CTV Playbook for Performance Marketers

Connected TV (CTV) advertising gives performance teams premium, full-screen video inventory with more addressability than linear TV. You can use it to influence direct response outcomes (site visits, leads, purchases, app actions) while also lifting demand that shows up later in search, social, and email.

You’ll also hear OTT advertising used alongside CTV. OTT refers to streaming content delivered over the internet; CTV refers to the connected-TV screen/device where ads are served. In practice, many campaigns buy OTT streaming inventory that’s viewed on CTV devices—so the terms often travel together.

This playbook follows the practical order that keeps CTV accountable:

Define attribution before buying media.
Build a three-tier audience architecture.
Develop conversion-focused creative.
Choose buying methods that match your goals.
Set budgets and frequency with enough room to learn.
Optimize against downstream outcomes (not vanity metrics).
Validate lift with incrementality and media mix modeling.

Start with attribution before you buy

Most underperforming CTV programs share one root issue: measurement was treated as reporting, not a launch requirement.

Before you spend, define what you’re trying to change (and how you’ll prove it). CTV advertising can influence direct visits, branded search, app activity, lead submissions, purchases, store visits, and assisted conversions. If you judge it only on last-click, you’ll often under-credit its contribution and overreact to early results.

Your measurement plan should answer:

Primary outcome: what event matters most (purchase, qualified lead, subscription, pipeline milestone)?
Secondary outcomes: what leading indicators are acceptable early (site visits, engaged sessions, add-to-cart, form starts)?
Attribution window: what time horizon matches the buying cycle?
Identity and exposure: how will you determine who was exposed (household/user) and when?
Baseline/control: what will you compare against (holdout, geo control, matched households, pre/post with controls)?

Platform dashboards are useful for directional optimization, but they’re rarely the full truth. Build a plan that combines what the platform can report with what your business needs to know.

A practical CTV attribution stack often includes:

Pixel/tag-based web events (and/or app SDK events)
Server-side events (where available) to reduce signal loss
Exposure logs (or equivalent reporting) that support exposed vs. control analysis
CRM/offline conversion uploads when lead quality matters
Incrementality testing for lift validation

For a deeper measurement framework, read our CTV attribution guide.

Build a three-tier audience architecture

CTV targeting performs best when each audience has a job. Instead of one broad segment, build a three-tier architecture so you can control frequency, learn faster, and assign the right creative to the right intent level.

Tier 1: Core conversion audiences

Highest intent, closest to action. These audiences should get the clearest conversion message and the most explicit next step.

Recent site visitors who didn’t convert
Cart/checkout abandoners or lead form starters
CRM lists (prospects, lapsed customers, churn-risk segments)
High-intent modeled audiences (e.g., likely converters)
Category researchers with strong purchase signals

Tier 1 is often small. Manage household saturation aggressively and rotate creative sooner than you would in broader tiers.

Tier 2: Qualified expansion audiences

Scale without going fully broad. These audiences are close enough to demand that you can still evaluate performance with downstream outcomes.

Lookalikes based on purchasers or qualified leads
In-market segments aligned to your category
Adjacent-interest audiences with strong fit
Contextual viewers consuming relevant content
High-performing geos based on historical demand

Tier 3: Strategic reach audiences

Broader reach for future demand creation and category presence. Tier 3 should be evaluated through lift and mix contribution—not only immediate conversions.

Broad household segments (with guardrails)
Geo-first targets (priority markets, store radius, service areas)
Premium content environments aligned to brand context
Seasonal/event-driven audiences when timing matters

For inventory, targeting, and supply-path fundamentals, see our connected TV advertising guide.

Develop creative that’s built for conversion

Performance results usually come from creative discipline, not just targeting. The common failure mode is running a generic brand video and hoping the platform “finds” converters.

Conversion-focused CTV creative typically includes:

A clear problem or promise in the opening seconds
Brand shown early (not just at the end)
One primary message (avoid stacked claims)
Simple proof points (without stretching credibility)
A single call to action and a friction-reducing next step
A visible mechanism to act: short URL, search prompt, QR code, offer, or app instruction

Design for sound-off comprehension, but don’t assume sound-off behavior. The goal is clarity either way.

Match creative to tier intent

Tier 1: direct response messaging (finish a quote, book a consult, start a trial, claim an offer).
Tier 2: benefit-led education (pain point → solution → differentiator → low-friction next step).
Tier 3: category and brand positioning (who it’s for, why now, and a simple action for later recall).

Test creative with a decision rule

Creative testing works when you define what “winning” means before you launch. If budget allows, test at least:

Problem-led hook vs. benefit-led hook
Offer CTA vs. educational CTA
QR code vs. short URL/search prompt
15s vs. 30s (and keep message density appropriate)
Demo-focused vs. scenario/story-focused

Judge tests using downstream behavior (qualified actions, conversion rate, cost per qualified outcome), not just completion rate.

Choose the right buying method

Buying method determines control, transparency, optimization levers, and how quickly you can learn. For performance marketers, the best choice is the one that supports clean measurement and repeatable iteration.

Buying method	Best for	Strengths	Tradeoffs to watch
Direct publisher	Premium placement, specific publishers, tight brand context	Clear inventory source; stronger environment control	Higher minimums; fragmented reporting; slower cross-partner optimization
Programmatic CTV (DSP)	Structured testing and scalable performance programs	Broader supply access; flexible targeting; pacing and frequency tools; centralized reporting	Supply-path discipline required; quality varies; fees and intermediaries need review
Private marketplace (PMP)	Curated supply with programmatic controls	Better transparency than open exchange; negotiated access to premium inventory	Deal quality varies; limited scale; requires ongoing performance validation
Managed service	Teams that need operational support or faster setup	Lower internal lift; help with trafficking and optimization	Less hands-on control; reporting and optimization logic may be less visible

Note: “OTT advertising” often describes the streaming content/inventory layer, while “CTV advertising” describes where those OTT ads are viewed. Your buying method can touch both.

Set budgets and frequency for learning (not just launch)

Underfunded CTV campaigns produce inconclusive results: not enough reach, not enough frequency, and not enough outcome signal to make decisions.

Start by defining the campaign’s job:

Conversion capture (often Tier 1-heavy)
Prospecting into qualified audiences (Tier 2 emphasis)
Regional demand generation (geo and tier blend)
Strategic reach (Tier 3 with lift validation)

Then allocate budget by tier based on audience size and the signal you expect to measure. If you need to prove incremental lift, reserve budget and time for a clean test design.

Frequency guidance

Frequency should be high enough to create recall and action, but capped to avoid household waste.

Tier 1: tighter caps; watch saturation quickly; refresh creative more often.
Tier 2: moderate caps; test creative and segments before scaling.
Tier 3: plan frequency across a longer window; evaluate with lift and mix contribution.

Common fatigue signals:

Delivery concentrates into the same households
Downstream action rate declines while spend holds
Site visitation lifts flatten despite stable reach
Creative performance drops across segments

Campaign setup steps (in the order that prevents waste)

Objective: define the business outcome and the decision you want to make after the test.
Conversions: lock primary/secondary events and confirm they’re measurable.
Tracking: validate pixels/server events, offline uploads, and reporting access before launch.
Audiences: build Tier 1/2/3 and document each audience’s job.
Creative mapping: assign creative versions by tier intent.
Buying method: choose direct/programmatic/PMP/managed based on control and measurement needs.
Inventory standards: define exclusions, brand safety requirements, and supply-path preferences.
Budgets and caps: set pacing, frequency, and a learning window you won’t cut short.
Reporting views: set breakdowns by tier, creative, geo, inventory, and outcome.
Test plan: define what will change if results are strong, weak, or mixed.

Use naming conventions that make analysis easy (tier + audience + creative + market + inventory + test cell). Clean structure is a performance lever.

Measure what happens after exposure

CTV advertising should be evaluated on business movement after exposure—not only impressions and video completion rate.

Metrics that tend to matter:

Exposed household/user site visits and engaged sessions
View-through conversions (with a defined window)
Assisted conversions and path influence
Qualified leads/purchases (and quality signals when available)
New customer share (when measurable)
Geo lift or market-level movement when running regional tests

Use platform reporting to diagnose delivery and directional performance, then validate impact with your broader measurement plan.

Prove incrementality (not just attribution)

Attribution assigns credit; incrementality answers whether outcomes increased because the ads ran. For performance marketers, this is the difference between “tracked” and “true.”

Common options:

Holdout tests: keep a clean control group unexposed.
Geo tests: compare exposed markets vs. control markets.
Matched household tests: match exposed to similar unexposed households.
Pre/post with controls: time-based analysis with careful baseline selection.
Media mix modeling (MMM): quantify channel contribution in the full mix.

To evaluate CTV alongside search, social, linear TV, audio, and other channels, use our media mix modeling guide.

Optimize against downstream outcomes

Optimization should follow the customer journey. Early on, you may use delivery quality and visitation as leading indicators; as signal builds, move decisions closer to revenue, qualified leads, or other primary outcomes.

A practical optimization cadence:

Confirm inventory quality, tracking, and pacing.
Manage reach/frequency by tier to avoid saturation.
Compare audiences by downstream action rate (not just completion rate).
Compare creative by qualified outcome efficiency.
Reallocate budget toward the best audience × creative combinations.
Audit performance by geo and supply source.
Re-check incrementality as you scale.
Refresh before fatigue shows up across tiers.

Don’t overreact to short windows. CTV needs a learning period. The discipline is making fewer, higher-quality changes based on outcomes you trust.

Common mistakes (and how to prevent them)

Launching before attribution and lift validation are defined
Optimizing to completion rate as a proxy for business impact
Using one broad audience with no tiered structure
Over-retargeting and saturating small Tier 1 pools
Running brand-only creative with no clear action step
Buying low-quality inventory without guardrails
Stopping the test before the buying cycle has time to show movement
Reporting platform conversions without an incrementality check

Launch checklist

Objective, primary KPI, and decision rule documented
Attribution window and methodology approved
Tracking tested (web/app) and offline/CRM path confirmed if needed
Tier 1/2/3 audiences built, named clearly, and assigned a job
Suppression applied (existing customers, recent converters, exclusions as needed)
Creative mapped by tier; CTA visible and easy to act on
Buying method and inventory standards documented (including exclusions)
Budget, pacing, and frequency caps set to support learning
Reporting views ready (tier, creative, geo, inventory, outcome)
Incrementality approach planned (holdout/geo/matched/MMM)
Weekly optimization owner and review schedule assigned

Key takeaways

Plan CTV advertising like a performance channel with upper-funnel influence—measure it accordingly.
Define attribution and lift validation before you buy inventory.
Use a three-tier audience architecture and map creative to intent.
Choose buying methods based on control, transparency, and your measurement requirements.
Optimize toward downstream outcomes and re-check incrementality as you scale.

FAQs

What is connected TV advertising?

Connected TV advertising is video advertising delivered into streaming apps and content viewed on internet-connected TVs (smart TVs or streaming devices). It’s commonly used to reach households with premium video and then measure downstream actions.

What’s the difference between CTV advertising and OTT advertising?

CTV refers to the connected-TV device/screen; OTT refers to streaming content delivered over the internet. Many OTT ads are watched on CTV devices, which is why the terms are often used together in media planning.

Can CTV advertising drive direct response conversions?

Yes—when your attribution plan, audience tiers, and creative are designed for action. The strongest programs measure beyond last-click and optimize toward qualified downstream outcomes.

What KPIs should performance marketers use for CTV?

Use completion rate and reach/frequency to diagnose delivery, but prioritize exposed traffic, qualified conversions, cost per qualified outcome, assisted conversions, and incremental lift.

How do you control frequency in CTV campaigns?

Set frequency caps by tier, monitor delivery concentration into the same households, and rotate creative before fatigue appears. Tier 1 typically needs tighter caps than broader prospecting tiers.

Is programmatic CTV better than direct publisher buying?

Programmatic is often better for structured testing, pacing, and scaling. Direct buys can be better when you need specific publishers or premium placement. The “best” option is the one that supports your measurement requirements and optimization workflow.

How do you prove incrementality for CTV?

Use holdouts, geo tests, matched household designs, and/or media mix modeling, depending on your footprint and what you need to prove. Define the method before launch so you don’t retrofit the test later.

Put the playbook into action

If you want CTV to perform, keep the sequence: define attribution first, build tiered audiences, run creative built for action, choose buying methods with the right controls, and optimize to downstream outcomes validated by lift.

Explore our CTV advertising services or continue with the connected TV attribution guide.