Reference

The sports publisher monetization stack

A visual map of how sports publishers turn attention into revenue — and where the system usually breaks. Start with the symptom you see. Trace the mechanism underneath it. Find the revenue line, infrastructure layer, or workflow where the fix belongs.

Built from operating experience across sports data, consumer product, video, commerce, and fan-revenue systems. Evidence markers link to supporting artifacts. The map has three layers, and each answers a different question.

Layer 1
Revenue lines
Where money enters the business.
Layer 2
Infrastructure
What has to be true for a line to earn anything. Where monetization is won or lost.
Layer 3
AI capabilities across the stack
Not a revenue line. It acts on the other two.
Layer 1
Revenue lines
Advertising
Performance & betting
Audience revenue
Commerce
Content & data
↑
makes them earn
Layer 2
Infrastructure
Identity & data
Ad stack
Yield & experience
Demand for the page
Sports constraints
Operations & delivery
Measurement
Layer 3
AI capabilities across the stack
Not a revenue line. It acts on both of the layers beside it, plus one genuinely new surface.
←
lifts what the lines earn
Yield intelligence
Content operations
Subscription intelligence
Conversational surfaces
Agents & workflows
Markers are evidence, not credentials. They say what kind of contact the work behind this map has had with each node — and where a claim stops.
Operated
Direct product ownership of the line or the surface.
Study node
Everything unmarked. Included because the stack cannot be understood without it — no personal-work claim attached.
Contributed
Moved the number through product without owning the function.

How to use this map

01
Find a symptom your business recognises.
02
Read the likely mechanism beneath it.
03
Locate the layer where the evidence and the intervention belong.
04
Follow an evidence marker to see how it has been applied in practice.
A symptom is not a diagnosis. Declining ad revenue can originate in traffic quality, identity loss, page performance, inventory design, demand mix, measurement, or commercial operations — and the wrong one gets blamed most of the time.

Where it usually hurts

Twenty-two failures that recur across sports publishers of every size. Each one is stated as the symptom an owner would describe, then the mechanism actually causing it, then where in the stack the fix has to happen. The mechanism is the part worth knowing.

Traffic

Search traffic collapsed and nothing has replaced it

Referrals are down roughly a third globally and closer to 38 percent in the US, with publishers expecting another 43 percent over three years. The number is not the problem. The problem is that every staffing plan, inventory forecast and revenue model underneath it assumed the old one, and none of them have been rebuilt.
Layer 2 · Demand for the page

The content is read but never visited

An answer quotes the work and the reader never arrives. The dashboard records that as a decline because it only knows how to count clicks, so the organization responds by chasing traffic that is not coming back rather than measuring the reach it is actually getting.
Layer 2 · Demand for the page

Seasonality whiplash

The audience arrives in spikes tied to a schedule and disappears between them. Commitments — staff, inventory, guaranteed ad deals — are annual. That mismatch is structural rather than a forecasting error, and it is why sports publishers over-hire in September.
Layer 2 · Operations and delivery

Ad revenue

The density spiral

CPMs fall, so units are added to hold revenue flat. The units cost load performance, which costs sessions and ranking, which requires more units. Every individual step is locally rational and the direction only goes one way.
1
CPMs fall
the open auction reprices
→
2
Units get added
to hold revenue flat
→
3
The page slows
LCP and CLS degrade
→
4
Sessions drop
and so does ranking
↺ Back to step 1. Every step is locally rational — the direction only goes one way.
Layer 2 · Yield and experience

Brand safety strips the best-performing content

Blocklists written for news treat ordinary sports language as unsafe: shooting, attack, penalty, beat, kill. Demand quietly leaves the pages that perform best, and the report shows a CPM decline with no cause attached to it.
Layer 2 · The ad stack

Direct sales costs more than the deals return

A direct team is a fixed cost sitting against deals that keep shrinking. Below a certain scale the cost of selling exceeds the premium direct earns over programmatic, and the honest answer is to stop selling that way.
Layer 2 · Measurement and commercial

Performance and betting

The biggest line is the least managed

Affiliate carries most of the revenue at emerging scale and gets reviewed once a year. Offers expire, operators exit states, links lose their tracking parameters. This revenue does not fail loudly. It just stops, and the monthly total absorbs it.
Layer 1 · Performance and betting

You are a variable cost in someone else's budget

Sportsbook affiliate spend is a marketing line for the operator. When they hit a customer-acquisition target or need a profitable quarter, the line they cut is yours, and the decision has nothing to do with your performance.
Layer 1 · Performance and betting

Regulatory volatility

A state changes its rules and a revenue line ends that week. The real exposure is concentration by operator and by state, and most publishers have never measured either one.
Layer 1 · Performance and betting

Subscriptions

They subscribe for the run and churn in February

Sports subscription demand is event-shaped. Annual cohort retention describes a business that does not behave like this one, so the renewal forecast is wrong in a direction you could predict and price for, and almost nobody does.
Layer 1 · Audience revenue

Most churn was never a decision

A quarter to a third of churn is typically failed payments rather than cancellations. It is the cheapest churn to recover and the least likely to be owned by anyone, because it sits between finance and product and belongs to neither.
Layer 2 · Operations and delivery

The paywall is placed by rule, not by reader

A fixed meter blocks the reader who would have paid and waves through the one who never will. The rule is identical for everybody because modelling per reader was harder than picking a number, and the number has not been revisited since.
Layer 1 · Audience revenue

Data and identity

Third-party data gone, first-party never built

The audience was known through someone else's cookie. That went away and nothing replaced it, so targeting, measurement and audience sales all degrade at the same time and each looks like a separate problem.
Layer 2 · Identity and data

Measurement

Nobody can say which line actually makes money

Revenue by line is reported. Cost to run each line is not, so contribution margin is unknown and investment follows the biggest number rather than the best one. The largest line is often the least profitable.
Layer 2 · Measurement

Three systems claim the same reader

One session touches an ad, an affiliate link and a paywall. Three systems each take credit, the totals do not reconcile, and the practical result is that nobody trusts any of them enough to act.
Layer 2 · Measurement

The AI era

The archive is training data and nobody paid for it

It has already been used. The current shift is toward pay-per-use and content marketplaces rather than one-to-one deals, which matters most because the majority of publishers will never be offered a deal at all and need a different route to getting paid.
Layer 1 · Content and data as product

The answer cites somebody else

Being the source an assistant quotes is now a distribution question. The levers — structured data, citation tracking, crawler policy — sit with three different teams, none of whom think of it as their revenue problem.
Layer 2 · Demand for the page

AI gets pointed at the wrong problem

It is aimed at content generation, where the risk and the reputational cost live, instead of at yield, metadata and revenue operations, where it is low-risk and compounds quietly. The visible use is the worst one.
Layer 3 · AI capabilities

Agentic commerce is aimed straight at affiliate

The largest line at emerging scale depends entirely on a human clicking a tracked link. Nobody has modelled what replaces attribution when the buyer is an agent, and the exposure sits under the revenue line least able to absorb it.
Layer 3 · AI capabilities

Organizational

Every team optimizes correctly and the total moves sideways

Ads wants density, subscriptions wants the paywall, editorial wants reach, commerce wants the click. Each is right within its own remit. Nobody owns the number they collectively produce, so the trades between them never get made deliberately.
Layer 2 · Measurement and commercial

The talent gap

People who understand sports audiences generally cannot build. People who can build generally do not know how sports makes money. Roles needing both are growing faster than the supply of people who have both, which is why this map is harder to staff than it looks.
All three layers
The map as a PDF

Take the twenty-two with you

Every revenue line, the infrastructure underneath it, and all twenty-two failures with the mechanism behind each one. Eleven pages you can mark up and put in front of a room.

Thanks — the PDF is on its way to your inbox.
Something went wrong. Email channing7@gmail.com and I will send it over directly.
One email with the link. No sequence, no list.
Layer 1

Revenue lines

Where money enters. Sports publishers run more of these at once than almost any other vertical.

Advertising

Display
Open programmatic, direct-sold, PMPs. Under the most structural pressure.
Contributed · USA Today Sports
Video
Instream, outstream, CTV and FAST. The one ad format still growing.
Operated · Fox Sports
Native and branded content
Studio-produced, sold on relationship rather than auction.
Operated · Fox Sports
Sponsorship and franchise
Presented-by on a recurring column, podcast or live show.
Newsletter inventory
Owned audience, no auction. CPMs hold when display does not.
Podcast and audio
Host-read versus programmatic insertion. Video podcast is where investment moved.
Off-platform
YouTube, social, syndication. Revenue that never touches your pageview.
Library: X / Meta / TikTok
Programmatic guaranteed and preferred deals
Reserved, fixed-price programmatic. A different sell from PMPs and often confused with them.
FAST channel ownership
Running the channel rather than selling inventory into someone else's.

Performance and betting

Sportsbook affiliate
CPA, revenue share or hybrid. Around 70 percent of emerging sports publisher revenue, and almost nobody audits the links.
Retail, gear and ticketing affiliate
The non-betting half. Survives regulatory shifts that hit sportsbook affiliate.
BetTech integrations
Odds widgets, embedded betslips, free-to-play. Outperforms display per session, and is a different mechanic from affiliate.
Regulatory exposure
State-by-state legality, age gating, advertising standards. The risk that makes this group volatile.
Commerce editorial
Best-of and review formats built for transaction intent. A discipline, not just a link placement.
Lead generation
Non-affiliate lead sale: app signups, ticket alerts, insurance. Paid per qualified lead rather than per conversion.
Networks versus direct deals
Going through an affiliate network or contracting the operator directly. Changes rate, terms and who owns the data.

Audience revenue

Subscriptions
Hard, metered, dynamic and freemium paywalls. Propensity modelling, price testing, bundling, churn.
Contributed · PFF
Memberships and community
Access and identity rather than content access. Different retention curve.
Micropayments and day passes
Single-game access. Fits sports demand spikes better than a monthly plan.
Fantasy products
Tools that act as subscription driver, affiliate funnel and engagement loop at once. The rare product that touches three revenue lines.
Operated · PFF
Free-to-play, pick-ems and brackets
Contest mechanics as an engagement and data-capture engine, sponsorable on top.
Library: Sports Trivia Generator
Operated · Sports Lab Arcade
Sweepstakes and sponsored contests
Prize-led acquisition. Cheap list growth, with its own regulatory surface.

Commerce

Owned commerce
Merch and direct purchase against an audience you already have.
Operated · Walt Disney World
Marketplace and dropship
Inventory-light commerce with no working capital tied up.
Ticketing and resale
Highest-intent transaction in sports, closest to the event itself.
Operated · Ticketmaster
Events
Watch parties, live shows, conferences. Went from 8 to 23 percent of publisher revenue sources in a year.

Content and data as product

First-party data and audience extension
Selling reach against your own logged-in and modelled audience, off your own site.
Syndication and licensing
The old licensing business. Still paying, still underrated.
AI licensing
The shift is from lump-sum deals to pay-per-use royalties and content marketplaces. Most publishers will never be offered a one-to-one deal.
Bot and crawl monetization
Pay-per-crawl at the CDN layer. A revenue line that did not exist eighteen months ago.
B2B data and analytics products
Selling to teams, agencies and books rather than to fans. Highest-margin line a sports publisher can build.
Services and agency arm
Selling the newsroom's own capability to brands. Revenue that does not depend on traffic at all.
Courses and education products
One-off or cohort-based, priced against outcomes rather than access.
Layer 2

Infrastructure

What has to be true for any line above to earn anything. This is where monetization is won or lost, and where it becomes clear whether someone has run it or read about it.

Identity and data

First-party capture
Where in the journey you ask, what you ask for, and what it costs in session. The one asset that does not move when Google does.
Operated · Sports Lab Arcade
Alternative IDs
UID2, ID5, and contextual as the fallback. What actually fills the addressability gap.
Clean rooms
How advertisers match audiences without either side handing over the file.
Consent and privacy
TCF and CCPA, and the revenue cost of a badly built consent flow, which is usually larger than anyone measures.
The ad server
GAM and its alternatives. The layer under header bidding that decides what actually serves.
Library: Google Ad Manager
Video ad serving
VAST, server-side ad insertion, player configuration. Video revenue is gated by this and it is specialized work.
Invalid traffic and ad fraud
A direct revenue leak and a reputational one. Usually discovered late.
Brand safety and keyword blocking
The classic sports revenue killer. Blocklists built for news strip demand from ordinary sports language.
Content taxonomy and metadata
IAB categories and internal taxonomy. What contextual targeting is actually built on.

The ad stack

Header bidding and Prebid
Wrapper configuration, timeout tuning, demand partner mix.
Floors and price rules
Auction dynamics, bid shading, unfilled and passback behaviour.
Curation and curated PMPs
Packaging your own supply into deals instead of leaving it in the open auction.
Supply path and hygiene
SPO, ads.txt and sellers.json, and staying clear of made-for-advertising classification.

Yield and experience

RPM and RPS as the unit
Revenue per thousand and per session. The numbers that make different lines comparable.
Contributed · through product
Ad density against Core Web Vitals
Every added unit trades against load performance, sessions and rankings. Usually an unmeasured net negative.
Viewability and refresh policy
What lifts CPM, and what quietly burns advertiser trust.

Demand for the page itself

AI search visibility
Structured data, citation tracking, being the source an assistant quotes. With referral traffic down a third, this is a monetization problem now.
Zero-click analytics
Measuring reach you never get a session for. The old dashboard reports it as a decline.
Owned channel growth
Newsletter, app, notifications. Distribution nobody can reprice on you.

Sports-specific constraints

Data rights and feed costs
Sportradar, Genius and the rest. What the feed costs decides what you can build before you build anything.
Library: Sportradar / SportsDataIOESPN endpoints
Highlight, logo and mark rights
What you are allowed to show. Shapes the product more than most people outside sports expect.
Player likeness and NIL
Who you can name, picture and build a product around, at pro and college level.
League and team partner status
Official designation changes what you can sell and at what price. A commercial asset, not a badge.

Operations and delivery

Billing, dunning and recovery
Failed payments are the largest silent churn driver in most subscription businesses, and the cheapest to fix.
Publishing workflow and speed to live
Sports is a latency business. Minutes between whistle and published decide whether the traffic arrives at all.
Library: Webflow / WordPress embed
App versus web economics
Store fees, push as an owned channel, and why the same subscriber is worth different amounts in each.
Data warehouse and BI
Where the numbers live, and whether anyone outside the analytics team can reach them.

Measurement

Testing as infrastructure
A permanent capability rather than a per-launch activity.
Operated · across roles
Incrementality
Whether a line added revenue or moved it from somewhere else. Most publisher reporting cannot tell.
LTV by acquisition source
Cohort retention, and which channels produce subscribers who stay.
Cross-line attribution
One reader touching an ad, an affiliate link and a paywall in a single session. Almost nobody models this well.
Rate card and pricing strategy
What direct inventory is worth and how it is defended in a negotiation.
Sell-through and inventory forecasting
How much you have, how much you will sell, and what the unsold remainder is worth.
Contribution margin by line
Which lines make money after their cost to run. Almost nobody computes this, and it changes what you would do next.
Layer 3

AI capabilities across the stack

Agents, LLMs and agentic workflows are not revenue lines. They act on the two layers above, plus one genuinely new surface.

Acting on existing lines

Grounded LLM surfaces
Assistants over your own data that cite sources and refuse when the answer is not there.
Try it live →Library: Grounded Sports Q&A
Operated · Channing & Co
Yield intelligence
Dynamic floors, bid shading, demand mix decided continuously rather than quarterly.
Library: Yield & Pricing Assistant
Content operations
Metadata, tagging, summarization, translation. Unglamorous, and it directly increases sellable inventory.
Library: Monetization CopilotRecap-to-Social
Subscription intelligence
Churn propensity, and a paywall that moves per reader instead of per rule.
Commerce intelligence
Recommendation and dynamic merchandising against event and team context.
Personalization and chat-style homepages
The feed-style homepage is on its way out. What replaces it is assembled per reader.
RAG over your own archive
A product in its own right, and the groundwork for licensing the archive later.
Dynamic and synthetic creative
Ad creative generated against team, moment and context rather than trafficked flat.

Agents and workflows

Agentic revenue operations
Reconciliation, anomaly detection, reporting. Where most real publisher value lands first.
Operated · Channing & Co
Agents that take actions
Planning loops, tool design, and the judgment about where a model belongs in a pipeline and where it does not.
Operated · Channing & Co
Conversational surfaces as inventory
The idea almost nobody has followed through. An assistant is an ad and commerce surface. What is the unit, and what does it cost?
Agentic media trading
When the buy side is agents, auction dynamics change.
Agentic commerce
When the shopper is an agent, what happens to affiliate attribution and the merchandising layer entirely?
Bot management and crawler policy
Which crawlers to allow, block or charge. Now a revenue decision rather than an infrastructure one.
AI in ad operations
Campaign setup, creative QA, trafficking. Unglamorous, and it compounds.
AI content policy and disclosure
What gets labelled, what search rewards, and what readers forgive.
Library: Compliance Engine

Five questions nobody has answered

Every model on this page assumes a human arriving on a page. That assumption is the one currently breaking.
01

What is a publisher worth when the reader is an agent?

Remove the page and the whole stack needs rebuilding. Nobody has a working answer, and every publisher is about to need one.
02

What is a crawl actually worth?

Pay-per-use pricing exists now at the CDN layer. There is no benchmark, no rate card, and no method for valuing one corpus against another.
03

What is the ad unit inside an assistant?

Conversational surfaces have no established format, no viewability standard and no disclosure norm. Whoever defines the unit shapes the category.
04

How does affiliate survive agentic commerce?

Affiliate carries most emerging sports publisher revenue and depends entirely on a human clicking a tracked link. Agents do not click.
Today
Reader
arrives on the page
→
Clicks a tracked link
parameters intact
→
Attribution fires
the click is the event
→
Publisher paid
CPA or revenue share
When the buyer is an agent
Agent
reads the same page
→
Completes the purchase
on the reader's behalf
→
No click, no parameters
nothing to attribute
→
Nothing fires
the line goes quiet
05

What happens to live moments?

Sports monetization runs on simultaneity, everyone wanting the same thing in the same ninety seconds. That is the one property an assistant cannot flatten, and the best argument that sports has more defensible ground than general news.
Channing & Co
channing7@gmail.comLinkedIn
© 2026 Channing & Company