Paid Media Agency: How to Choose the Right Partner in 2025 (Pricing, Scorecards, and Top Picks)

You’re choosing a paid media agency while juggling budget certainty, speed to impact, and stakeholder trust. This expert guide gives you clear pricing ranges, a decision scorecard, platform playbooks, and a 90‑day onboarding plan so you can hire with confidence.

What a Paid Media Agency Actually Does (and When You Need One)

When performance stalls or you’re scaling into new channels, a paid media agency can unlock your next growth curve. The right partner blends strategy, buying, creative, and analytics to drive business outcomes like ROAS, CAC/LTV, and pipeline contribution.

Core services: planning, buying, optimization, creative testing, analytics/reporting

A modern performance marketing agency aligns channel plans to your revenue model and margins, then manages execution end to end. Expect hypothesis-driven testing, weekly optimizations, and transparent reporting that ladders up to CFO-ready KPIs. The goal is to turn media into a repeatable system, not a series of one-off tactics. For example, strong agencies connect experiments to forecasted outcomes and document decision gates. That rigor lets you predict impact and scale with fewer surprises.

Media planning and forecasting
Campaign builds, trafficking, and QA
Bid strategy and budget pacing
Creative strategy, production, and iteration loops
Analytics, dashboards, and experimentation (A/B, geo, holdouts)
Data integrations (pixels, CAPI, offline conversions, CRM)

Great agencies publish verifiable case studies with attributable impact and explain their optimization loop (inputs, cadence, decision gates). If you can’t see the “how,” you can’t judge whether results are repeatable.

Channels covered: search (Google/Microsoft), social (Meta, TikTok, LinkedIn), programmatic/CTV, Amazon

Top paid media agencies operate full-funnel across search, social, marketplaces, and programmatic. They select channels based on audience fit, unit economics, and incrementality—then orchestrate them so each role in the funnel is clear. You should see a rationale for how channels work together and what each is expected to contribute. For instance, search may harvest demand while CTV builds reach measured via lift tests. That channel clarity makes budget allocation and stakeholder reporting straightforward.

Search/PPC: Google Ads & PMax, Microsoft Ads, brand defense, Shopping
Paid social: Meta (Advantage+), TikTok Spark Ads, LinkedIn, X/Reddit (selected)
Programmatic & CTV: The Trade Desk, DV360, YouTube, Hulu/CTV marketplaces
Retail media: Amazon Ads, Walmart Connect, Instacart
B2B extras: LinkedIn Lead Gen, content syndication, ABM, offline conversion uploads

Ask for platform certifications (e.g., Google Partner, Meta Business Partner, Amazon Ads Verified Partner) and how they’ve adapted to privacy changes (iOS, cookies) by channel. This is where many “set-and-forget” shops fall behind, and where experienced agencies will show concrete mitigation steps.

In-House vs Paid Media Agency vs Hybrid: A Decision Framework

The wrong operating model costs time and money; the right one compounds learning and results. Use this framework to choose in-house, an external media buying agency, or a hybrid approach based on constraints and goals, not preferences.

Scorecard: team skills, creative bandwidth, tooling, speed-to-value, total cost of ownership

Start by scoring your current state and needs. Weight what matters most in the next 2–3 quarters, not in theory. Then pressure-test assumptions with real budget scenarios and deadlines. This keeps the decision anchored to impact and risk, not org charts. Revisit your score quarterly to reflect changing goals.

Team skills and seniority: strategy, platform depth, experimentation, analytics
Creative capacity: volume, variety, and speed of asset production
Tooling and data: CMP, MMP, clean room access, dashboards, QA
Speed-to-value: onboarding pace, decision cadence, test velocity
Total cost of ownership: salaries, benefits, tools, fees, creative, ramp time

If your team lacks two or more of the above for the next big goal (e.g., PMax rollout + CTV test + creative sprints), an agency or hybrid model typically wins on time-to-impact. Revisit the score quarterly as needs shift.

When hybrid works best: internal strategy + external execution

Hybrid is often optimal for brands with strong in-house strategy and analytics but limited channel bandwidth or creative volume. You own roadmap and business context while the agency brings channel firepower, production scale, and cross-account pattern recognition. This split protects institutional knowledge while accelerating execution.

Use hybrid when you want to keep brand stewardship, data modeling, and vendor contracts in-house, while outsourcing day-to-day buying, testing, and asset iteration. This model also reduces vendor risk; if the agency underperforms, your internal team still holds the keys and IP.

How Much Does a Paid Media Agency Cost?

Pricing opacity causes decision drag and misaligned expectations. Here are the common pricing models, typical ranges by spend tier, and what’s usually included so you can budget and negotiate with clarity.

Pricing models: % of spend, flat fee, hourly/retainer, performance/hybrid

Pricing shapes incentives and risk allocation. Choose the model that best aligns to your goals and budget predictability. Consider how variable your monthly spend is and how many channels you’ll run. Also clarify what’s inside the fee versus billed separately. These details prevent bill shock and protect ROI.

Percent of spend (most common): 8–20% of media spend; aligns to scale but can overpay at high budgets without stair-steps or caps.
Flat fee/retainer: predictable monthly fee tied to scope; great for stable budgets or multi-channel programs with variable mix.
Hourly: transparent time-based billing; useful for audits, consulting, or overflow work, but not ideal for ongoing buying.
Performance/hybrid: base retainer plus bonus on ROI/CAC or revenue; aligns incentives but requires clear baselines and data integrity.

Negotiate floors/ceilings, channel-specific rates, and fee reviews at budget inflection points. For complex stacks, hybrid models reduce misaligned incentives while protecting downside.

Typical ranges by spend tier (SMB, mid-market, enterprise) and minimums

Ranges vary by complexity, channel mix, and creative scope, but these benchmarks will anchor your planning. Align the fee to the level of hands-on management and experimentation you expect. Ask for a stair-step model as spend scales. This keeps effective rates fair as you grow.

SMB (monthly ad spend $10k–$75k): 12–20% of spend or $2.5k–$10k flat; common minimum ad spend $5k–$15k; light creative included.
Mid-market ($75k–$500k): 8–15% or $8k–$40k flat; tiered/stair-stepped fees; common minimums $20k+ for programmatic/CTV.
Enterprise ($500k–$5M+): 5–12% or $40k–$200k+ flat; often hybrid with in-house analytics; creative billed separately or via SOW.

Programmatic advertising agencies and CTV partners may add tech/platform fees (e.g., DSP seat, verification) of 5–15% of media. Ask for an all-in effective rate card to avoid hidden add-ons.

What’s included: reporting cadence, meetings, creative iterations, experiments/month

Scope clarity prevents “empty retainer” syndrome. Define what success requires and lock it into your MSA/SOW. Spell out the operating cadence, deliverables, and owners. Quantify creative volume and testing commitments so resourcing is unambiguous. This makes performance and accountability measurable.

Weekly dashboards and monthly/quarterly business reviews tied to CFO metrics
Named strategist + buyer + creative lead with backup coverage
Creative volume targets (e.g., 10–20 net new ads/month) and iteration cadence
Experimentation quota (e.g., 2–4 tests/month across audience, bids, creative)
Analytics support: pixels/CAPI, offline conversions, attribution guidance

Ask for a 30-60-90 plan with milestones and the first 3 experiments upfront. If a PPC agency can’t outline test design before kickoff, expect more management than optimization.

How to Evaluate a Paid Media Agency (With a Downloadable RFP Template)

A structured evaluation lowers risk and speeds alignment. Use the criteria below and standard RFP sections to compare paid media agencies apples-to-apples and defend your recommendation.

Must-have proof: platform certifications, case studies with verifiable KPIs, client tenure

Evidence beats promises. Require third-party verifications and proof you can audit. Look for math you can replicate and tenure that signals stability, not churn. Reference checks should match your stage and use case. This due diligence separates marketing from salesmanship.

Certifications: Google Partner, Meta Business Partner, Microsoft Ads Partner, Amazon Ads Verified, The Trade Desk Edge Academy
Case studies with math you can replicate (baseline, tactic, KPI lift, time frame)
Client tenure and retention rate by industry and budget tier
References you can call, ideally matched to your use case
Named team bios and roles, not just logos

Validate claims by requesting read-only access to redacted accounts or log-level reports where feasible. If results can’t be tied to attribution choices, treat them as anecdotes, not proof.

Data ownership, tech stack, and transparency (dashboards, account access)

Control of data and accounts is non-negotiable. You should own ad accounts, pixels, audiences, and dashboards. Document who pays for which tools and how data flows. Ensure export rights at any time, not just at termination. That control protects continuity and compliance.

Ensure all accounts are in your org with your billing; agency gets partner access
Require real-time dashboards (Looker, GA4, BigQuery, Power BI) with clear source notes
Clarify data retention policies and export rights at termination
Document tools used (verification, brand safety, MMM/MTA, clean rooms) and who pays

If a media buying agency resists transferring ownership or limits access, consider it a red flag on both ethics and continuity risk.

Communication and accountability: who runs the account, SLAs, escalation paths

Great results come from a predictable operating cadence. Lock service levels into the contract. Define who pushes buttons daily and how decisions are recorded. Pre-commit to test velocity and review forums so learning compounds. Clarity here prevents drift and finger-pointing.

Response SLAs: e.g., critical issues 4 business hours, standard 1 business day
Meeting cadence: weekly working session, monthly strategy review, QBRs
Experiment cadence: minimum tests/month, pre-registered hypotheses, success criteria
Change windows: bid/creative changes logged and time-stamped
Escalation: named sponsor, backup coverage, and executive escalation path

Ask who actually pushes the buttons daily and the average seniority of that role. Senior strategy without daily rigor is a common failure mode.

Interview questions to ask (and red flags to watch for)

Use interviews to expose the real optimization process and confirm senior talent will touch your account. Push for specifics, artifacts, and examples where things didn’t go to plan. You want a learning culture, not bravado. The answers should map to your 90-day plan and KPIs.

Ask:

Walk me through your 90-day plan for a brand like ours. What are the first three tests?
How do you use PMax/Advantage+ while maintaining control and brand safety?
Show an example where MMM and MTA disagreed—how did you decide budget?
What’s your creative iteration loop and weekly decision meeting format?
How do you measure incrementality on CTV or upper-funnel channels?

Red flags:

Vague “we let the algorithm learn” without test plans
No read-only account examples or unwillingness to discuss failures
Overreliance on last-click or vanity metrics
Creative treated as a one-time deliverable vs a weekly system
No clear owner for analytics and data QA

Platform Playbooks: What ‘Good’ Looks Like by Channel

Surface-level tips won’t move your KPIs. These proven practices keep control while leveraging automation on each major platform, with guardrails that protect brand and profit.

Google Ads & PMax: signals, tROAS/tCPA, asset groups, brand protection

Performance Max thrives on clean signals and clear structure. Feed it high-quality audiences, product feeds, and conversion signals, then constrain where needed. Start with achievable targets and expand as signal density improves. Keep brand control separate to avoid cannibalization. Treat it like a portfolio with rules, not a black box.

Build asset groups by product theme or intent cohort; exclude brand into separate Search for control
Use tROAS/tCPA with realistic targets; start looser, tighten as signal quality improves
Connect first-party data (Customer Match), offline conversions, and enhanced conversions
Add brand negatives to PMax and protect with exact-match brand Search
Monitor search term insights and placement reports; use URL expansion controls

Source: Google Ads Help Center and field insights show PMax performance correlates with feed health, signal density, and realistic conversion lags. Treat PMax as a portfolio strategy with guardrails, not a black box.

Meta Advantage+: creative iteration loops, signal quality, conversion APIs

Meta’s algorithm rewards rapid creative learning and robust server-side signals. Build a weekly creative lab and harden your data plumbing. Use broad targeting once signals are strong; otherwise, start narrower. Test hooks and offers deliberately and retire losers quickly. The fastest learners usually win on Meta.

Implement Conversion API with deduplication; verify Events Manager quality diagnostics
Use broad or Advantage+ audiences once signals are strong; start with stacked interests otherwise
Run creative in themes; test hooks, formats, and offers systematically (3–5 new ads/week)
Keep pixel hygiene strong: prioritized events, aggregated event measurement, 7–28 day windows by goal
Separate prospecting vs remarketing budgets; protect branded terms with exclusions where relevant

Source: Meta Business Help Center and creator-led testing show iterative creative velocity often drives larger lifts than micro-targeting. Scale what wins quickly; retire losers ruthlessly.

LinkedIn for B2B: audience expansion, lead quality safeguards, offline conversions

LinkedIn is powerful for B2B when you prioritize quality over cheap form fills. Build for pipeline, not CTR. Start with a tight ICP and expand as signal quality improves. Close the loop with offline conversions and value-based optimization. This keeps spend aligned to revenue, not vanity volume.

Start with tight ICP (titles, functions, company size, industries), then layer audience expansion
Use Lead Gen Forms judiciously; enrich via CRM and score with downstream stages (MQL→SQL→Closed Won)
Upload offline conversions and use value-based optimization when eligible
Protect quality with exclusion lists (competitors, students, job seekers) and firmographic filters
Optimize creative for clarity: problem-solution, stat proof, and a single CTA

Source: LinkedIn Marketing Solutions guidance emphasizes offline conversions and matched audiences for high-intent optimization. Expect higher CPCs but better pipeline density.

Programmatic & CTV: brand safety, audience curation, incrementality testing

Programmatic and CTV shine for reach and mid-funnel lift—if you curate inventory and measure incrementality. Avoid spray-and-pray by using curated deals and verification. Control frequency to protect efficiency and brand experience. Then validate with holdouts or geo tests. This approach proves impact beyond last-click.

Use curated PMPs and program lists; apply GARM suitability, IAS/DoubleVerify/MOAT verification
Build first-party and contextual segments; test retail media data overlays where relevant
Cap frequency by channel; unify via DSP where possible
Run geo or PSA holdouts to quantify lift and guard against attribution inflation
Align creative to TV norms: clear branding in 2–3 seconds, subtitles, and sound-off comprehension

Source: IAB and The Trade Desk best practices show lift is real when brand safety, audience quality, and frequency are controlled—and when testing isolates incrementality.

Measurement That Holds Up: MMM vs MTA, Attribution Windows, and Incrementality

Attribution is messy; your budget shouldn’t be. Pick a measurement approach that fits your data reality and validates investment beyond clicks so finance and growth can align.

Choosing the right model for your data reality

Marketing Mix Modeling (MMM) and Multi-Touch Attribution (MTA) solve different problems. MMM is great for cross-channel planning at a macro horizon; MTA helps at the micro level where identity resolution is viable. Most teams benefit from a hybrid that pairs models with experiments. Align reporting windows with your sales cycle to avoid false negatives. Then socialize a single source of truth for planning.

Use MMM when spend >$1–2M/year, multiple channels, and privacy limits user-level tracking; update quarterly with weekly reads
Use MTA where signals are strong (search, app, logged-in ecosystems) and for granular optimization
Adopt hybrid: MMM for budget allocation, MTA for in-channel decisions, and experiments to calibrate both
Match attribution windows to buying cycle; report in parallel (e.g., 7-day click + MMM-lift) to align finance and growth teams

The best paid media agencies show a point of view on trade-offs and will help you align stakeholders on a single source of truth for planning and reporting.

Designing clean incrementality tests (geo, PSA, holdouts)

Experiments are your BS detector. Design for isolation, power, and business realism. Choose the test type that fits channel and budget. Pre-register hypotheses and guard against contamination. Then report deltas with confidence intervals tied to CAC/LTV.

Pick a test type: geo split (city/region), PSA holdout (charity ads), ghost bids, or audience holdouts
Power the test: estimate sample size and duration to detect a meaningful lift (e.g., +10% conversions)
Pre-register the hypothesis, success metrics, and stop conditions
Control contamination: frequency caps, exclusion lists, and creative consistency
Report deltas with confidence intervals; translate lift into CAC/LTV and budget guidance

Run at least one lift test per quarter on big-budget channels (Meta, CTV, YouTube). Tests create organizational conviction and protect your budget during planning cycles.

Privacy, Brand Safety, and Compliance

Privacy rules and suitability standards affect how you target and what you can measure. Build compliance into your media operations to protect brand and performance and safeguard long-term data quality.

Data retention, consent, and clean room basics

Consent is the new currency for signal quality. A compliant setup increases match rates and improves optimization. Document data flows and retention so audits are straightforward. Use clean rooms where advanced analysis is needed without exposing PII. Treat brand safety standards as default, not optional.

Consent management: implement a CMP, honor GDPR/CCPA, and configure Google Consent Mode v2
Data retention: define lookback windows, access controls, and deletion SLAs in your DPA
Clean rooms: use platforms like Ads Data Hub, Amazon Marketing Cloud, or Infosum for privacy-safe joins and cohort analysis
Brand safety: apply GARM standards, blocklists/allowlists, and verification (IAS/DoubleVerify/MOAT)
Documentation: maintain a data map, DPIAs where needed, and incident response steps

Your agency should brief you on how privacy changes impact targeting, measurement, and creative—and propose mitigation plans, not excuses.

The First 90 Days With a Paid Media Agency (Onboarding Plan)

A dependable 30-60-90 day plan makes value visible and accountable. Use this timeline to set milestones, deliverables, and decision gates so everyone knows what “good” looks like.

Day 0–30: audits, access, baseline reporting, quick wins

The first month is about foundations and early impact. Expect deep audits, instrumentation, and low-risk wins. Lock naming conventions and dashboards early to avoid rework. Prioritize fixes that cut waste and improve signal quality. This creates momentum while the bigger tests queue up.

Secure access to ad accounts, analytics, product feeds, and creative libraries
Complete audits (structure, queries, audiences, creative, tracking) with a prioritized fix list
Implement pixels/CAPI, offline conversions, enhanced conversions, and naming conventions
Launch quick wins: brand defense cleanup, wasted spend cuts, creative refreshes, and feed fixes
Stand up dashboards and agree on baseline KPIs, budgets, and experiment slots

If you don’t see a written audit and prioritized action plan by Day 21, escalate. Sloppy foundations compound into waste.

Day 31–60: experimentation roadmap, creative sprints, bid strategy shifts

Month two should show measured lifts and faster iteration. This is where good agencies separate from order-takers. Run structured tests with pre-registered hypotheses and clear stop criteria. Shift bidding as signal quality improves, and add automation with guardrails. Pilot a secondary channel once core KPIs stabilize.

Launch 2–3 structured experiments (audience, offer, format, bidding)
Start a weekly creative lab: 3–5 new ads, tight feedback loop, and pre/post metrics
Transition to tROAS/tCPA or value-based bidding where signal quality allows
Introduce platform automation (PMax, Advantage+) with guardrails and brand protection
Run a pilot on a secondary channel (e.g., YouTube, TikTok, LinkedIn) if primary KPIs are stable

Insist on a mid-point review with learnings, reallocations, and hypotheses for the next 30 days. Progress should be traceable to specific changes.

Day 61–90: scale decisions, budget reallocation, KPI re-baselining

By month three you should be compounding what works and killing what doesn’t. Document how learning translates into scale and forecast. Increase creative velocity around winning themes and expand audiences as signals strengthen. Schedule a lift test for upper/mid-funnel if budget allows. Close with a cross-functional plan for next quarter.

Reallocate budgets toward proven campaigns and audiences; pause underperformers
Increase creative velocity around winning themes; refresh landing pages if needed
Expand automation guardrails and audience breadth as signals strengthen
Decide on a lift test for upper/mid-funnel (YouTube/CTV) or a new market rollout
Re-baseline KPIs with finance; plan next quarter’s roadmap and test slate

Close with a QBR that includes a refreshed forecast, risk register, and an updated scorecard. This locks accountability for the next quarter.

Who’s Best for What? Agencies by Use Case and Industry

Different industries require different muscles. Use these “best for” markers to narrow your shortlist of top paid media agencies without guesswork and avoid misfit partners.

eCommerce and retail: catalog scale, feed management, creative velocity

For eCom, product feed excellence and rapid creative iteration drive most of the lift. Look for retail-ready processes and marketplace fluency that translate into ROAS and contribution margin. Expect rigorous brand protection in Shopping and PMax. Tie creative and CRO to product economics. These capabilities separate catalog movers from true growth partners.

Proven Google Shopping/PMax and Advantage+ Shopping chops, with brand protection
Feed management (schema, attributes, promotions) and retail media experience (Amazon, Walmart)
High-volume creative workflows (UGC, short video, dynamic templates) and CRO chops
LTV modeling and contribution margin-aware bidding
Returns, inventory, and price-change signals integrated into campaigns

Ask for category-specific benchmarks (CPC/CPA/ROAS) by AOV band and seasonality expectations.

B2B SaaS and lead gen: pipeline KPIs, lead quality, sales alignment

B2B winners optimize to revenue stages, not lead counts. Favor agencies that integrate with your CRM and sales process. Insist on offline conversion uploads and value-based optimization. Guard quality with enrichment and scoring. The outcome to watch is cost per qualified opportunity, not just MQLs.

Offline conversion uploads, value-based optimization, and robust UTM governance
LinkedIn ICP targeting, content offers that map to buying committee roles, and ABM options
Lead quality safeguards (forms, enrichment, scoring) and clear handoff SLAs to SDRs
Long-cycle attribution fluency (7–28+ day windows, view-through treatment) and MMM alignment
Thoughtful content syndication and retargeting nurtures without fatigue

Expect pipeline benchmarks (SQL rate, cost per qualified opportunity) and proof they’ve lifted sales acceptance rates, not just MQL volume.

Enterprise and regulated: governance, compliance, procurement fit

Enterprises need governance, audit trails, and cross-functional alignment. Choose a performance marketing agency built for complexity. Confirm privacy posture and documentation standards upfront. Ensure procurement requirements won’t stall execution. The right partner will show mature controls without slowing test velocity.

Documented controls: brand safety standards, incident response, change logs, and approvals
Privacy compliance and data governance (DPA, DPIA, clean rooms, role-based access)
Procurement readiness: indemnification, insurance, SOC 2/ISO references where applicable
Multi-region ops, translation/localization, and center-of-excellence playbooks
Executive-ready reporting and planning integration with finance

Ask how they handle conflicts, escalations, and regional playbook variations. Governance failures are expensive.

Comparison Matrix and Selection Methodology

Standardize your evaluation to compare top paid media agencies fairly. Use the methodology below to build a defensible shortlist and secure stakeholder buy-in.

Methodology: sources, criteria, scoring weights

Gather inputs from multiple sources, score consistently, and weight what matters for your goals. Balance outcome proof with operating rigor and creative velocity. Validate claims via a pilot or audit before final selection. Document assumptions and evidence so the decision stands up to scrutiny.

Sources: client references, read-only account reviews, case studies, platform certifications, Glassdoor (talent quality), and your pilot results
Criteria:
Outcomes and proof (30%)
Strategy and testing rigor (20%)
Channel and platform depth (15%)
Creative capabilities and velocity (15%)
Analytics/measurement maturity (10%)
Governance, SLAs, and culture fit (10%)
Process: score 1–5 for each criterion per vendor; run a pilot or audit to validate; require a written 90-day plan before final selection

Document assumptions and keep evidence in a shared folder. Transparency strengthens stakeholder buy-in.

Download: Vendor scorecard + RFP template

Copy this structure into your doc or sheet to speed selection. Standardize the brief so responses are comparable and easy to score. Request artifacts (dashboards, test plans, sample QBR) with each submission. This reduces back-and-forth and surfaces execution quality early.

Company profile: industries, stage fit, min spend, average client tenure
Team: named roles, seniority mix, availability, backup coverage
Services: channels, creative scope, analytics, experimentation cadence
Pricing: model, inclusions, change thresholds, creative costs
Tech/data: dashboards, integrations, data ownership, privacy posture
SLAs: response times, meetings, experiment cadence, escalation
Case asks: 2–3 relevant case studies with data you can verify
RFP questions: see interview list above; add a 90-day plan and a sample QBR

Share the same brief, KPIs, and budget scenario with all vendors so answers are comparable.

FAQs

What are typical paid media agency pricing ranges by spend tier, and which model fits which scenario?
SMB often pays 12–20% of spend or $2.5k–$10k flat; mid-market 8–15% or $8k–$40k; enterprise 5–12% or $40k–$200k+. Percent of spend scales but needs caps; flat fees suit stable budgets; hybrids align incentives.
How do I decide between in-house, agency, or a hybrid model for paid media?
Score team skills, creative capacity, tooling, speed-to-value, and TCO. If you’re missing two or more for the next growth goal, go agency or hybrid; keep strategy and data in-house if you have them.
What should a 30-60-90 day onboarding plan with a paid media agency include?
Day 0–30: audits, tracking fixes, dashboards, quick wins. Day 31–60: structured tests, creative sprints, bidding upgrades. Day 61–90: scale winners, reallocate budgets, plan next quarter.
Who owns the ad accounts, data, and dashboards when working with an agency?
You should. Place all accounts and billing under your org, grant partner access, and ensure export rights and data retention are in your DPA/SOW.
How do MMM and MTA differ, and which should my team use?
MMM guides cross-channel budget at macro level; MTA supports micro optimizations where identity exists. Use hybrid: MMM for planning, MTA for in-channel, experiments to calibrate.
What interview questions reveal an agency’s true optimization process and seniority on my account?
Ask for a 90-day plan, how they control PMax/Advantage+, how they reconcile MMM vs MTA, their creative loop, and a recent lift test. Push for specifics and artifacts.
How do agencies implement Google PMax and Meta Advantage+ without losing control of targeting and brand safety?
They structure asset groups, use brand negatives and exact-match brand, set realistic tROAS/tCPA, maintain exclusions, and enforce verification and suitability standards.
What SLAs should be in my paid media agency contract?
Response times (4 hours critical), weekly meetings, experiment quotas, dashboard refresh cadence, change logs, and executive escalation paths.
How do privacy regulations (GDPR/CCPA) and clean rooms affect media measurement and targeting?
Consent governs signal quality and remarketing reach; clean rooms enable privacy-safe aggregation and cohort analysis. Configure a CMP and consider clean-room workflows for advanced analysis.
What KPIs and benchmarks should I expect by industry and budget level?
Benchmarks vary by AOV, LTV, and cycle. Ask for category-specific CPC/CPA/ROAS or SQL/CAC ranges and seasonality by budget tier; insist results map to profit or pipeline.
When is programmatic/CTV worth it, and how do I measure incrementality credibly?
When you’ve saturated lower-funnel channels and need reach; measure with geo/PSA holdouts and verification controls to isolate lift and avoid double-counting.
What red flags indicate an agency is over-automating or under-testing creative?
Vague “let the algo learn,” no written tests, low creative volume, no brand safety plan, and one-size-fits-all attribution.