Part I · Chapter 4

AdTech Economics and Business Models

Status: reviewed · Prerequisites: What AdTech Is and Why It Exists, AdTech Ecosystem Participants, Core Advertising Metrics

AdTech Economics and Business Models

The Core Idea: the price follows the contract, not the label in the dashboard

CPM1, CPC2, or CPA3 in an interface does not yet explain what is being paid for. The same label can be a contractual price, a bid4, an optimization target5, or a post-event average. The monetary obligation is created by the contractual pricing basis6 together with the definition of a billable event7 — a qualifying event after which an amount can be charged. Even an auction win8 does not have to be billable in itself (in programmatic9 buying, chapters 15–18).

To read the economics of a deal, break it down into six independent layers:

Layer Main question Example
Contractual pricing On which basis was payment agreed? $10 CPM per thousand qualifying impressions10
Billing Which event creates the charge, and for what amount? Render or another billing signal per contract policy
Optimization Which outcome changes the system’s future decisions? target CPA on a subscription
Participant ledger11 For whom is the amount spend12, cost, payout, or retained amount13? $100 advertiser14 spend; $80 media cost15 of the intermediary16
Accounting presentation What does a particular company recognize as revenue17 and cost? Gross or net presentation
Settlement18 When did the accrual become an invoice, and the obligation cash? Month-end invoice, paid later

The $100 and $80 come from the chapter’s running example: a managed network19 (an ad network that buys media itself and sells the advertiser a finished result) resells for $100 media it bought for $80; the full walkthrough is in the section “The path of the instructional $100”.

Pricing basis is the contractual rule for calculating the price: per impression, click20, action, share of proceeds21, a period of access to software, or another agreed object. Billable amount is the amount calculated under this rule after counts, rates, validation, and contractual adjustments. It is not necessarily invoiced yet, recognized as revenue, or paid in cash.

This decomposition explains the whole chapter: the parties choose not a pretty metric but the boundary of an observable event, a price, and a risk allocation; then each side records the result in its own ledger.

CPM, CPC, and outcome pricing: who bears the risk until the next event

The funnel from an ad opportunity22 can stop after any step:

opportunity → qualifying impression → qualifying click → install/lead/action → business value

The deeper the billable event, the longer the payee23 — the party awaiting payment — finances delivery without a guaranteed charge. But the risk does not “migrate” wholesale: after the paid event, the buyer still risks the quality of the next outcome.

Pricing model Contractual billable event Payee risk before the charge Advertiser risk after the charge What must be defined
CPM A thousand qualifying impressions Receive and confirm the chosen impression event Click, conversion24, and subsequent value may never occur served25, rendered26, viewable27, or another impression basis; filters and who keeps the count
CPC Qualifying click Impressions may not lead to a click The click may be low-quality and not produce a conversion Click point, user initiation, deduplication, invalid-click filters
CPA An agreed action, such as a paid subscription Bring the user to the action and obtain valid attribution28/validation The action may have low margin, retention, or lifetime value29 Exact action name, attribution rule, validator, approval state
CPI30 Valid attributed install Delivery and clicks may not lead to a counted install The install may not launch, retain, or monetize Install definition, who performs attribution/validation, reinstall policy
CPL31 Qualifying lead Traffic may not produce an accepted lead The lead may not pass further qualification or become a customer Lead fields and criteria, duplicate/rejection rules, who accepts the lead

The table shows a shift of risk, not its disappearance: the deeper the billable event, the more delivery, conversion, and measurement32 risk the payee takes on, while the advertiser keeps the risk of the quality and monetization of the outcome. Fraud, attribution disputes, and caps can change the accepted count; their mechanics are covered later.

The billable event is set by contract policy, not automatically by the protocol. In the programmatic flow, IAB Tech Lab distinguishes the auction win from the event after which the auction becomes billable. OpenRTB also separates the win notice (nurl) from the billing notice (burl): a win does not guarantee delivery, and the billing signal depends on the policy of the publisher33 or exchange34. This is a programmatic example of the general rule, not a universal contract for all advertising.

The same instructional billable amount of $100 can be produced by five alternative, not simultaneous, contracts:

Contract Qualifying volume and rate Where the risk boundary passes
CPM 10,000 impressions at $10 CPM → $100 Impression
CPC 200 clicks at $0.50 CPC → $100 Click
CPI 20 installs at $5 CPI → $100 Attributed install
CPL 25 leads at $4 CPL → $100 Accepted lead
CPA 10 subscriptions at $10 CPA → $100 Attributed subscription

An equal amount does not mean equal risk. In CPM the buyer already owes $100 even if there are no subscriptions. In CPA for a subscription, the payee receives no charge for impressions and clicks that did not reach the accepted action.

The model is chosen by who sees which events and whose count is trusted. The publisher and the SSP35 observe the impression in their own systems but cannot see an install or a subscription inside the advertiser’s app, so they usually sell on CPM — the billable event lies inside their measurement boundary. An advertiser who does not trust the quality of someone else’s traffic but is willing to send conversion postbacks moves the billable event deeper — to CPI or CPA — and shifts the delivery risk onto the payee, usually in exchange for a higher price per event. The deeper the event, the more the payee depends on someone else’s count, so the contract must name the validator and the rejection rules.

Billing, optimization, and the realized metric are different objects

Imagine a product-specific flow: the advertiser actually pays for qualifying clicks, while the bidding system optimizes decisions toward a target CPA. After the period, the dashboard can compute the realized CPM, CPC, and CPA from actual spend and events.

optimization target: target CPA on a subscription
billing basis:       CPC per qualifying click
realized metrics:    observed CPM, CPC, and CPA after the period

Here target CPA does not turn the contract into CPA billing. A bid also expresses a decision input or willingness to pay, and does not prove an actual charge. The payment model establishes the contract; the realized average describes what has already happened; the optimization target influences future decisions.

How intermediaries and software providers earn

An intermediary can retain a portion of the media transaction, resell media with a markup36, or receive separate payment for a service/software. These models cannot be added together until the contract boundary and the calculation base are known.

Model Who pays Calculation rule Relation to the media dollar Main risk or caveat
Revenue share37 The party passing on proceeds under the contract A share of a named revenue/proceeds base for a period Usually divides the amount at a specific media boundary The base, allowable deductions, and adjustments must be named
Fixed fee38 The client of the service provider A pre-agreed amount for a period, scope of work, or deliverable Usually fully separate The provider carries the risk that the agreed scope will require more work
SaaS pricing39 The customer of the software platform Subscription, seats, usage tier, or processed volume Usually a separate software/service flow A usage event does not automatically become a media event
Markup The buyer pays the selling price to the intermediary Spread40 relative to the acquisition/base cost Yes, if the intermediary resells media The percentage is computed on cost (acquisition cost in the denominator), not on the selling amount
Take rate41 Depends on the boundary Retained amount / named transaction flow Usually characterizes retention at the chosen boundary Without numerator, denominator, and included fees, the rate is not comparable

Revenue share: a percentage without a base is undefined

Proceeds is the amount of transaction inflow before the agreed distribution at the chosen boundary. Revenue share is the contractual distribution of a named revenue or proceeds base. The contract must answer: whose proceeds, at which boundary, for which period, before or after which agreed adjustments, and who performs the calculation.

In the running example, the SSP receives $80 gross proceeds at its boundary and, under the seller contract, retains 15%:

SSP retained amount = $12 / $80 SSP-boundary proceeds = 15%
publisher share      = $68 / $80 SSP-boundary proceeds = 85%

The numerator of the first percentage is $12, the denominator $80. That is not 15% of the original advertiser spend of $100: relative to the original $100, the same $12 is 12%. The word gross here describes the transaction base before the split, not automatically gross revenue42 in financial reporting.

Markup, spread, and take rate

Spread is the absolute difference between the selling amount and the acquisition/downstream amount at a named boundary. If a managed network buys media for $80 and sells it to the advertiser for $100, the spread is $20.

Markup measures the uplift over the base cost:

network markup = $20 spread / $80 acquisition cost = 25%

Take rate measures the retained amount relative to a named transaction flow at the same boundary:

network take rate = $20 retained / $100 advertiser spend = 20%

25% markup and 20% take rate describe the same $20 spread with different denominators. Neither percentage can be compared with someone else’s rate without the participant boundary, numerator, denominator, and the list of included fees. Markup is also not gross margin43: gross margin uses accounting revenue and cost of revenue44, not arbitrarily chosen buy/sell amounts.

Fixed fee and SaaS: separate money flows

A fixed fee does not depend on the volume of the media transaction. Typical cases: an agency45 retainer for running campaigns46; a setup or managed-service fee; a minimum monthly commitment that the client pays even when volume falls short; a flat sponsorship fee for a placement over a period. It is chosen for budget predictability and to remove the “fee grows with spend” conflict of incentives: a provider paid a percentage of spend benefits from spending more. The risk is allocated differently than in media models: the provider risks that the agreed scope of work will take more effort, the client that it pays the same amount for low volume or a weak result.

A SaaS flow often runs alongside the media flow. Suppose the advertiser pays an ad server47 or measurement platform $2,000/month for software access. The contract shapes this in one of two ways: a fixed $2,000/month regardless of volume (a fixed fee), or a usage tier where the amount depends on the number of processed impressions (usage pricing, not a fixed fee — but also not a CPM purchase of inventory48). In both variants the amount:

  • is a service/software charge under a separate contract;
  • is not deducted from the instructional $100 media flow;
  • does not become a publisher payout49 or an intermediary take rate.

Including the $2,000 in the media dollar is valid only if a specific contract genuinely combines these charges. Otherwise you get two money flows: the media transaction and the SaaS subscription.

One transaction, different ledgers and points in time

Ledger — a particular party’s records of accrued amounts and obligations. A single transaction has no single neutral monetary label: one party’s inflow is another party’s outflow, and the intermediary simultaneously sells upstream and buys downstream. The same amount changes its label along the chain:

advertiser:    $100 spend / media cost
  → network:   $100 billed (receivable) | $80 downstream media cost (payable)
  → SSP:       $80 proceeds (receivable) | $68 publisher payout (payable)
  → publisher: $68 receivable
  • Advertiser spend — the amount charged to the advertiser for a chosen set of media/services. The dashboard may include or exclude platform/data fees, taxes, credits, and adjustments; the set must be named.
  • Media cost — the cost of acquired media for a particular party: for the advertiser in our narrow example $100, for the network the downstream acquisition cost of $80. Without a named owner, media cost is undefined.
  • Publisher payout — the amount that, under the seller contract, must be transferred to the publisher after the share, fees, and adjustments. For the SSP it is a payable; for the publisher, a receivable and part of its economics, and which part it recognizes as publisher revenue50, and in which period, is determined by its accounting policy.

Receivable — an amount a counterparty owes the company. Payable — an amount the company owes a supplier or another party. These words describe the direction of an obligation, not the moment cash moves.

It helps to distinguish states:

accrued  → qualifying events reconciled; the amount is earned or incurred under the contract
billed   → the amount is included in an invoice; a formal payment request exists
settled  → the obligation is discharged in cash or another agreed way

Taxes, credits, refunds, and FX can change the invoiced or settled amount, but their mechanics are not covered here.

Event and data flow

10,000 qualifying impressions
  → contract-specific billable count at the network and the SSP
  → participant event records with their own measurement boundaries
  → period reconciliation
  → ledger amounts: $100, $80, and $68 at different boundaries

Events can travel from publisher-/sell-side systems to buyers and reporting systems asynchronously. They prove the basis for the calculation, but no money moves along this path itself.

Invoice and cash flow

network → invoice $100 → advertiser
SSP     → invoice $80  → network
SSP     → statement $68 → publisher   (or publisher → invoice → SSP, per contract)

later, cash:
advertiser → $100 → network → $80 → SSP → $68 → publisher

Who issues the invoice and how settlement is arranged is set by the contract; the event path, invoice path, and cash path do not have to match in direction, participants, or timing.

Gross/net revenue and gross margin: a separate accounting boundary

Reporting entity — the company whose financial statements we are reading. Recognized revenue — the amount it presents as revenue under the applicable accounting policy; it is not just any invoice, collection, or cash inflow. Gross billings51 — a broader amount invoiced to customers or passed through collection before the exclusion of supplier components; gross billings are not a synonym for recognized revenue.

Gross revenue and net revenue are not “money before expenses” and “money after expenses”. The presentation in the financial statements depends on whether the company is a principal52 or an agent with respect to the specific promised good or service:

  • Specified good/service — the specific promise to the customer being assessed: for example, to provide advertising delivery or to arrange access to supply53.
  • A principal controls the specified good/service before it is transferred to the customer; under that conclusion, the agreed payment is usually presented gross as revenue, with the supplier component presented separately under the applicable cost policy.
  • An agent arranges for another party to provide the specified good/service; under that conclusion, revenue is usually the net fee or commission.
  • Control here means the ability to direct the use of and obtain substantially all of the benefits from the good or service before transfer — not the existence of an API, an invoice, or a product name.
What did the company promise the customer?
  → which specified good/service is transferred?
  → did the company control it before transfer?
      principal conclusion → revenue usually presented gross
      agent conclusion     → usually presented as net fee/commission

Primary responsibility, inventory risk, and pricing discretion can serve as indicators, but they do not form a mechanical scoring test. The conclusion is made for a specific promise and transaction: one AdTech company54 can be a principal in one flow and an agent in another.

One transaction economics, two mutually exclusive presentations

Take only the network boundary: $100 sell / $80 supplier component. The table is illustrative; in real reporting, one cannot choose a presentation for a prettier margin.

Illustrative conclusion Recognized revenue Supplier component as cost of revenue Gross profit before other CoR Gross margin before other CoR
Principal / gross $100 $80 $20 20%
Agent / net $20 fee $0 (the supplier’s $80 is outside revenue and CoR) $20 100%

Cost of revenue (CoR) — the expenses the reporting entity classifies as directly associated with obtaining recognized revenue under its policy. It is not just any cash outflow. Gross profit and gross margin are computed on one company and one accounting basis:

gross profit = recognized revenue − cost of revenue
gross margin = (recognized revenue − cost of revenue) / recognized revenue

In the net row, 100% is a mechanical result of the instructional isolation: $20 recognized revenue and $0 of other shown cost of revenue. It is not a benchmark and not a claim that a platform runs for free: a real company incurs infrastructure, data, and operations expenses, some of which may fall into cost of revenue under its policy, while operating expenses, interest, and taxes sit below gross profit and likewise do not turn a retained amount into profit.

The numerical coincidence of the 20% gross margin in the gross row with the network’s 20% take rate is also instructional: both times the calculation is $20 / $100, because there is no other cost of revenue. The definitions differ: take rate divides the retained amount by the transaction flow, while gross margin divides gross profit by the recognized revenue of the same company; under net presentation the same economics yield a 100% margin at the same 20% take rate.

Public filings show the contract-specific nature of this boundary:

  • In its Form 10-K for 2025, Magnite described net presentation for most platform transactions — the company does not control the inventory and does not set the auction price — and gross presentation for some insertion-order campaigns, where it considers itself the primary obligor for delivery.
  • In its Form 10-K for 2025, The Trade Desk separated gross billings and the related receivables/payables from net recognized revenue, with an agent conclusion for supplier components.

These are examples of specific companies’ policies for a specific period, not a universal classification of an SSP or DSP55. ASC 606, IFRS 15, the contract, and professional accounting judgment determine the actual conclusion; the chapter provides a mental model, not accounting advice.

Arbitrage: a spread in exchange for accepted risk

Arbitrage56, in the broad AdTech sense, arises when an intermediary buys media/traffic cheaper and resells it at a higher price, or buys on one pricing basis and sells on another. Its economics depend on the spread and the risk between boundaries.

A simple variant:

network buys batch media for $80
network sells the same agreed scope to the advertiser for $100
buy/sell spread = $20

In the chapter’s main example, the network buys per impression at $8 CPM and pays only for impressions actually delivered: it carries no volume risk, only price risk. Volume risk appears with a prepaid batch or a guaranteed commitment: if the network committed to pay downstream $80 but did not sell the whole agreed scope, or the customer did not accept the delivery, the spread shrinks or becomes a loss. Then it carries inventory/volume, fulfillment, and reconciliation risk to the extent set by the contracts.

Basis mismatch57 adds performance risk. An intermediary can buy impressions on CPM and sell clicks or actions on CPC/CPA. Its revenue then depends on the realized CTR58/CVR59 and validation. If clicks or actions occur worse than expected, the original media cost remains, and the downstream billable amount does not cover it.

Not every intermediary fee is arbitrage, and arbitrage does not automatically mean abuse. Transparent fees, revenue share, or a resale spread can pay for aggregation, reach60, service, and accepted risk. The opacity problem arises when the customer does not understand the buy/sell basis, the size of the spread, the intermediary’s function, or the conflict of incentives. The CMA’s 2020 study illustrates this mechanism in UK open display, but its rates cannot be used as a current benchmark.

The path of the instructional $100 from advertiser to publisher

Let us assemble the layers in one scenario. A subscription-app advertiser buys media in a mobile publisher’s app through a managed network and an SSP. All numbers are instructional, not a market benchmark and not a claim about a typical chain length.

Assumptions:

  • 10,000 identical qualifying billable impressions are needed only to isolate the economics;
  • the advertiser/network contract uses $10 CPM, the network/SSP contract $8 CPM;
  • the SSP/publisher contract splits the $80 proceeds: 15% SSP, 85% publisher;
  • the currency is USD; taxes, FX, refunds, rebates, fraud adjustments, and credit risk are excluded;
  • amounts are accrued after reconciliation, but invoices and cash are not yet settled;
  • the separate $2,000/month SaaS is not part of the media flow.

Contract arithmetic

Advertiser → network: 10,000 / 1,000 × $10 CPM = $100
Network → SSP:        10,000 / 1,000 ×  $8 CPM =  $80
Network spread:                                 =  $20
SSP retained share:             15% × $80       =  $12
Publisher payout:               85% × $80       =  $68
Check:                         $20 + $12 + $68   = $100

On one original advertising dollar, the contract economics are 20 cents retained by the network, 12 cents retained by the SSP, and 68 cents of publisher payout. In the full example, the intermediary retained amount is $20 + $12 = $32. This is the decomposition of specific contracts, not an industry benchmark and not accounting profit.

Denominators

network markup on acquisition cost = $20 / $80  = 25%
network take on advertiser spend    = $20 / $100 = 20%
SSP take on its boundary            = $12 / $80  = 15%
SSP share of original spend         = $12 / $100 = 12%
combined intermediary share         = $32 / $100 = 32%
publisher payout share              = $68 / $100 = 68%

Every line names its numerator and denominator; that is why SSP take = 15% and SSP share of original spend = 12% are true at the same time and answer different questions.

Each participant’s ledger

Participant Contractual basis Inflow / receivable Outflow / payable Economic remainder What the record does not prove
Advertiser $10 CPM × 10,000 — $100 advertiser spend/media cost — Invoice date, cash date, or business return
Managed network Customer charge $100; downstream $8 CPM $100 billed/receivable $80 downstream media cost/payable $20 spread That $100 is recognized revenue or $20 is profit
SSP Revenue-share base $80 $80 proceeds at the SSP boundary/receivable $68 publisher payout/payable $12 retained amount That $80 is recognized revenue or that 15% refers to the $100
Publisher 85% of SSP-boundary proceeds $68 payout receivable — $68 publisher economics That the amount is already settled cash or how exactly revenue is recognized

The labels change not because of inconsistency, but because of different reporting entities and counterparties. Transaction amount, retained amount, recognized revenue, gross profit, invoice, and cash answer different questions — about the contract, retention, reporting, profitability, the presented claim, and settlement; no record proves its neighbor.

Checklist for analyzing vendor economics

  1. Name the customer, payer, payee, counterparties, and the promised product/service.
  2. Record the pricing basis, qualifying billable event, rate, who performs validation, and the billable amount; separately, the optimization target and the realized metrics.
  3. Draw the participants’ ledgers: where the advertiser spend, media cost, publisher payout, receivable, and payable sit.
  4. For each fee or share, state the absolute amount, numerator, denominator, boundary, and allowable adjustments; separate media money from a fixed/SaaS charge.
  5. Check the accounting policy for the specific transaction type: specified good/service, principal/agent conclusion, recognized revenue, and cost of revenue; then compute gross profit and gross margin.
  6. Place accrual, reconciliation, invoice, and cash settlement on a timeline; do not infer the accounting presentation from who issues the invoice or from the direction of payment.

Common Confusions

  • Pricing model, optimization target, and realized metric are not the same thing. Target CPA can govern bids under CPC billing, while the observed CPA is computed only after the period.
  • An auction win is not necessarily a billable event. The charge arises after the event and validation defined by the contract; the win notice and the billing notice can be different signals.
  • A percentage without a numerator and denominator is undefined. An SSP share of 15% of $80 is the same $12 as 12% of the original $100; one spread of $20 yields a 25% markup on a cost of $80 and a 20% take rate on a spend of $100.
  • Gross/net revenue is not “before/after expenses”. The presentation follows a contract-specific principal-agent assessment; invoices and gross billings do not by themselves determine the conclusion.
  • Gross margin is not the take rate. The former uses recognized revenue and cost of revenue of one reporting entity; the latter, a retained amount and a named transaction flow. In the instructional gross row they coincide numerically (20%), but not by definition.
  • Not every intermediary income is arbitrage. Arbitrage presupposes a buy/sell spread or basis mismatch and the corresponding risk; a transparent fee or revenue share can pay for a separate function.

What to Remember

  1. Economics starts with the contract: the pricing basis and the billable event define the charge and allocate risk; the model is chosen by who sees the event and whose count is trusted.
  2. Revenue share, markup, and take rate are comparable only after the numerator, denominator, and participant boundary are disclosed.
  3. One flow of $100 → $80 → $68 simultaneously creates advertiser spend, an intermediary’s media cost, and a publisher payout; the retained $20 and $12 are transaction economics, not automatically recognized revenue or profit.
  4. Fixed fee and SaaS pricing are a separate money flow; they reduce the media dollar only if the contract combines them.
  5. Gross or net presentation is determined for a specified good/service through a principal-agent assessment, not by a product name; gross profit and gross margin are computed on the recognized revenue and cost of revenue of one company.
  6. Arbitrage yields a spread in exchange for volume, performance, or reconciliation risk and can produce a loss.
  7. The event flow, participants’ ledgers, invoice flow, and cash settlement are analyzed separately; the instructional $20 + $12 + $68 = $100 shows a method, not a typical market share.

Check Yourself

  1. An advertiser optimizes a campaign toward a target CPA but pays for qualifying clicks. What are the pricing basis, billable event, optimization target, and realized metric here?
  2. Why can $20 / $80 = 25% and $20 / $100 = 20% simultaneously and correctly describe the network’s economics?
  3. Can a company issue a $100 invoice, recognize $20 net revenue, and later receive $100 cash without a contradiction? Which layers does each amount describe?
  4. What contract, ledger, accounting, and settlement data are needed before comparing the gross margin and take rate of two AdTech vendors?

Sources and Further Reading

  1. IAB Tech Lab — Programmatic Auction Definitions, Final and OpenRTB 2.x Implementation Guidelines, sections 7.8–7.9 — definitions of Billable Event, the transaction ledger, and the distinction between nurl and burl; protocol guidance does not replace contractual billing policy.
  2. Google Ads — CPC, CPM, and Bidding basics — product-specific examples of pricing and optimization; this is not a vendor-neutral contract.
  3. FASB — Revenue from Contracts with Customers, Topic 606, paragraphs 606-10-55-36–55-40 and IASB — IFRS 15 PIR: Principal versus Agent Considerations, AP6B — the principal-agent mental model; application depends on the reporting framework, the contract, and professional judgment.
  4. FASB — GAAP Taxonomy Implementation Guide: Revenue from Contracts with Customers — gross profit as the difference between revenue and cost of revenue/cost of goods sold; the composition of cost of revenue remains company-specific.
  5. Magnite — Form 10-K for the year ended 2025-12-31 — company-specific examples of percentages, fixed CPM/monthly pricing, and different gross/net presentations; not an industry benchmark.
  6. The Trade Desk — Form 10-K for the year ended 2025-12-31 — the distinction between platform fee, gross billings, receivables/payables, and net revenue; the conclusions relate to this company’s contracts and policies.
  7. UK Competition and Markets Authority — Appendix M: Intermediation in open display advertising and official landing page — a historical analysis of UK open display from 2020; describes the mechanics of arbitrage and fee opacity, but its rates cannot be used as a current 2026 benchmark.

Sources checked on 2026-08-28.

Footnotes

  1. CPM (Cost per Mille) — the cost of a thousand counted impressions for the buyer: CPM = advertiser cost or spend / counted impressions × 1,000. The term is ambiguous without a contract: it can mean a pricing basis, a rate, a bid, or a post-fact observed average CPM; the observed CPM answers the question “how much did the buyer spend per thousand counted impressions.” ↩

  2. CPC (Cost per Click) — the cost or spend per qualifying clicks under a disclosed rule: CPC = cost / qualifying clicks. A bid/target CPC is not the actual average CPC: the former is a bid or an optimization target, the latter is a post-fact average. ↩

  3. CPA (Cost per Action) — the cost per advertiser-defined qualifying actions: CPA = cost / qualifying actions. The action is defined by the campaign contract — a trial, an order, a lead, or a subscription, not necessarily a new customer — so a CPA cannot be read without the name of the action or compared between campaigns directly. ↩

  4. Bid — a buyer’s offer to pay a certain price for a specific ad opportunity, usually expressed as a CPM. A bid is a pre-event decision: it is neither the final price (clearing price) nor the billable amount; in programmatic, the DSP forms a bid for each bid request separately. ↩

  5. Optimization target — the outcome toward which a bidding system tunes its future decisions: most often an optimization event (install, paid subscription) plus a goal for it. Not to be confused with a strategy parameter: target CPA $4 is a bid-strategy constraint, not the event the system predicts. It does not turn the contract into CPA billing: the payment model establishes the contract, the realized metrics describe what happened post-fact, and the optimization target influences only decisions before the event. ↩

  6. Pricing basis — the contractual rule for calculating a price: the object for which the buyer pays — an impression, click, action, install, lead, a share of proceeds, a period of access to software, or another agreed result. A metric’s name in an interface (CPM, CPC, CPA) does not by itself fix the basis: the same name can be a contractual price, a bid, or the actual average after the events, so only the basis named in the contract creates a monetary obligation. ↩

  7. Billable event — a qualifying event defined by the contract, after which the payee is entitled to accrue the billable amount — the sum under the pricing basis after counts, rates, validation, and contractual adjustments; it is not necessarily invoiced, recognized as revenue, or paid. The event is set by contract policy, not automatically by the protocol: an auction win, a win notice, and a billing notice are different points, and a win does not guarantee delivery or a charge. ↩

  8. Auction — the mechanism for choosing a winner among several bids for one ad opportunity and for determining the price (clearing price) under the auction’s rules — first-price, second-price, and their variants. An auction win is the event of choosing a winner; it is not an impression and does not by itself make the impression billable. ↩

  9. Programmatic — automated buying and selling of advertising: decisions about each opportunity are made by systems, not people. RTB is only one of its mechanisms; programmatic also includes automated reserved and direct mechanics without an auction. Programmatic is part of AdTech, but not all of AdTech. ↩

  10. Impression — a delivery or display event for the selected ad, recorded by a system under a specific rule; a counting unit, not a fact of human perception. Served, rendered, and viewable are different delivery states, so the counters of different participants in the same deal may not match. ↩

  11. Ledger boundary — the boundary within which an entity recognizes and records amounts and events under its own rules: the same amount is revenue on one boundary and a downstream media/traffic cost on another. Because of different ledger boundaries, the reports of the advertiser, the platform, and the publisher about the same deal naturally do not match and do not have to match. ↩

  12. Media spend — the money an advertiser allocates and spends on buying advertising. It passes through the media sellers and ultimately forms publisher revenue, shrinking along the way by fees; it is not a synonym for publisher revenue, nor for intermediary fees — these are different amounts in one chain. ↩

  13. Retained amount — the part of the transaction flow remaining with a participant on its boundary before its own costs. It characterizes transaction economics but is not recognized revenue or profit: the accounting presentation and the expenses are determined by the entity’s separate policy. ↩

  14. Advertiser — a market participant that funds advertising for a business result: installs, subscription purchases, growth in awareness. It buys not a person but the opportunity to address an audience in a specific context; it creates demand and sets goals, budget, and creative. ↩

  15. Cost — the expenses of the entity whose economics are being analyzed, in the chosen scope: for an advertiser, the media cost; for a platform, the downstream cost; for a business, a broader set of expenses. Not a synonym for spend: spend is the buyer’s outlay on media, while cost can include non-payment and out-of-media components and is always tied to a perspective. ↩

  16. Intermediary — a participant that connects market sides or performs a specialized function: agencies, DSPs, ad networks, exchanges, SSPs, ad servers, measurement, verification, and data providers. It is justified when the value of its function (reach, aggregation, interoperability, speed, measurement, risk control) exceeds the cost and complexity; it need not be present in every deal and does not necessarily retain a share of the media money — some intermediaries receive a separate service or SaaS fee. ↩

  17. Revenue — the amount recognized by the chosen entity under its accounting rules on its ledger boundary. Estimated dashboard value, invoiced amount, settled amount, and recognized revenue can be different states of one chain; for comparison, the period, timezone, currency conversion, taxes/refunds, and participant must be agreed. How much of the money passing through a platform it recognizes as revenue depends on its principal or agent role (gross/net presentation), not on the size of the invoice. ↩

  18. Settlement — the actual movement of money between participants for a period, as distinct from the invoice (the issued claim) and from accrual. One amount passes through three states: accrued — recorded from events, billed — issued in an invoice or statement, settled — paid; at each step the amount may change because of adjustments, fraud filtering, credits, and FX. For one side it is a receivable (expected inflow), for the other a payable. ↩

  19. Ad network — an intermediary that aggregates advertising opportunities (inventory) from many publishers and sells them to advertisers as a single packaged product: the network takes over supply selection, pricing, and delivery, while the buyer gets one manageable offering instead of a dozen separate contracts. Prices inside the network are usually opaque to the buyer. Historically, many ad networks added auction mechanics and rebuilt themselves as ad exchanges, so both functions are often combined under one brand. ↩

  20. Click — a qualifying user-initiated interaction with an ad, recorded under a specific system’s rule. Even a single action has several measurement points: initiated, measured (click tracker), received (destination), and resolved (landing page) clicks need not match due to dropped connections, redirect chains, and filters. ↩

  21. Proceeds — the amount of the transaction inflow before the agreed distribution on the chosen boundary. The word gross next to proceeds describes the base before the split, not automatically gross revenue in financial reporting: for calculating shares, the boundary, the period, and the permitted deductions matter. ↩

  22. Ad opportunity — a specific opportunity to select and show an ad that arises when a user reaches a placement. A short-lived event: in a digital environment it can vanish faster than a person can make a decision, so systems process it in milliseconds, and in programmatic it is put up for bidding as a separate unit. ↩

  23. Payee / payer — the two sides of one payment obligation: the payer pays for the billable event, the payee receives the payment. Risk allocation in a pricing model is described through them: under CPM the payee earns the right to charge early in the funnel and the post-impression risk sits with the payer; under CPA the payee takes on all the risk up to the conversion. One company can be the payee in one contract and the payer in another. ↩

  24. Conversion — a valuable action or outcome that the advertiser has defined as significant (install, trial, first payment, renewal, purchase) and that a system has counted under its measurement and attribution rules. Conversions cannot be summed without an explicit counting rule: one click can lead to several conversions. An observed conversion is a recorded event; an attributed conversion is an event to which an attribution rule assigned credit; neither proves causality. ↩

  25. Served impression — an impression counted on the fact that the ad server or selling platform returned the creative (or a reference to it) in response to a request; it proves sending, but not loading, rendering, or visibility. The earliest and largest number in the served → rendered → viewable ladder. ↩

  26. Rendered impression — an impression counted after the creative has loaded and been rendered on the user’s device (usually on a signal from client-side code: a pixel or an SDK event after render). It proves rendering, but not that the creative entered the visible area of the screen or was noticed; the MRC standard for display counts an impression precisely on client-side load/render. ↩

  27. Viewable impression — an impression that met the applicable geometry/time criteria of viewability (the classic MRC benchmark: at least 50% of the creative’s pixels on screen for one continuous second for display and two for video). Viewable means a standardized opportunity to see, not a person’s attention; non-viewable (the measurement took place, the criteria were not met) is not the same as non-measurable (the status could not be determined). ↩

  28. Attribution — the rules by which the available advertising touchpoints (impression, click) are matched to outcomes (install, registration, purchase) and receive credit for the conversion. Attribution answers the question “which source was credited with the action under the chosen rules,” not “whether the ad was the true cause”: credit does not prove causality. Attribution providers are a separate cross-cutting category; in mobile, the MMP. ↩

  29. LTV (Lifetime Value) — the cumulative observed or predicted value of a cohort’s members: LTV = cumulative observed or predicted value / cohort members. It requires a cohort, a horizon, an identity rule, a value basis (revenue or contribution), and an observed/predicted status; revenue LTV does not subtract costs, and contribution LTV is calculated on the agreed marginal basis. Without these definitions, “a user’s LTV” is not a defined scalar. ↩

  30. CPI (Cost per Install) — the cost of mobile acquisition per qualifying installs: CPI = cost / qualifying installs. The install count depends on the measurement and attribution rules, so a CPI is comparable only under named install-counting rules. ↩

  31. CPL (Cost per Lead) — the cost per business-defined qualifying leads: CPL = cost / qualifying leads. Lead definitions and quality differ between businesses, and a lead is not yet a customer, so a low CPL is useless without qualification. ↩

  32. Measurement — the function of recording and aggregating advertising events (impressions, clicks, outcomes) for reporting and optimization. Measurement providers give an independent point of observation; different participants have different timestamps and counting rules, so reports do not have to match perfectly. In mobile, the dedicated category is the MMP. ↩

  33. Publisher — the owner or operator of a media environment where advertising opportunities arise: a website, a mobile app, a video service, a game. It monetizes the environment with advertising, creates supply, and decides which placements, formats, and buyers to admit, balancing revenue against user experience. ↩

  34. Ad exchange — the transaction layer that connects buyers and sellers in an automated deal, usually through a real-time auction (RTB). The difference from an ad network is transparency: the buyer sees the description of an individual opportunity and decides for themselves how much to pay for it, and the price is set by the auction rather than by the network’s price list. An exchange is not necessarily a separate company between the DSP and SSP: exchange and SSP functions are often combined in a single product. ↩

  35. SSP (Supply-Side Platform) — a platform that helps publishers manage the sale of inventory: it sets buyer access rules, connects programmatic demand (DSPs, exchanges), and runs or forwards the auction. SSP and exchange functions are often combined in a single product. ↩

  36. Markup — a surcharge on the base cost expressed as a share of the cost denominator: markup = (selling amount − acquisition cost) / acquisition cost. It is not the spread (the absolute difference) and not the gross margin (which uses the reporting entity’s accounting revenue and cost of revenue); comparable only when the buy/sell amounts are named. ↩

  37. Revenue share — the contractual distribution of a named revenue or proceeds base between the parties in a specified share. Without disclosure of the base it is undefined: you need to know whose proceeds, on which ledger boundary, for which period, before or after which agreed adjustments, and who performs the calculation; a share of one boundary does not equal a share of the original advertiser spend. ↩

  38. Fixed fee — an amount agreed in advance for a period, a scope, or a deliverable that does not depend on the volume of media transactions: an agency retainer, a setup or managed-service fee, a minimum monthly commitment, a flat sponsorship fee. It is usually a flow separate from media money: it is not deducted from the media dollar and does not become a publisher payout or a take rate — unless a specific contract bundles software and media charges into one invoice. It is chosen for predictability and to remove the “fee grows with spend” conflict; the provider bears the risk that the agreed scope will require more work. ↩

  39. SaaS pricing — the payment model for a software platform: subscription, seats, usage tiers, or processed volume for access to the product. It forms a separate software/service flow: the platform’s usage event does not automatically become a media event, and a SaaS charge does not turn the relationship into a purchase of inventory. ↩

  40. Spread — the absolute difference between the selling amount and the acquisition/downstream amount on a named boundary. The economic core of an intermediary’s resale and arbitrage models; the markup and the take rate are computed from the same spread with different denominators and are not comparable without the named boundary. ↩

  41. Take rate — the retained amount relative to a named transaction flow on the chosen boundary: take rate = retained amount / named transaction flow. Comparable only when the numerator, denominator, and included fees are disclosed: the same amount yields different rates relative to the boundary proceeds and the original advertiser spend. ↩

  42. Gross/net revenue (presentation) — a variant of how the reporting entity presents revenue: gross shows the agreed payment in full, net only the fee or commission. The choice follows the contract-specific principal-agent assessment for the specific specified good/service, not the logic of “money before/after expenses”: the product’s label, the invoice, or gross billings do not by themselves determine the conclusion. ↩

  43. Gross margin — the share of recognized revenue remaining after the cost of revenue of the same reporting entity on the same accounting basis: gross margin = (recognized revenue − cost of revenue) / recognized revenue; gross profit is the numerator of this formula. Not the same as the take rate, which divides the retained amount by a named transaction flow. ↩

  44. Cost of revenue (CoR) — the expenses that the reporting entity classifies as directly connected with earning the recognized revenue under its policy. Not just any cash outflow: the composition is determined by the company’s policy, and operating expenses, interest, and taxes sit below gross profit. ↩

  45. Agency — a contractor company that plans and executes marketing and media work on behalf of an advertiser: develops strategy, prepares creative, selects channels, launches a campaign, and reports on results. An advertiser can keep media buying in-house, hand it to an agency in full, or split the responsibility. An agency is an organization and a contractual role, as opposed to a media buyer — an operational function that can also work inside an advertiser. ↩

  46. Campaign — an organized set of an advertiser’s advertising activity with a goal, a budget, delivery rules, and creatives. Before delivery, its conditions are communicated to the buying systems; objects and events are linked by identifiers such as campaign_id. ↩

  47. Ad server — a system that selects, delivers, and logs advertising under given rules. There are publisher-side ad servers (matching direct campaigns, networks, and programmatic demand, managing delivery) and advertiser-side ones (storing creatives, counting impressions and clicks); the term denotes a function, not a market side. ↩

  48. Advertising inventory — the set of available or forecast opportunities to show advertising, bounded by the environment, the slot, the format, the time, the publisher’s rules, and the permissible audience context. Inventory is not a list of users and not a warehouse of pre-produced impressions: an opportunity arises and disappears at the moment the product is used. ↩

  49. Publisher payout — the amount that, under the seller’s contract, must be transferred to the publisher after revenue share, fees, and adjustments; for the SSP or the network it is a payable, for the publisher a receivable. The payout is a contractual and payment quantity; which part of it, and in which period, the publisher recognizes as revenue is determined by its accounting policy (see publisher revenue). ↩

  50. Publisher revenue — the amount the publisher recognizes as revenue under its own accounting rules from media payments for its inventory; in the simple case it equals the publisher payout, but recognition (accrued / billed / settled) follows the publisher’s policy, not the moment of payment. It differs from advertiser spend by the sum of all fees in the chain; participants’ service and SaaS fees outside the media money do not count toward it. ↩

  51. Gross billings — the amount invoiced to clients or passed through collection before excluding supplier components; a broader quantity than recognized revenue and not its synonym. Billings describe billing, not the principal/agent accounting conclusion. ↩

  52. Principal / agent — the accounting assessment of the reporting entity’s role with respect to a specific promised specified good/service: the principal controls it before transferring it to the customer and presents the payment gross; the agent arranges for its provision by another party and presents the net fee/commission. Control is the ability to direct the use and obtain significant benefits before transfer; primary responsibility, inventory risk, and pricing discretion are evidence, not a mechanical score; one company can be a principal in one flow and an agent in another. ↩

  53. Supply — the advertising opportunities available from publishers; the market side around which the sell side / supply side groups — the people and systems that help describe, sell, and deliver advertising (SSPs, publisher ad servers). Supply is the set of inventory for sale, not the name of a platform. An ad network aggregates both supply and demand, so it sits between the sides rather than inside one of them. ↩

  54. AdTech (advertising technology) — the set of technologies that coordinate the buying, selling, selection, delivery, measurement, and payment for advertising between advertisers, publishers, and intermediaries: from the ad server and DSP/SSP to measurement and verification systems. Programmatic is only part of AdTech: direct deals, reserved buying, and closed platforms also run on AdTech. The boundary with MarTech is drawn by primary purpose: AdTech coordinates paid media, MarTech the customer lifecycle. ↩

  55. DSP (Demand-Side Platform) — a platform that automates the purchase of advertising opportunities for an advertiser: it aggregates access to several supply sources, evaluates each opportunity, and decides whether to bid and how much to pay. It belongs to the demand side not because it “creates demand” but because it acts in the buyer’s interest. ↩

  56. Arbitrage — an intermediary model that buys media/traffic cheaper and sells it dearer, or buys on one pricing basis and sells on another. The economics rest on the spread in exchange for accepted risk: volume/fulfillment, basis mismatch, and reconciliation risk can turn the spread into a loss. Not every intermediary fee is arbitrage, and arbitrage is not automatically abuse; the problem is the opacity of the buy/sell basis and the spread to the customer. ↩

  57. Basis mismatch — buying and selling the same media on different pricing bases, for example buying impressions on a CPM and selling clicks or actions on a CPC/CPA. It adds performance risk to arbitrage: the intermediary’s revenue depends on the realized CTR/CVR and validation, while the original media cost remains if the downstream billable amount does not cover it. ↩

  58. CTR (Click-Through Rate) — the share of selected impressions after which the selected clicks were recorded: CTR = qualifying clicks / counted impressions × 100%. The number depends on the click/impression basis (served or rendered, filtered or all) and the population; a high CTR by itself does not indicate quality traffic, conversions, or causal lift. ↩

  59. CVR (Conversion Rate) — the share of qualifying conversions among the declared eligible precursor events: CVR = qualifying conversions / declared eligible precursor events × 100%. A bare CVR hides the transition: only named transitions such as click-to-install or install-to-trial are comparable; a value above 100% is possible with several conversion actions on one interaction. ↩

  60. Reach — the de-duplicated number of unique entities (persons, devices, households, accounts, or a modeled audience) that received at least one qualifying exposure in a period, or the share of such entities in an explicitly defined population: reach rate = reached unique entities / declared universe × 100%. Comparable only with identical entity definition, exposure rule, population, de-duplication scope, and period. ↩