TV Measurement and Attribution: How it works and how Tatari Gets it Right

What Is Measurement for TV Advertising and Why Is It So Hard to Get Right?

This guide is for performance marketers, brand advertisers, and agency buyers who want to understand how TV advertising measurement works — and how to choose the right methodology. It covers the core concepts behind TV attribution, the difference between linear and streaming measurement, how Tatari's platform handles the unique challenges of convergent TV, and what metrics actually matter for performance-oriented TV investment.

In this guide:

• Why TV attribution is harder than digital — and how it's solved

• The difference between linear, CTV, and convergent TV measurement

• How Tatari's three attribution models work and when to use each

• What incrementality testing is and why it matters more than attribution

• How DragFactor accounts for delayed TV response

• Key metrics performance TV advertisers use instead of GRPs

TV measurement is the practice of connecting ad exposures on television to real business outcomes like visits, installs, purchases, and revenue. Done well, it tells you not just how many people saw your ad, but how many acted because of it.

The challenge is that TV (including streaming TV) is primarily a passive medium. Viewers don’t click. They watch, and then seconds, minutes, hours, or days later — they might search for your brand, open your app or website, or place an order. That delayed, indirect path makes attribution fundamentally harder than digital channels, where every click leaves a traceable record.

The problem is compounded by three structural realities:

Multi-device households. Let’s say a TV ad airs in a household of four. Any one of them might convert on any device — phone, tablet, laptop — minutes or weeks later. Without a device graph that connects household members, you’ll miss most of that signal.

Shared IP addresses. Traditional streaming TV attribution matched conversions to ad exposures by IP address. But apartment buildings, offices, and campus networks share IPs across hundreds of people, creating false matches that inflate results.

Passive, delayed response. Unlike a paid search click, TV’s effect decays over time. Some categories see the bulk of response within two minutes of an ad airing. Others see it spread across two weeks. Using the wrong attribution window either overcounts (by capturing organic conversions) or undercounts (by cutting off the tail).

Tatari’s measurement stack was built to solve all three. This deep dive explains how.

The Fundamentals: What are we talking about when we say “TV”?

Before diving into measurement methodology, it helps to understand what you’re measuring. The TV ecosystem in 2026 spans three distinct delivery types — and each has different measurement implications.

What Is Linear TV?

Linear TV is traditional broadcast and cable television: scheduled programming delivered via satellite or cable networks. Ads air at fixed times within fixed shows. You buy a slot in Thursday Night Football or The Today Show, and your ad reaches everyone watching at that moment in that market.

Measurement implications: Linear TV has no user-level data. Historically, reach was estimated through Nielsen panel ratings, which is a sample of households used to extrapolate total viewership. You never knew exactly who saw your ad; you knew what show aired and what the ratings estimated. This is the world of GRPs (more on that below).

What Is CTV and Streaming?

Connected TV (CTV) refers to television content delivered over the internet to a TV screen — via smart TVs, Roku, Amazon Fire TV, Apple TV, or gaming consoles. The ads that run within that content are bought digitally (both direct and programmatically) with targeting and measurement closer to digital advertising than to linear TV.

OTT (over-the-top) is the broader term for streaming video delivered independently of traditional pay TV. CTV is a subset of OTT specifically referring to content viewed on a TV screen (as opposed to a phone or laptop).

Measurement implications: Because CTV is delivered digitally, each impression is logged with a device ID, IP address, and timestamp. That creates the foundation for deterministic attribution — matching an ad exposure to a downstream conversion — that linear TV cannot offer natively.

What Is Convergent TV?

Convergent TV is the umbrella term for the combined ecosystem of linear and streaming, reflecting the fact that most households use both. Tatari measures TV holistically across both delivery types, isolating the true incremental impact of the full TV investment rather than measuring each silo separately.

What Is a GRP — and Why Are Advertisers Moving Away From It?

GRP stands for Gross Rating Point. It is the traditional currency of linear TV advertising, calculated as: Reach % × Frequency = GRPs.

A campaign that reaches 40% of the target audience an average of 3 times delivers 120 GRPs. The metric tells you how much exposure you purchased. It says nothing about what that exposure caused.

GRPs were the standard because, for decades, there was no alternative. Linear TV had no user-level data. Panel-based ratings were the only signal available. GRPs gave buyers a common language for comparing placements and estimating audience size.

Why brands are moving away from GRPs

GRPs measure inputs, not outputs. A brand can accumulate 500 GRPs in a quarter and have no idea whether the ads had a direct impact on revenue. In a world where digital advertising channels can measure key business KPIs like cost-per-click, cost-per-install, and ROAS in real time, a metric that only estimates exposure has become increasingly hard to defend to a CFO.

Brands are also moving away from GRPs because the metric compresses too much information into a single number. A GRP of 120, for example, could represent vastly different realities: broad reach with moderate frequency (40% reach × 3 frequency) or extremely narrow reach with excessive repetition (1% reach × 120 frequency). These scenarios have fundamentally different implications for campaign effectiveness, yet appear identical in GRP terms.

As a result, GRPs are limited both in diagnosing campaign quality — how reach and frequency are distributed — and in evaluating true business impact. This makes them less useful for modern, performance-oriented advertisers.

The shift is also being driven by the rise of CTV. Streaming inventory is bought on CPM (cost per thousand impressions) and measured with digital-style attribution, making GRP-based planning feel incompatible with modern convergent TV buys. Performance advertisers in particular have pushed for outcome-based metrics: cost per visit (CPV), cost per install (CPI), customer acquisition cost (CAC), and ROAS — all of which Tatari reports natively.

GRPs haven’t disappeared. They remain useful for brand reach planning on linear, and most national linear buys are still negotiated in GRP terms. But they are no longer sufficient as the sole measure of TV effectiveness, particularly for performance-focused advertisers.

How Does TV Attribution Actually Work?

TV attribution is the process of linking a TV ad exposure to a downstream conversion, such as a website visit, app install, or purchase. Because viewers don’t click through on TV ads, TV attribution relies on probabilistic and deterministic matching rather than click-tracking.

Linear vs. Streaming Measurement: A Key Methodological Difference

The measurement approach for linear and streaming TV differs significantly, reflecting the fundamental difference in how audiences experience each format.

For linear TV, Tatari uses a probabilistic, baseline-lift methodology to determine immediate incremental lift. Because linear is a broadcast medium — many people see the same ad at the same moment — a measurable spike in site visitors is expected within a short window (typically 5 minutes) after airing. Tatari compares observed traffic against a dynamically updated baseline to isolate the lift from an ad, above what would have occurred organically. Critically, this does not require individual user identification via IP matching or device graph matching — the lift is calculated at the aggregate level.

For streaming TV, Tatari uses impression-level signals such as IP address, device information, and timestamps to connect ad exposure to response patterns. Because streaming audiences see ads at different times, the methodology combines matching and sessionization with baseline or exposed-vs-expected comparisons rather than relying on a single aggregate traffic spike.

For deep-funnel conversions (purchases, sign-ups, installs) across both formats, Tatari uses sessionization and available identity signals, including device graph data where applicable, to connect exposures and downstream events over time. This helps link a TV exposure to a conversion that may occur on a different device or days later without treating every match as causal on its own.

The Two Primary Matching Methods for CTV

IP matching: When a viewer is exposed to a CTV ad, the IP address of the household is logged. When a conversion occurs from a device sharing that IP, it is attributed to the exposure. This method is simple but prone to false matches from shared IPs (apartment buildings, offices) and misses conversions that happen on mobile networks.

Device graph matching: A more sophisticated approach uses a proprietary identity graph to connect multiple devices within a household. This allows Tatari to match a TV exposure logged at the household level to a conversion that happens on a phone, tablet, or laptop used by anyone in that home — substantially improving accuracy.

Both methods must also account for the baseline: the conversions that would have happened anyway, without any TV ad. Attribution that doesn’t subtract the baseline overcredits TV.

Tatari’s core attribution methodology is protected by USPTO patents. The foundational patent covers Tatari’s systems and methods for attributing TV conversions — what enables Tatari to identify exactly when a TV airing caused a downstream action, rather than crediting organic traffic.

Tatari’s Three Attribution Models

Tatari offers three distinct attribution models built on this patented infrastructure.

Tatari View-Through Attribution

The only view-through model designed specifically for TV. It uses Tatari’s proprietary device graph and filters out shared IPs to provide accurate household-level attribution, accounting for multi-device households and the delayed response patterns unique to TV. This is the recommended model for most established performance advertisers measuring TV on its own terms.

The scale of the communal IP problem is larger than most advertisers expect. In one Tatari client case study, only 28% of impressions were delivered across communal IPs — yet 57% of 1-day view-through responses originated from those same communal addresses, indicating substantial over-attribution risk when communal IPs are not filtered. Tatari View-Through corrects for this, resulting in a materially lower — and more accurate — response count that reflects genuine household-level causality rather than spurious IP matches.

Digital View-Through Attribution

Aligns with digital platform conventions by using IP matching and shorter attribution windows. This model exists specifically to enable apples-to-apples comparisons between TV and channels like Meta or Google — useful when you need to present TV performance in the same framework your digital team uses.

Incremental Lift Attribution

Rather than attributing individual conversions to exposures, Incremental Lift measures net-new conversions by comparing observed performance against a predictive baseline. It answers the question: “What would have happened if we hadn’t aired at all?” This is not a view-through model. It is a causal model, and it is the most defensible measure of TV’s true contribution.

How to Choose the Right Attribution Model

Tatari recommends choosing the attribution model based on campaign objective, traffic scale, paid-media mix, and the amount of household-matched signal available:

• Tatari View-Through is recommended when there is enough website traffic and household-matched signal to distinguish true response from noise. It is most useful for established advertisers measuring TV on its own terms.

• Digital View-Through is appropriate when the goal is a direct comparison of TV performance against other digital channels or CTV providers — it uses the same IP-matching conventions as platforms like Meta and Google, enabling apples-to-apples benchmarking.

• Incremental Lift is recommended when traffic volume or signal density makes household-level view-through attribution less reliable. The baseline methodology can be more useful when the goal is to understand absolute TV contribution.

For most established performance advertisers, Tatari View-Through is the default recommendation — with Incremental Lift used alongside it as a validation layer to estimate TV’s absolute contribution.

Attribution Model Comparison

Decision guide:

• Need to benchmark TV against your digital channels (Meta, Google)? → Digital View-Through

• Have sufficient traffic and want the most accurate household-level TV measurement? → Tatari View-Through

• Lower traffic volume, or need to answer "does TV actually drive growth?" → Incremental Lift

• Running an established performance campaign? → Tatari View-Through as primary + Incremental Lift as validation layer

What Is Incrementality Testing — and Why Does It Matter More Than Attribution?

Attribution tells you which conversions occurred after an ad exposure. Incrementality tells you which of those conversions likely happened because of the ad — and which would have happened anyway.

The distinction is critical. A retargeting campaign targeting people already browsing your product page will show excellent attribution numbers: many people who saw the ad converted. But some of them were already planning to buy. The ad didn’t cause the conversion. You were just giving TV credit for it.

The same dynamic applies to TV. A brand with strong organic search volume will see web traffic spike immediately after an ad airs — because some of that audience was already searching. If your attribution model doesn’t account for the organic baseline, you’ll overstate TV’s impact.

Incrementality testing solves this by establishing a counterfactual: what would have happened without the ad? The most rigorous approaches use controlled experiments:

• Geo holdouts: Test and control geographic markets are designated. TV runs in test markets; control markets receive no TV. The difference in conversion rates between markets, adjusted for baseline differences, represents the incremental lift.

• Audience holdouts: A randomly selected portion of the target audience is withheld from ad exposure. Their conversion behavior forms the control group baseline.

Tatari’s Incremental Lift attribution model uses a predictive baseline approach rather than requiring a live holdout test — making it practical to run on every campaign without sacrificing media efficiency. The model predicts expected traffic for each time period based on historical patterns and attributes only the lift above that prediction to TV.

The baseline calculation itself is protected by USPTO patent — the technology that allows Tatari to attribute only the lift genuinely above the organic baseline to TV, rather than overcrediting the channel. A related patent covers media creative attribution specifically: the process of determining a response profile for each individual TV spot within an attribution window. In plain terms, this is the technology that allows Tatari to tell you not just that TV drove lift, but which specific spots drove it and by how much.

What Is DragFactor? Understanding TV Response

One of the most underappreciated problems in TV measurement is how to account for delayed response. Some viewers act within seconds of an ad airing. Others respond days later. Waiting 30 days to capture the full picture makes in-flight optimization impossible. Tatari’s DragFactor solves this by expressing the relationship between immediate and delayed response as a single multiplier:

Immediate Response × DragFactor = Total Response

To understand DragFactor precisely, three related but distinct concepts need to be separated:

• The response window is Tatari’s fixed window for measuring immediate response after an airing — the spike in site visitors observed in the minutes directly following a spot. This window is fixed by Tatari’s methodology; it is not selected or configured by the marketer.

• Drag Factor operates beyond the response window. It is a multiplier that scales immediate response up to an estimate of delayed response over a longer measurement horizon, enabling next-day reporting without waiting for the full response to mature.

• The attribution window is a separate concept: how long after a session start a conversion event (purchase, sign-up, install) can be attributed to that session. This governs whether a conversion that happens hours or days after a viewer first engages is still credited to the originating TV exposure.

In practice, Tatari uses fixed response windows — marketers do not configure or select them. The response curve varies by category and creative format: direct-response spots generate a sharp spike within the first few minutes, followed by rapid decay; brand-oriented creative produces a flatter, longer tail with conversions distributed more evenly across hours and days. DragFactor models this variation, enabling accurate next-day reporting regardless of category.

Category and Custom DragFactors

All new clients begin with a category DragFactor, calculated using data from comparable companies in similar industries. If insufficient category data exists, a default DragFactor is applied.

Once a campaign has accrued sufficient data, Tatari calculates a custom DragFactor specific to that advertiser, refreshed monthly.

How Does Tatari Collect Data? Pixel vs. Server-to-Server

Accurate measurement starts with accurate signal collection. Tatari offers two implementation paths.

Tatari Pixel (TTM — Tatari Tag Manager)

A lightweight browser-based tag that records page visits, app installs, or purchases and links them back to ad exposures. It is the simplest implementation path and integrates natively with Shopify for e-commerce brands — deployable in a few clicks.

Events tracked by TTM are automatically sent to the Vault Data Clean Room, where events are tokenized and matched before results are passed to Tatari’s measurement pipeline.

Vault Server-to-Server (S2S)

Rather than placing a pixel on the client side, conversion events are sent directly to Vault’s privacy clean room infrastructure.  Similar to the TTM use case, user events are tokenized and matched before results are passed to Tatari’s measurement pipeline. PII is separated at the clean room layer and never reaches Tatari’s measurement environment.

The advantages over pixel-based tracking:

• Immune to browser-based tracking restrictions (Safari ITP, ad blockers, iOS privacy changes)

• More accurate across devices — Vault can link multiple devices within a household, improving attribution coverage

• Privacy enhanced — no client-side data exposure

Both methods feed the same attribution models, so advertisers can choose the implementation that fits their infrastructure without sacrificing measurement consistency.

What Are the Key TV Advertising Metrics?

Performance TV advertisers track a different set of metrics than traditional TV buyers. Here is what matters — and what each metric tells you.

• Cost per Visit (CPV): The cost of driving one incremental website visit via TV. Useful for direct-response campaigns focused on driving site traffic.

• Cost per Install (CPI): The cost of driving one app install attributable to TV. The standard metric for mobile-first brands using TV to scale user acquisition.

• Customer Acquisition Cost (CAC): Total TV spend divided by new customers acquired. The broadest measure of TV efficiency for subscription or DTC businesses.

• Return on Ad Spend (ROAS): Revenue (or gross margin) attributable to TV divided by media spend. The primary profitability signal for e-commerce advertisers.

• Reach, Frequency, and Impressions: Exposure metrics that remain relevant for brand-building objectives and for managing frequency overlap between linear and CTV.

Tatari’s dashboard lets you cut these metrics by network, creative, DMA, device, platform, and attribution model — and export raw event and attribution data to S3 for integration with your own BI tools.

The TV Advertising Halo Effect: How Does TV Affect Other Marketing Channels?

TV advertising doesn’t only drive direct conversions. It creates downstream effects across other channels — a phenomenon called the halo effect — that standard attribution misses entirely.

There are two types of halo effects: direct (within-user halo) and indirect.

Direct effects impact the same user exposed to an ad, increasing their likelihood to respond or convert across other marketing channels due to that prior exposure.

Indirect effects occur when users not exposed to the ad are influenced to higher response rates, either through platform-mediated dynamics (e.g., improvements to quality scores, keyword rankings, or delivery efficiency) or broader network effects.

Some direct effects include:

• Branded search lift: The increase in Google searches for your brand name in the minutes and hours after a TV spot airs. This signal can be observable and correlated with airtime when advertisers make timestamped search data available.

• Social engagement lift: Increases in social media mentions, followers, or engagement correlated with airtime.

• Direct traffic lift: Increases in direct (typed-in URL) website traffic following a spot.

• Third-Party Retail (Amazon) search lift: Increases in product searches on retail platforms in the hours following a spot, where the advertiser can provide timestamped data.

These effects matter for two reasons. First, they represent real business value that conversion-only attribution misses — meaning TV is almost certainly undervalued in models that only count direct visits and purchases. Second, they reveal how TV amplifies other channels: a paid search campaign will perform better in the 24 hours after a TV spot airs because branded search intent is elevated.

In fact, among 100 brands for which Tatari measured this effect for social marketing-driven traffic:

• Nine out of 10 brands saw higher CVRs from website visitors exposed to TV ads

• 60% of companies saw a CVR lift of over 50%

• 33% saw a lift of over 100%

Measuring halo effects accurately requires timestamped data from each downstream channel, correlated with airtimes. Tatari can support this analysis; brands are encouraged to surface data from search, social, and retail platforms alongside TV exposure data to build a complete picture.

How Do You Manage Frequency Across Linear and CTV?

One of the most common — and costly — mistakes in convergent TV advertising is serving the same household the same ad too many times by failing to coordinate frequency across linear and CTV buys.

Linear and CTV are typically bought through separate systems. A household might be served four exposures through a national cable buy and another four through a programmatic CTV campaign — delivering eight total impressions when four would have been sufficient and more cost-efficient. Beyond wasted spend, overexposure drives ad fatigue and declining response rates.

Effective frequency management in a convergent environment requires:

• Household-level identity resolution: The ability to recognize that the same household is being reached by both the linear and CTV buys — which requires a cross-channel device graph, not just siloed channel-level reporting.

• Unified frequency caps: Setting an overall household frequency target and coordinating delivery across both channels to stay within it, rather than capping each channel independently.

• Reach vs. frequency analysis: Understanding whether incremental spend is adding new households (reach) or adding more exposures to already-reached households (frequency) — and optimizing accordingly.

Tatari holds a USPTO patent specifically covering optimal frequency determination — a method that rolls up IP-level impression data to the household level, then correlates household-level impressions with response and conversion rates across frequency bins. This is the technical foundation that allows Tatari to tell advertisers not just how many times a household was reached, but at what frequency level their specific campaign stops paying off.

Reach and frequency reporting is available natively in the Tatari dashboard, allowing buyers to identify where frequency is accumulating and rebalance accordingly.

Should I Advertise on Linear TV, CTV, or Both?

The question most performance advertisers ask when entering TV is where to start. The answer depends on your objectives, budget, and target audience — but the data consistently points toward a hybrid approach outperforming either channel alone.

• Linear TV strengths: Unmatched reach for live events and cultural moments, including many of the most-watched U.S. telecasts each year. Strong performance with audiences 50+. Lower CPMs at scale. The “legitimacy effect” — TV still confers brand credibility in ways digital does not.

• CTV/Streaming strengths: Deterministic measurement. Addressable targeting (reach specific audience segments rather than relying on show demographics as a proxy). Faster optimization cycles — creative and placement can be adjusted weekly based on real-time performance. Incremental reach into cord-cutter households that linear doesn’t touch.

• The case for both: Many households use both streaming and traditional TV, while others skew heavily toward one environment. A linear-only buy can miss cord-cutters; a CTV-only buy can miss heavy linear viewers. Efficient campaigns often use linear for broad reach and scale, and CTV for precision, targeting, and impression-level measurement — then use a convergent measurement platform to understand the combined impact.

Tatari’s own data indicates that balancing CTV for determinism and linear for scale yields optimal results — a principle that applies broadly across verticals.

Linear vs. CTV vs. Convergent TV

Integrating TV With Your Broader Measurement Stack

Television doesn’t exist in isolation. Tatari integrates with the measurement ecosystem most performance brands already use.

Marketing Mix Modeling (MMM)

MMM uses multivariate regression to estimate how TV, alongside other channels, promotions, and external factors, influences sales over time. Because it operates on aggregate data rather than user-level matching, it is resilient to privacy restrictions and signal loss. Tatari supplies household-level TV exposure data to MMM partners including Keen, Measured, Marketing Attribution, and LiftLab.

Multi-Touch Attribution (MTA)

MTA tracks user-level events across digital media and connects them to conversions, modeling the full customer journey. Tatari integrates with Northbeam and Rockerbox, which model how TV influences lower-funnel actions alongside digital and out-of-home channels.

Design of Experiments (DOE)

Controlled geo-holdout or audience-holdout tests quantify lift with scientific rigor. Partners like LiftLab run matched-market experiments and agile mix models, producing estimated lift, CPA/ROAS, and diminishing returns curves by channel.

App Attribution & Mobile Measurement

Tatari integrates with AppsFlyer, Branch, Adjust, Kochava, and Singular to surface app installs and in-app purchases driven by TV in the Tatari dashboard.

Amazon and Retail Sales Measurement

• Amazon Sales Analysis: Tatari can analyze Amazon sales signals alongside airtime and website response to estimate TV-driven retail impact. The methodology and data requirements depend on the client’s available Amazon data and product setup..

• Geo-Based Retail Testing: For eligible brick-and-mortar clients, Tatari can run geo-based incrementality tests using DMA-level sales data to measure how TV influences in-store performance. This requires sufficient sales data at the right geographic and time granularity; no PII is required.

• Retail Modeled Conversions: Tatari can use statistical modeling to estimate retail outcomes where direct signals are sparse. These analyses are data-dependent and should be scoped with the Tatari team.

Best Practices for TV Measurement

Answer the question before choosing the tool

Start with the business question: are you trying to prove TV works, optimize within TV, or justify TV spend relative to other channels? The answer determines which measurement methodology to prioritize.

Use incrementality as your primary source of truth

View-through attribution is useful for optimization — it tells you which placements and creatives are performing relatively better. But for the absolute question of whether TV is driving growth, incremental lift is the more defensible answer. Use view-through to optimize; use incrementality to justify.

Understand your response window and DragFactor before reporting

Before reporting ROAS or CPV, ensure your DragFactor is calibrated for your category. The response window is fixed by Tatari’s methodology — marketers do not configure it. But understanding whether your category responds immediately or over days is essential context for interpreting results. A direct-response brand and a brand-awareness campaign will have very different response profiles, and DragFactor captures that difference.

Implement signal collection correctly before spending at scale

The quality of your measurement is only as good as your signal collection. Whether you use TTM or Vault S2S, test and verify implementation before scaling spend. A misconfigured pixel on a high-spend campaign produces confident-looking wrong numbers.

Blend streaming and linear intentionally

Use CTV for deterministic measurement and targeting precision; use linear for reach and scale. Measure them together in a convergent framework rather than evaluating each channel in isolation.

Don’t rely on a single source of truth

The most sophisticated TV advertisers triangulate. Avoid declaring a 'winner' methodology without cross-validating. MMM, attribution, and holdout tests each have blind spots; if two methodologies disagree, investigate the discrepancy before acting on either. Tatari’s attribution for campaign-level optimization, MMM for long-term channel-level planning, and DOE for periodic causal validation.

Measure halo effects alongside direct conversions

Monitor branded search lift, direct traffic, and Amazon search trends correlated with airtimes. TV’s true impact is consistently larger than direct attribution captures, and quantifying halo effects makes the business case for TV investment more complete and more defensible.

Treat campaigns as continuous experiments

Set hypotheses (this creative will drive a faster response than that one), run the test, measure the outcome, and iterate. The brands getting the most from TV are the ones that have built a learning agenda, not just a buying calendar.

Glossary of TV Measurement Terms

Attribution: The process of connecting an ad exposure to a downstream conversion. TV attribution uses IP matching or device graph matching to link TV airings to website visits, app installs, or purchases.

Attribution Window: How long after a session start a conversion event (purchase, sign-up, install) can be attributed to that session. Distinct from the response window and from DragFactor. Governs whether a conversion that happens hours or days after a viewer first engages is still credited to the originating TV exposure.

Baseline: The predicted level of conversions that would have occurred without any TV advertising. Used to isolate the incremental lift caused by TV.

CAC (Customer Acquisition Cost): Total spend divided by new customers acquired. The primary efficiency metric for subscription and DTC businesses.

Clean Room: A privacy-preserving environment where two parties can match data without either party exposing raw PII to the other. Vault is Tatari’s clean room partner for TV measurement.

CPM (Cost Per Mille): The cost of 1,000 ad impressions. The standard buying currency for CTV; used alongside GRPs in linear TV planning.

CPI (Cost per Install): The cost of driving one app install attributable to TV advertising.

CPV (Cost per Visit): The cost of driving one incremental website visit via TV.

CTV (Connected TV): Internet-connected television screens — smart TVs, Roku, Amazon Fire TV, Apple TV, gaming consoles — and the ads delivered within streaming content on those devices.

DragFactor: Tatari’s proprietary multiplier that quantifies the relationship between immediate and delayed TV response. Expressed as: Immediate Response × DragFactor = Total Response. Operates beyond the fixed response window to estimate total response over 30 days, enabling next-day measurement. Distinct from the attribution window, which governs how long after a session start a conversion can be credited to that session. New clients start with a category or default 2.2x DragFactor; custom DragFactors are calculated after sufficient spend accrues (approximately 4 months/$1M on linear; 6 weeks/$250K on streaming).

GRP (Gross Rating Point): The traditional currency of linear TV advertising. Calculated as Reach % × Frequency. Measures exposure volume, not outcomes.

Halo Effect: Downstream impact of TV advertising on other channels — including branded search lift, direct traffic, and social engagement — that does not appear in direct conversion attribution. Can be direct (affecting exposed users) or indirect (affecting platform dynamics and broader network effects).

Incremental Lift: The net-new conversions caused by TV advertising above the baseline that would have occurred organically. The most defensible measure of TV’s true contribution.

Incrementality Testing: A controlled experiment (geo holdout, audience holdout) that quantifies incremental lift by comparing an exposed group to an unexposed control group.

iROAS (Incremental ROAS): Return on ad spend calculated on truly incremental revenue only — what you’d lose if you turned TV off. More conservative and more actionable than attributed ROAS.

Linear TV: Traditional broadcast and cable television delivered on a fixed schedule via satellite or cable networks.

MMM (Marketing Mix Modeling): A statistical technique using multivariate regression to estimate the contribution of each marketing channel — including TV — to sales over time. Operates on aggregate data; not reliant on user-level matching.

MTA (Multi-Touch Attribution): A user-level attribution methodology that models the full customer journey across digital and TV touchpoints to assign credit to each.

OTT (Over-the-Top): Streaming video delivered over the internet independently of traditional pay TV. Includes content viewed on any screen; CTV is the TV-screen subset of OTT.

ROAS (Return on Ad Spend): Revenue attributable to advertising divided by spend. For TV, typically calculated from revenue or gross margin attributed to TV exposures.

Response Window: Tatari’s fixed window for measuring immediate response after a TV airing — the spike in site visitors observed in the minutes directly following a spot. Fixed by Tatari’s methodology; not selected by the marketer. Distinct from DragFactor (which estimates total delayed response beyond this window) and from the attribution window (which governs conversion crediting).

Response Curve: The distribution of conversions over time following a TV airing. Shape varies by category, creative, and daypart. Captured by DragFactor.

Server-to-Server (S2S) Tracking: A method of sending conversion events directly from the advertiser’s server to the measurement platform, bypassing browser-based tracking limitations. Tatari’s implementation is called Vault.

TTM (Tatari Tag Manager): Tatari’s browser-based pixel for tracking page visits, app installs, or purchases. The simplest implementation path for signal collection.

View-Through Attribution: Attribution that credits a TV exposure with a downstream conversion, regardless of whether the viewer took any direct action in response. Tatari’s View-Through model uses a proprietary device graph to improve accuracy over IP-only matching.

FAQs

1. How is TV attribution different from digital attribution? Digital attribution uses click-through tracking — every action leaves a direct record. TV attribution is probabilistic: viewers don't click, so measurement relies on matching IP addresses or device graphs to downstream conversions, then subtracting the organic baseline to isolate TV's contribution.

2. What's the difference between view-through attribution and incrementality? View-through attribution credits a TV exposure with a downstream conversion. Incrementality measures only the net-new conversions caused by TV — what you'd lose if you turned TV off. Incrementality is more conservative and more defensible; view-through is more useful for optimization.

3. What data do I need to get started with TV measurement on Tatari? At minimum: a Tatari pixel (TTM) or server-to-server Vault integration for conversion tracking, and sufficient website or app traffic to generate a measurable signal. Your Tatari team will help assess whether your traffic volume supports view-through attribution or whether Incremental Lift is the right starting point.

4. Can I measure TV if I use Shopify? Yes. The Tatari pixel (TTM) integrates natively with Shopify and can be deployed in a few clicks. No engineering resources required for basic implementation.

5. What's the difference between CTV and OTT? OTT (over-the-top) is any streaming video delivered over the internet, viewable on any screen. CTV (Connected TV) is the subset of OTT watched on a television screen. For TV advertising purposes, CTV is the relevant term.

6. Does Linear TV measurement work without cookies or user-level data? Yes.  Tatari's linear measurement uses a probabilistic, aggregate-level baseline-lift methodology that does not require IP matching or device graph matching — it measures traffic spikes at the population level. For streaming, impression-level signals (IP, device ID, timestamp) are used but processed through Tatari's privacy clean room (Vault), which separates PII before data reaches Tatari's measurement environment.

7. How do I know which attribution model to use? Start with Tatari View-Through if you have sufficient traffic and household-matched signal. Use Digital View-Through if you need apples-to-apples comparisons with Meta or Google. Use Incremental Lift if traffic volume is lower or if you need the most defensible causal answer to "is TV working?"

Ready to measure TV performance? Contact Tatari to discuss your measurement setup, review implementation options (pixel vs. server-to-server), and determine which attribution model fits your campaign objectives.