A paid media budget that includes AI visibility starts from a revenue goal and a break-even ROAS, then splits spend across search, social and programmatic by job. AI visibility is funded as a separate line because it is earned through accurate pages and independent coverage, not bought. McKinsey’s October 2025 survey found 44 percent of AI search users call it their primary source of insight.
Last updated November 1, 2026.
On this page:
- Set the budget from a revenue goal
- Split spend across channels
- Why AI visibility belongs
- What the AI line covers
- Improve ROAS first
- Measure paid, PR and AI together
- Review cadence
- Frequently asked questions
How do you set a paid media budget from a revenue goal?
A paid media budget is set by working backward from the revenue paid media must produce, using two numbers: the target return on ad spend (ROAS) and the break-even ROAS. ROAS is the revenue attributed to ads divided by the ad spend that produced it. Break-even ROAS is one divided by gross margin, because each ad dollar must return enough revenue to cover the cost of the product sold.
A brand with a 40 percent gross margin breaks even at a ROAS of 2.5. Any return above 2.5 is profit on the first purchase. Brands with strong repeat purchases can accept a lower first-order ROAS because later orders cover the gap. The paid media glossary entry defines the channels the budget covers: search, social, display, programmatic, video and connected TV.
Separate the line items from the start. Media spend buys impressions and clicks. Creative production, tracking tools and management time make that spend work, and a budget that leaves them out usually produces weak ads and unreliable ROAS figures.
Worked example of setting a paid media budget
A brand with a 40 percent gross margin needs $1,200,000 in paid-attributed revenue this quarter and targets a ROAS of 4. Media spend is $1,200,000 divided by 4, which is $300,000. Break-even ROAS for the same brand is 2.5, so a ROAS of 4 leaves room for profit.
| Bucket | Job | Share | Spend |
|---|---|---|---|
| Paid search | Capture existing demand | 35% | $105,000 |
| Paid social | Create demand | 30% | $90,000 |
| Programmatic and CTV | Extend reach | 15% | $45,000 |
| Retargeting | Convert warm audiences | 10% | $30,000 |
| Test budget | New channels, formats and messages | 10% | $30,000 |
How should spend be split across search, social and programmatic?
Spend should be split by the job each channel does, not by habit or historical allocations. Paid search captures people already looking for the product. Paid social creates demand by reaching people before they search. Programmatic buys display, video, audio and connected TV placements across many sites and apps, which extends reach beyond the large platforms. Media planning and buying services set these shares channel by channel; the media planning and buying practice covers that work.
| Channel | Job | Strongest when | Main measure |
|---|---|---|---|
| Paid search | Capture demand | Buyers already search the category | ROAS, cost per acquisition |
| Paid social | Create demand | The brand needs new audiences | Blended ROAS, assisted conversions |
| Programmatic and CTV | Extend reach | Creative is strong and scale matters | Reach, view-through conversions, lift |
| Retargeting | Convert warm audiences | The site already has traffic | ROAS |
The best-known evidence on balance comes from the Institute of Practitioners in Advertising. According to the IPA’s review of Binet and Field’s research, their 2013 report found an optimum budget ratio of 60:40 between brand building and activation, and their 2018 follow-up found 62:38. Reach-oriented channels such as programmatic video, connected TV and paid social usually do the brand-building work, while paid search and retargeting usually do the activation work. Treat the ratio as a starting reference and let the brand’s own results move it.
Keep a protected test budget. Without one, every dollar chases the current winner and the next channel never gets a fair trial.
Why does AI visibility belong in the paid media budget?
AI visibility belongs in the paid media budget because a growing share of buyers now research purchases inside AI platforms, and the sources those platforms draw on are mostly pages that ad spend does not produce. McKinsey’s October 2025 AI search report, based on an August 2025 survey of 1,927 U.S. consumers, found that about half of consumers use AI-powered search and that 44 percent of its users call it their primary and preferred source of insight, ahead of traditional search at 31 percent and retailer or brand websites at 9 percent.
The same report explains why a strong ad program does not guarantee a presence in AI answers. McKinsey’s analysis found that a brand’s own websites make up only 5 to 10 percent of the sources AI search references, and that in consumer packaged goods and financial services more than 65 percent of sources are publishers, user-generated content and affiliate sites. McKinsey also reported that only 16 percent of the brands it surveyed systematically track AI search performance.
Two budget consequences follow. First, buyers who ask an AI platform for recommendations may never see an ad, so a brand absent from the answer loses consideration before any click is available to buy. Second, the work that changes the answer is accurate owned content and independent coverage, which belongs on its own line. McKinsey expects AI platforms to begin weighing paid ad formats in the future, but the sources behind answers today are mostly owned, third-party and community content. A category error is one concrete way to lose a recommendation, as the analysis of AI misclassification documents.
What does an AI visibility line item cover?
An AI visibility line item covers four kinds of work: a baseline check of how AI platforms describe the brand, accurate owned pages, independent coverage and ongoing measurement. The AI visibility glossary entry defines the term, and the table shows what each part buys and how to check that it worked.
| Line item | What it buys | How to check it worked |
|---|---|---|
| Baseline check | A record of how AI platforms answer a fixed set of buyer questions about the category | The same questions re-run each month |
| Owned page accuracy | Clear category wording, product facts and consistent descriptions on the brand’s own pages | AI descriptions match the approved wording |
| Independent coverage | Articles, reviews and community mentions on the publisher, affiliate and user-generated sources McKinsey found AI search relies on | Count of new independent pages that name the brand |
| Measurement | Tracking of AI mentions and AI referral traffic | Share of answers that mention the brand, plus referral sessions |
Because this work is earned and owned rather than bought, a practical approach is to fund it from the brand-building share of the budget or as a separate line, and to judge it on share of answers and branded demand instead of ROAS. Size the line after the baseline check, since the gaps it finds decide how much owned and independent work is needed.
How do you improve ROAS before adding spend?
ROAS improves fastest by fixing what surrounds the ad, because most weak returns come from broken tracking, thin creative testing or a landing page that does not convert. Work through these steps in order:
- Confirm conversion tracking, revenue values and UTM tags before reading any ROAS number.
- Add negative keywords in search, exclude weak placements in programmatic and pause ads whose cost per acquisition sits above break-even.
- Test one creative variable at a time, such as hook, offer or format, and keep a log of winners.
- Improve the landing page conversion rate, which lifts ROAS in proportion without extra media spend.
- Move budget toward the channel where the next dollar returns the most.
The last step matters most. A channel’s average ROAS flatters it because early spend reaches the best audiences. Marginal ROAS, the return on the next dollar, falls as spend grows. When marginal ROAS drops below break-even, stop adding spend to that channel.
How do you measure paid media, PR and AI visibility together?
Paid media, PR and AI visibility are measured together by scoring all three against the same business outcomes, not each channel’s own scorecard. Paid teams watch ROAS, PR teams watch coverage and AI visibility has its own counts, which hides how the three feed each other. The guides to measuring PR effectiveness and integrated marketing across PR, paid media and content cover the PR and planning sides.
| Measure | Where to read it | What it shows |
|---|---|---|
| Branded search volume | Search Console, Google Trends | Whether coverage and ads raise demand for the brand name |
| Conversion rate on paid traffic during coverage windows | Ad platform, analytics | Whether press coverage lifts paid performance |
| Assisted conversions | Analytics attribution reports | How earned and paid touchpoints combine before a sale |
| Referral traffic from AI platforms | Analytics referral sources | Visits that AI answers send to the site |
| Share of AI answers that mention the brand | A fixed set of buyer questions, re-run on a schedule | Whether the brand appears where buyers research |
| Incremental lift | Holdout or geo tests | What paid media caused beyond what would have happened anyway |
A holdout test is the cleanest check on paid results: withhold ads from a matched audience or region and compare it with the exposed group. For wider data on how AI search is changing discovery, see the State of AI Search 2026 research.
How often should the paid media budget be reviewed?
The paid media budget should be checked weekly for problems, rebalanced monthly and reset quarterly. Weekly checks catch tracking breaks and runaway spend. Monthly rebalancing moves money toward channels with the best marginal return. Quarterly resets revisit the revenue goal, the margin assumption and the split between brand building and activation. Written rules keep the review from turning into a debate:
- Move budget out of a channel whose marginal ROAS stays below break-even for two review cycles.
- Promote a test to a funded channel only after it beats break-even on a meaningful number of conversions.
- Re-run the AI answer baseline each month and compare it with the previous month before changing the AI visibility line.
- Record every change and its reason so the next quarter starts from evidence.
Frequently asked questions
What percentage of revenue should go to paid media?
No single percentage of revenue applies to every brand. Set the budget backward from the revenue paid media must produce, the target ROAS and the gross margin. A brand with thin margins needs a higher ROAS to break even, so it often spends less per revenue dollar than a high-margin brand.
Does paid media change what AI platforms say about a brand?
Paid media does not directly change what AI platforms say about a brand, based on findings from McKinsey. Its analysis shows AI search answers draw mostly on publishers, user-generated content, affiliate sites and a small share of brand-owned pages, and it expects paid formats to enter AI answers later. Check the answers directly instead of assuming ad spend changes them.
What is a good ROAS?
A good ROAS is any figure comfortably above break-even ROAS, which equals one divided by gross margin. At a 40 percent margin, break-even is 2.5, so a ROAS of 4 leaves room for profit. Brands with strong repeat purchases can accept a lower first-order figure.
Where should AI visibility sit in the budget?
AI visibility should sit in its own line, funded from the brand-building share of the budget, because the work is earned and owned rather than bought. Judge the line on the share of AI answers that mention the brand and on branded demand, not on ROAS.
How do you measure incrementality?
Incrementality is measured by running a holdout test. Withhold ads from a matched audience or region and compare its results with the exposed group. The gap is the revenue the ads caused. Platform lift studies and geo experiments both support this method.
