in the same direction.
Funnel analysis: are the leads or trials from your paid campaigns converting downstream? If paid trial users convert to customers at 10% but organic trial users convert at 25%, your paid targeting may be bringing in the wrong people.
Common Campaign Mistakes
Running campaigns without conversion tracking. If you cannot measure conversions, you cannot optimize toward them. Verify that your conversion events are firing correctly before spending. Use Google Tag Assistant, Meta's Test Events tool, and platform-specific debugging tools to verify.
Too many campaigns competing for the same audience. Running five campaigns targeting the same audience segments splits your budget and creates auction overlap. Consolidate campaigns and let your budget concentrate.
Pausing campaigns and restarting them. Every time you restart a Meta or Google campaign after a pause of more than a few days, the algorithm resets its learning. Stopping and starting campaigns wastes learning data. If a campaign is underperforming, adjust the creative or audience before pausing.
Evaluating performance before sufficient data exists. Making bidding or budget decisions after three days of data is unreliable. Give new campaigns two to four weeks and at least 30 conversion events before drawing conclusions.
Ignoring landing page quality. Your ad drives the click. Your landing page drives the conversion. Optimizing the ad without addressing landing page quality is like improving your restaurant's signage while leaving the kitchen broken.
Attribution Models and What They Mean
Attribution tells you which touchpoints get credit for a conversion. The model you choose changes which campaigns appear profitable.
Last click: the final touchpoint before conversion gets all credit. This favors brand search and retargeting, which tend to be the last click. It undervalues top-of-funnel campaigns that initiated awareness.
First click: the first touchpoint gets all credit. This favors prospecting and upper-funnel channels. It ignores the role of later touchpoints.
Linear: each touchpoint in the path gets equal credit. More balanced but still arbitrary.
Data-driven (Google) or Algorithmic: the platform's algorithm assigns fractional credit based on observed contribution to conversion. The most accurate but the least transparent.
No attribution model is perfect. Use a consistent model across your campaigns so comparisons are apples-to-apples. Supplement platform attribution with multi-touch attribution tools (Triple Whale, Northbeam, Rockerbox) for a more complete picture, especially when you run across multiple platforms.
Which attribution model is your team using, and have you deliberately chosen it, or is it just the platform default?
Reporting That Drives Decisions
Your paid advertising reports should answer one question: what should we do differently next week?
Build a weekly one-page dashboard that shows:
- Total spend by platform
- Conversions by platform
- CPA or ROAS by platform
- Week-over-week change in CPA or ROAS
- Top-performing campaigns and ads
- Campaigns or ads that should be paused or adjusted
Share this with relevant stakeholders every week. Consistency builds trust. Surprises destroy it.
Monthly, add a second layer of reporting that covers downstream quality: lead-to-customer conversion rates by paid channel, average deal size by acquisition source, and LTV trends for cohorts acquired through paid.
This downstream data is what separates paid advertising teams that drive business results from teams that optimize for platform metrics that do not connect to revenue.